The Money Mustache Community
General Discussion => Welcome and General Discussion => Topic started by: jamesplease on April 16, 2020, 12:19:02 AM
-
In my free time I've been working on a new calculator that simulates retirements using historical data. The idea is similar to the esteemed firecalc or cFIREsim calculators, if you're familiar with them. My calculator
is still early on (it's come a long way and has a lot of features now), but I'd love to get your feedback!
You can check it out here: https://ficalc.app (https://ficalc.app)
Here are the main reasons I made it:
- The existing calculators don't have mobile support. This one is designed to work on phones as well as laptops/desktops
- I wanted to improve the user experience. Existing calculators have amazing functionality – but you first need to learn how to use them to unlock their potential. I find it to be the case that mastering these apps can be a challenging task. Part of this is the inherent complexity of what they offer, but I also think the design of the apps contributes to the challenge. I hope that the design of FI Calc helps make the process of mastering it easier.
There are still a lot of features I need to add. Here's some (but not all) of what I'm planning:
- additional portfolio asset types (i.e.; bonds, cash) (all added)
- additional withdrawal plans, including:
- VPW (added)
- CAPE/CAEY-based (added 4/16/20)
- Hebeler Autopilot (added 5/31/20)
- Min/max withdrawals for all %-based sims (added 4/16/20)
- the ability to share/save simulations (added 5/31/20)
- export the results to a CSV file (added 4/16/20)
- restrict results to certain years. For instance, if you only want to consider the period from 1920 - 2020 in your simulations, rather than from 1871 - 2020 (added 4/18/20)
- additional withdrawals for specific years, such as a future college expense, or other large purchases that may not be factored into a yearly withdrawal (added 4/19/20)
- additional income for specific years (added 4/20/20)
- a Monte-Carlo option (as an alternative to historical data)
- real estate investment support (guides added on the website explaining current support)
- user-customizable analysis. This will allow you to do basic analysis within the app of the things that matter to you
- US state/federal tax support
Potential future ideas (what do you think of these?):
- Additional data sources. While researching these withdrawal plans, I noticed that different studies use different data sources. It's not completely fair to compare any two studies without using the same exact data for them. I don't think that everybody will care to get this detailed, but it might be useful to someone.
If you have time to check this app out, give any feedback, or suggest new ideas, I would appreciate it! @themagicman has already helped a ton by giving me a few suggestions :)
Update on the following section: the app currently supports "saving" simulations through bookmark able URLs.
One thing I'm curious to know from you all: adding the ability to save simulations would give me a monthly cost. Although I'm not looking to make money from this project, I would want to at least cover upkeep if I were to add this feature, so that it's sustainable for me to keep the app running over the long run. Is that a feature you would potentially consider paying a small, one-time amount to gain access to? If so, what do you think would be a fair price for that?
If it helps you to pick a number, I'm anticipating the monthly costs to be $10-$20, plus $20 a year for the domain. Ads are another option to support the feature, but my preference would be to keep this thing ad-free.
-
The changes are looking good! A couple of additional things that would be nice! I was going to suggest the CSV file export, so I am glad it is on your list already!
All have the variable spending methods having a floor or ceiling. For example, Guyton Klinger or 95% rule having an inflation adjusted floor that the spending would never go below.
I think the additional data sources would be useful. Also, I am not sure how difficult it is to implement but I would find a monte carlo simulator option very useful (I have never come across a calculator that allows for variable spending with a monte carlo simulator)
One option is to have ads and pay a one time donation to remove those ads of $x
-
Just released an update: a CAPE-based withdrawal plan has been introduced
The changes are looking good!
Hey, thanks :)
All have the variable spending methods having a floor or ceiling. For example, Guyton Klinger or 95% rule having an inflation adjusted floor that the spending would never go below.
Seems reasonable. Added to the list!
I think the additional data sources would be useful.
Awesome!
Also, I am not sure how difficult it is to implement but I would find a monte carlo simulator option very useful (I have never come across a calculator that allows for variable spending with a monte carlo simulator)
This is totally doable. I had actually been considering it, but I figured I'd omit it from my todo list unless someone brought it up. I'll probably work on some more "core" features first and then revisit this.
One option is to have ads and pay a one time donation to remove those ads of $x
I like this idea a lot 👍
Thanks for the great suggestions, @themagicman !
-
Just used it on my phone for the first time! Love the interface on mobile. It was quick and very easy to use!
-
Just used it on my phone for the first time! Love the interface on mobile. It was quick and very easy to use!
That's what I want to hear! :)
Here are a few updates from tonight:
- all %-based withdrawal plans now support min/max withdrawals
- results can now be exported as CSV files
-
I saw the CSV output. It is clean and easy to read. One suggestion (stop me if my suggestions are annoying you ha), is to have this CSV output as well as a more detailed output for CSV. I would like to have a csv output like cfiresim, it shows the year by year withdrawal amount and the investment balance each year for each starting year.
I use this for two reasons,
1) To ensure that the software is preforming the way that I expected. I can see that the floor is being following or my guyton klinger guardrails are being followed, how I interpret it.
2) I can run some analysis outside of the system. For example, on variable spending plans, I like to look at what percentage of the overall time I spend within 10% of my starting amount, how much within 20% and how much is more than 30%
-
(stop me if my suggestions are annoying you ha)
Please, keep them coming!!! They've been great, and I appreciate them a ton!
One suggestion is to have this CSV output as well as a more detailed output for CSV.
Great idea. I'll add it!
The latest feature lets you restrict the calculator to only use some of the historical data. For instance, if you want you can only run sims between 1920-2020.
The next big features will be:
- additional withdrawals (added!)
- additional income (added!)
- supporting bonds in your portfolio
-
There might be an issue with your additional withdraw feature. I can't tell easily what the issue is, because I am not sure the total amount each year. but it is not working as expected. I am getting much larger fail rates then I expected . It could be user error but wanted to see if you were seeing any abnormalities with it
-
It’s entirely possible that there’s a bug with the algorithm right now. I need to write more tests around these new features so that I have more confidence in them.
If you have time, could you share with me some of the settings that you’re using that are producing unexpected results? i.e.; how many withdrawals and for how long? Also the fail rate that you’re seeing, and what might expect? If you’re busy, though, no worries – I’ll take a closer look.
Thanks for the report!
Edit: I also plan to update the CSV exports both with the feature you suggested above, as well as detailed info related to additional income/withdrawals, which should help with understanding what’s going on each year.
-
See attached! It might just be the plan actually fails and the tool has helped me discover that but just seems odd to me that these would not be relatively close. They both allow for $15k of drop at the same rate but one has the mortgage as an additional withdraw (for 25 years) and the other just has it in the normal spending amount. I guess the mortgage portion does not drop by the 3% in the second one, but it is also not inflation adjusted, so I am not sure. I could just be missing something as to why but I just do not understand how the chance of success would vary by that much.
Yes, for sure the CSV file will help! If I could see what it is doing year by year it would confirm that it is functioning how I thought the plan would or not!
-
Also, CFIRESIM does not have the option to do the 95% rule with an overall floor (thank you for putting that in! I have not seen that anywhere else) , so I am unable to verify there.
-
Amazing, thanks for the screenshots @themagicman ! I'll dig into this tonight to see what's up. And I might have time to do the CSV update, too.
Edit: quick update. Pretty sure this is a bug with my 95% Rule algorithm when combined with additional withdrawals. I'll release a fix tonight if it is, and write up a more detailed report here explaining what's going on :P
-
Alright, a fix should be deployed. Let me know if the results that you're seeing now are lining up more with your expectations.
Here's a summary of what was going on, if you're interested in reading about it:
The way I wrote the Guyton-Klinger and 95% rule implementation was creating a runaway withdrawal effect when combined with additional withdrawals. Your 25 year mortgage was a long enough retirement length that this runaway withdrawal caused nearly every simulation to fail.
Here's what was happening in more detail to create the runaway effect:
- Additional withdrawals are added to whatever the yearly withdrawal is determined to be. So first I run the 95% Rule, Guyton-Klinger, or whatever, and then tack on extra money. In your case, that was an extra $12k each year for the mortgage.
- With GK and 95% rule, each year's withdrawal is sometimes determined by the previous years' withdrawal. For the 95% rule, you can withdraw 95% of the previous year's withdrawal if it's greater than the % of your portfolio that you specify.
The combination of these two behaviors led to the runaway effect. For your calculation, 95% of (previousYearWithdrawal + $12k) was always bigger than 3.7% of the current portfolio value, so that was the path being taken each time. And so the spending just rose and rose until the portfolio was exhausted.
The fix that I implemented is that GK/95% rule are calculated using the "base" withdrawal of the previous year, which is the value of the previous year's withdrawal without additional withdrawals.
Also, a few CSV export updates:
- from the overview page, the CSV export is now a full export. I'll add back in a summary export soon.
- the CSV exports now include information about additional withdrawals (but not additional income. That and other information will be added soon)
-
That makes sense! The changes look much better now! Went from a 20% success rate to 95% ha.
I like the way the detail comes out in the csv. Not sure if you have seen CFIRESIMs or not but I like this much better. I like that it has continuous rows (without blanks) and theirs does not. They also have a lot more data then I feel is needed on the export which makes it not as clean.
The only suggestion that I would have (and this is nit picking) on the CSV is either replacing "simulation number" or adding an additional column with the starting year for the simulation. So the year that you click on, on the right "simulations by starting year" would be listed 30 times down that column, if that makes sense. Again, this is not really a big deal and something I (or anyone) could easily do themself after dumping to CSV (and I do it myself on the CFIRESIM CSV), but I find it useful because I can make more sense of the data when viewing in excel. For example, in my withdraw plan, the only time my spending would hit the spending floor is in the 1930's and 1970's or something like that.
As a side note, I am not sure why others have not been commenting on this calc yet. I am honestly a little perplexed ha. Maybe it is because it is in the welcome part, or maybe people have just not gone to the site yet.
This is already the best retirement calculator out there I have seen IMHO (and I have searched for a lot) and you are still adding to it and making improvements. I will make some posts about it on reddit. I think you have a really good opportunity to be "the fi calculator" that people use and bloggers talk about in their analysis.
-
That makes sense! The changes look much better now! Went from a 20% success rate to 95% ha.
Nice! :)
I like the way the detail comes out in the csv. Not sure if you have seen CFIRESIMs or not but I like this much better. I like that it has continuous rows (without blanks) and theirs does not. They also have a lot more data then I feel is needed on the export which makes it not as clean.
Yeah, I did reference cFIREsim's export after you mentioned it, and thought about how I could improve upon it in FI Calc.
The only suggestion that I would have (and this is nit picking) on the CSV is either replacing "simulation number" or adding an additional column with the starting year for the simulation. So the year that you click on, on the right "simulations by starting year" would be listed 30 times down that column, if that makes sense. Again, this is not really a big deal and something I (or anyone) could easily do themself after dumping to CSV (and I do it myself on the CFIRESIM CSV), but I find it useful because I can make more sense of the data when viewing in excel. For example, in my withdraw plan, the only time my spending would hit the spending floor is in the 1930's and 1970's or something like that.
No detail is too small or nitpicky. This is great feedback, and I agree that it should be there. I'll add it tonight!
As a side note, I am not sure why others have not been commenting on this calc yet. I am honestly a little perplexed ha. Maybe it is because it is in the welcome part, or maybe people have just not gone to the site yet.
This is already the best retirement calculator out there I have seen IMHO (and I have searched for a lot) and you are still adding to it and making improvements. I will make some posts about it on reddit. I think you have a really good opportunity to be "the fi calculator" that people use and bloggers talk about in their analysis.
Hey, thanks a lot! I've put a ton of effort into this already, and I plan to put a lot more. I was (and still am) a little bit worried that people won't be interested. I'm hoping over time, as I add more features, interest will grow...we'll see, I guess.
Feel free to post it on Reddit btw. I was planning to do the same once I added a few more "key" features (like supporting bonds), but I think it's in a good enough state if you wanted to share it. If you do end up posting it, ya gotta share the link here so I can see all of the comments! (hopefully they're not too mean)
Btw, your awesome suggestions have definitely gone a long way to getting this calculator to where it is today. I really appreciate you taking the time to let me know what things can be improved...so...thank you! :D
-
I've just given it a try, looks very slick, thanks for sharing it.
Is there an issue with the Historical Data? When I use the tick box "Use all available historical data" I get data from 1871 to 1975. If I set "Only use data between" 1900-2000 I only get data from 1900 to 1955 displayed? I had Length of Retirement set to 45 years, changing this variable changes how much data is displayed... Have I missed something or is it broken?
-
I've just given it a try, looks very slick, thanks for sharing it.
Is there an issue with the Historical Data? When I use the tick box "Use all available historical data" I get data from 1871 to 1975. If I set "Only use data between" 1900-2000 I only get data from 1900 to 1955 displayed? I had Length of Retirement set to 45 years, changing this variable changes how much data is displayed... Have I missed something or is it broken?
Hey @Shaz_Au ! Thanks for trying it out.
The behavior that you're seeing is expected. Let me explain it a bit more:
The calculator can only run simulations when there is enough historical data to cover the entire retirement length. What this means is that for a retirement that's 45 years long, the latest start year that you will see is 1975 (2020 - 45 years = 1975) when you're using all of the available data. Similarly, if you restrict the years to be 1900-2000, then the latest start year that you'll see will be 2000 - 45 = 1955.
If you click into one of the years that you are seeing (which, again, represents the start year of a particular retirement simulation), you will see more details, and it will show how the portfolio evolves over the duration of the specified retirement. So clicking into, say, the 1975 year from your first calculation will show a retirement spanning 1975 - 2020.
One other thing that might help you understand what is going on is to set the length of retirement to be 1 years. Then you'll see start dates all of the way up until 2019. But, again, 2020 won't be displayed because if that one-year retirement started in January 2019 then it ended in January 2020.
Does that help clarify things?
-
I've just given it a try, looks very slick, thanks for sharing it.
Is there an issue with the Historical Data? When I use the tick box "Use all available historical data" I get data from 1871 to 1975. If I set "Only use data between" 1900-2000 I only get data from 1900 to 1955 displayed? I had Length of Retirement set to 45 years, changing this variable changes how much data is displayed... Have I missed something or is it broken?
Hey @Shaz_Au ! Thanks for trying it out.
The behavior that you're seeing is expected. Let me explain it a bit more:
The calculator can only run simulations when there is enough historical data to cover the entire retirement length. What this means is that for a retirement that's 45 years long, the latest start year that you will see is 1975 (2020 - 45 years = 1975) when you're using all of the available data. Similarly, if you restrict the years to be 1900-2000, then the latest start year that you'll see will be 2000 - 45 = 1955.
If you click into one of the years that you are seeing (which, again, represents the start year of a particular retirement simulation), you will see more details, and it will show how the portfolio evolves over the duration of the specified retirement. So clicking into, say, the 1975 year from your first calculation will show a retirement spanning 1975 - 2020.
One other thing that might help you understand what is going on is to set the length of retirement to be 1 years. Then you'll see start dates all of the way up until 2019. But, again, 2020 won't be displayed because if that one-year retirement started in January 2019 then it ended in January 2020.
Does that help clarify things?
I was actually thinking about this a week ago. I feel like analysis that I do is somewhat flaws because it is not including those years, some of which are most likely going to be failures in my plans (2000,2001,2002,2007). This is how all calculators work, not just yours, but it would be nice to have an idea on those. One idea that I was thinking was a box that would fill in historical data for returns and inflation for the years that we do not have data for. For example, on a 30 year time horizon and that boxed checked. The year 2000 scenario would use 2000-2020 actual data and then from 2021-2030 it would use the historical average (or even a user input number) for returns and inflation. Thoughts?
-
I've just given it a try, looks very slick, thanks for sharing it.
Is there an issue with the Historical Data? When I use the tick box "Use all available historical data" I get data from 1871 to 1975. If I set "Only use data between" 1900-2000 I only get data from 1900 to 1955 displayed? I had Length of Retirement set to 45 years, changing this variable changes how much data is displayed... Have I missed something or is it broken?
Hey @Shaz_Au ! Thanks for trying it out.
The behavior that you're seeing is expected. Let me explain it a bit more:
The calculator can only run simulations when there is enough historical data to cover the entire retirement length. What this means is that for a retirement that's 45 years long, the latest start year that you will see is 1975 (2020 - 45 years = 1975) when you're using all of the available data. Similarly, if you restrict the years to be 1900-2000, then the latest start year that you'll see will be 2000 - 45 = 1955.
If you click into one of the years that you are seeing (which, again, represents the start year of a particular retirement simulation), you will see more details, and it will show how the portfolio evolves over the duration of the specified retirement. So clicking into, say, the 1975 year from your first calculation will show a retirement spanning 1975 - 2020.
One other thing that might help you understand what is going on is to set the length of retirement to be 1 years. Then you'll see start dates all of the way up until 2019. But, again, 2020 won't be displayed because if that one-year retirement started in January 2019 then it ended in January 2020.
Does that help clarify things?
I was actually thinking about this a week ago. I feel like analysis that I do is somewhat flaws because it is not including those years, some of which are most likely going to be failures in my plans (2000,2001,2002,2007). This is how all calculators work, not just yours, but it would be nice to have an idea on those. One idea that I was thinking was a box that would fill in historical data for returns and inflation for the years that we do not have data for. For example, on a 30 year time horizon and that boxed checked. The year 2000 scenario would use 2000-2020 actual data and then from 2021-2030 it would use the historical average (or even a user input number) for returns and inflation. Thoughts?
This is a complicated problem that all of these calculators must examine in some way (including truncating like this one does which is the easiest way to code it), but one worth thinking about. If you read EarlyRetirementNow's safe withdrawal rate series you can see that they've done this experiment you described (via spreadsheets) to get a lot more data. It's useful, particularly since this is the most recent and probably most relevant data.
A thing to consider though, is that you might want a way to tell the user / weigh results based on e.g. "95% of this run's results are using averages, not actuals" because as we all know that makes a huge difference. I.e. if you started a simulation retiring in 2017, 27 of the 30 (n, really) years or whatever from that run will be historical average and it's pretty straightforward to say it will succeed with 7% real returns and 4% withdrawals every year since you're skipping sequence of returns risk. Once you move away from "these actual returns that happened in this actual order" you're on a spectrum between all-actuals and monte carlo.
I think my preferred way to do this is, though I haven't thought too much, done the arithmetic, nor had any coffee today, is a hybrid approach:
1) if you have a full 30 years use those 30 years. e.g. 1900-1929. If this fails, it fails, if it passes, it passes. Add 1 or 0 to the numerator, and 1 to the denominator to calculate "% of times these inputs fail."
2) if you only have x (x < 30) years use those x years then pull every (30 - x) year segment and run the simulation out to 30 years and use results from that. Separate from your main "% of times these inputs fail" add up the percentage of simulations that pass or fail here. Add that number as a percentage of 1 to the global numerator, and 1 to the global denominator* to calculate the "% of times these inputs fail."**
An example:
You want to include the data from 1995-2019, but you're missing 5 years, so you can't know if the 1995 start year is a pass or a fail so you can't contribute a 1 or 0 to the numerator. Have a sub-calculation aimed at saying "if you have 1995 results for the first 25 years*** of retirement, what percentage of the time would you have failed in the last 5 years?****". Do that by modeling 1995-2019 then 1900-1904. Repeat using 1901-1905. etc. Looks like 97% of those 25 + 5 year models pass? Add 0.97 to the global numerator (so not 1 or 0, 0.97 to reflect uncertainty).
That makes the global success rate a little harder to interpret or reverse engineer, but maybe it bridges the gap between assumptions of independence of monte carlo simulations and the non-independence of the real world? In my head I guess I'm assuming you won't get a 5 year run of returns like 1929, 2008, 1929, 1929, 2008 and by bucketing into x year buckets we avoid that monte carlo result structurally while also providing a more realistic model than "6.97% real returns 5 years in a row" which is also very unlikely (although much less meaningfully unlikely than the other example in this paragraph).
* should this be weighted somehow? I don't know enough about this to answer the question of "how much should I care about this run given that x (or y or z) of these years didn't actually happen in this order".
** I haven't actually thought this through enough. I'm getting a nagging feeling that this is going to be arithmetically the same as 'use the average return rate for those years' making this all a waste of time =).
*** imo we can basically ignore this problem usually, because SORR tells us that you usually know in the first 10 years if you're going to fail or not. I think it is worth thinking this through though for people who want to model e.g. a large lump sum payment 15 years into retirement (kid goes to college, buying kid a house etc.). Even still, this is in danger of dipping into way too small to matter territory.
**** naturally of course this is kind of the same of truncating, just truncating at one level more recursive, so maybe there's not much value here. I.e. you're searching only for complete 5 year periods so you're not counting any of the years from 2015-2019. I guess that's better than ignoring 1995-2019.
In conclusion, this sounds like a lot of work for not much gain and might also have been covered in the ERN series above anyway =).
-
Would love to see real estate investment options added. A bunch of folks have between a few and a whole bunch of properties to fund retirement.
-
To add further ideas:
1) incorporating other features of popular calculators like graphing make a lot of sense as you get there =).
2) A down the road thing might be allowing users to customize the color coding - yellow for you might mean "money left at the end, but not much" and for me I might want to model it with yellow to mean "you've dropped below 75% of your inflation-adjusted stash" a.k.a. adjust the definition of failure. This gets away from the trinity study mindset but drifts towards real-world use of the calculator. Read through ERN to get a better sense for this (and for why things like "i'll just get a part time job if my stash drops below X%" aren't as great as they sound). It also blends the line between calculator and something more than that, e.g. the portfolio visualizer website.
3) This one depends on the core intention of the calculator. If this calculator exists to help people see how much they can withdraw each year, ignore this. If instead it exists to help them estimate how successful they'll be in retirement, I think we can do more.
[motivation]
In general these calculators are all built around entering your inputs and implicitly dealing with the 50%+ uncertainty inherit in modeling real life. That's why stressing while aiming for 100% success rates can be kind of silly - capital gains taxes might go up 10%, you might decide to add x$ in expense, maybe a war happens and everything changes etc. and all that calculation is for naught kind of thing. That said, they still have value because people have trouble envisioning long term stock returns and how much those matter for outcomes when withdrawing, and we can "server-side" model that helpfully because they are global/non-personal. However, none of these calculators really go any further on the spending side. Users come to the table with assumptions, but can we helpfully "server-side" model scenarios reflecting real life?
[/motivation]
All of this is non-vanilla of course and getting deep into too many features territory. But there might be simple things we could look at, such as asking people to input their previous 5 years of expenses and then model various scenarios based on that rather than just a simple mean - if I have 3 straight expensive years 5 years into retirement in 1929 that matters right? Can't we do better than "no one knows, just use the average"? In some ways that's the counterpart to the "7% returns means 7% SWR" argument on the savings side. Sequence-of-spending-risk (SOSR) matters (though it matters much less). "Use a buffer" is fine, but how much buffer? what if users don't think to include a buffer? Can we tell them "your mean withdrawal rate works in 87% of scenarios. If you get 30th percentile expenses results, it works in 32% of scenarios" and is there value in doing so?
On the "they can model that themselves with the inputs" side of things:
Certainly, but can we help? Maybe we could model real-life spending where x% of the years, distributed uniformly, someone has an unexpected one time x% increase in spending. Or maybe most people's spending goes up in their 40s and 50s, down from 70-80, and back up from 95-100. Of course, people can enter all of this themselves, but maybe a simple "real life population spending simulation" checkbox might be helpful to do this on the "server side" and protect ignorant users - no buffer, no black swan expenses etc. while making it easier for sophisticated users to model (instead of making them manually add 2 expensive years and run it for e.g. year 1 & year 2, then year 1 & year 3, etc.).
-
Love it!!! The only add-on thing I recommend is being able to adjust for taxes, so we can adjust for higher federal/state/city tax rates as appropriate (especially given all the deficit spending going on now).
-
Thanks for all of the thoughtful discussion, everyone!
I was actually thinking about this a week ago. I feel like analysis that I do is somewhat flaws because it is not including those years, some of which are most likely going to be failures in my plans (2000,2001,2002,2007). This is how all calculators work, not just yours, but it would be nice to have an idea on those. One idea that I was thinking was a box that would fill in historical data for returns and inflation for the years that we do not have data for. For example, on a 30 year time horizon and that boxed checked. The year 2000 scenario would use 2000-2020 actual data and then from 2021-2030 it would use the historical average (or even a user input number) for returns and inflation. Thoughts?
I've been thinking about this, too. I think it's OK to use averages. In fact – and you may already know this – William Bengen (who invented this style of retirement planning analysis afaik) did this in his original study for some of the retirement durations (he describes this in "Appendix: Assumptions of Computations of Portfolio Values" in "Determining Withdrawal Rates Using Historical Data" - William P. Bengen, 1994).
I agree with @swashbucklinstache that incorporating this steps the calculator in a direction toward becoming a Monte Carlo calculator. You suggested adding a full Monte Carlo option earlier, and I think that is a great idea, so I'm not sure if it makes sense to add this partial option. Does it provide much more value over a full Monte Carlo option? Would it make the UI overly complicated? I'm undecided right now, but I absolutely think it's worth considering further. Right now I'd say I'm leaning in the direction of having all 3 options:
- real data only
- real data + averages to fill in the gaps
- full Monte Carlo
A thing to consider though, is that you might want a way to tell the user / weigh results based on e.g. "95% of this run's results are using averages, not actuals" because as we all know that makes a huge difference.
Completely agree. If the results ever included average data, then that would need to be clearly stated.
I think my preferred way to do this is, though I haven't thought too much, done the arithmetic, nor had any coffee today, is a hybrid approach:
I really appreciate you taking the time to write out this algorithm in such detail! Thank you! I'm going to sit down in a bit and walk through it really make sure I fully understand it before giving a reply. I'll let you know if I have any questions.
Would love to see real estate investment options added. A bunch of folks have between a few and a whole bunch of properties to fund retirement.
Great idea, @Blindsquirrel ! I'll add this to the todo list. Thank you!
1) incorporating other features of popular calculators like graphing make a lot of sense as you get there =).
What kinds of graphs would you like to see? Other than the graphs that currently exist (portfolio over time and withdrawals over time), I was thinking of adding one to show the stock market performance over the period. i.e.; a chart showing how $1,000 that is 100% invested in the S&P 500 changes over the retirement, with no withdrawals/additions to it (kind of like Vanguard's performance charts for its funds).
At the moment I'm not considering adding in a graph that shows the portfolio values of the simulations on top of one another. I'm not really sure what information those graphs provide, but I'm open to hearing how others use that particular graph if anyone feels strongly that it should be a part of FI Calc!
2) A down the road thing might be allowing users to customize the color coding - yellow for you might mean "money left at the end, but not much" and for me I might want to model it with yellow to mean "you've dropped below 75% of your inflation-adjusted stash" a.k.a. adjust the definition of failure. This gets away from the trinity study mindset but drifts towards real-world use of the calculator. Read through ERN to get a better sense for this (and for why things like "i'll just get a part time job if my stash drops below X%" aren't as great as they sound). It also blends the line between calculator and something more than that, e.g. the portfolio visualizer website.
Yeah! I love this idea, and it's been the long-term vision for the color-coded years. It's definitely a down-the-road thing, but I'd love for users to be able to write their own little "algorithms" to color code years according to the things that they're interested in. Although I'm skeptical that I could ever cover all of the custom analysis folks do in spreadsheet apps, I think I might be able to get a good chunk of those use cases in the app? It's a cool feature to think about.
3) This one depends on the core intention of the calculator.
This is an interesting idea. It sounds kind of like Monte Carlo, but on the spending side rather than the market performance side? Is that right? I'll need to think more about it. If you have more thoughts on it over time ya gotta keep me updated! I'm definitely curious to hear more.
Love it!!! The only add-on thing I recommend is being able to adjust for taxes, so we can adjust for higher federal/state/city tax rates as appropriate (especially given all the deficit spending going on now).
Absolutely, this is a great idea @tk2356 . My main concern is getting the tax code for all 50 states + the federal government for as many years as possible. If I spend a long time reading the tax docs that they publish, I can usually figure them out, but doing it 51 times across 100+ years (or however many years are available)...that might take me ages, and I'm worried I would make some mistakes. I need to do more research into the feasibility of implementing this feature, but it is a really good idea. Maybe it's not even as complicated as I'm thinking. Either way, I'll add it to the list!
Other things I'm working on:
- Alternative withdrawal schedules. Right now, it is set to 1 year, but I am working on rewriting the algorithm to support monthly, quarterly, and twice-yearly.
- Customizing the CSV export columns. i.e.; you can add/remove columns based on what you find useful. These settings would be saved, so once you configure it once you don't have to change them again (unless you want to). In the distant future you might even be able to write calculations for custom columns (using a similar system to the custom color-coded years described above).
- an optional summary CSV export. Rather than year-by-year results, you would just see 1 row for each year.
I'm also likely to add a "Guides" section that goes into extreme detail how all of the algorithms in the calculator work.
One example of a guide could describe what happens in a simulation year. Bengen's analysis worked like this:
1. Year starts with a given portfolio value
2. Adjust portfolio for gains/losses
3. Withdrawal occurs based on initial portfolio value
4. Portfolio is rebalanced
FI Calc swaps 2 and 3, but there's no way for anyone to know that unless they're reading this thread. So I want to write that in a guide somewhere. Also, I think it might be interesting in an advanced setting to allow users to customize this order of things.
-
I was thinking more along the lines of an "average tax rate" like on this calculator: https://engaging-data.com/will-money-last-retire-early/
Then, the income listed will be pre-tax, and the individuals can set the tax rate at what's appropriate to them. This allows someone to see how far one's stache goes (tax-wise) in NYC versus, say, Texas or Florida.
I'm not sure how difficult that is to include though!
-
I tried out the calculator and just want to confirm, the extra income section that is extra yearly income not monthly correct?
Thank you for this.
-
I was thinking more along the lines of an "average tax rate" like on this calculator: https://engaging-data.com/will-money-last-retire-early/
Then, the income listed will be pre-tax, and the individuals can set the tax rate at what's appropriate to them. This allows someone to see how far one's stache goes (tax-wise) in NYC versus, say, Texas or Florida.
I'm not sure how difficult that is to include though!
Oh nice, that makes sense. Totally doable. Added to the todo list!
I tried out the calculator and just want to confirm, the extra income section that is extra yearly income not monthly correct?
Thank you for this.
That's correct, everything in the calculator is yearly. I'll update the extra income/withdrawal sections to make this more clear.
Thanks @BikeFanatic !
Edit: I updated the "additional income" help dialog to explicitly state that these are yearly sums. I also updated the add/edit income dialog to label the input as "Annual Amount" rather than "Amount". I hope that helps clarify things in the app, @BikeFanatic ! Let me know what you think.
-
I just messed around with this thing for the better part of an hour.
Things I wist I had:
- Reoccurring withdraws (ie. new car every 10 years)
- Visual of which year is the "best" year and more importantly "worst" year
- Ability to set a floor on the portfolio value which causes a "Fail" or "near failure" - if I start with 1,000,000 and get below 500,000 I am having a panic attack and getting a job.
- Wasn't clear how the income / withdraws were changing the results, had to put in very large values to 'see' the difference. Maybe a less dedicated person wouldn't take the time.
- ability to adjust portfolio
Also noted that if I had a "max/min" set from one withdraw style and switched to another style it would keep those max/min unless I manually went back and unchecked it. Probably should automatically un-check when the combo box value changes.
Looks good - congrats
-
I've just given it a try, looks very slick, thanks for sharing it.
Is there an issue with the Historical Data? When I use the tick box "Use all available historical data" I get data from 1871 to 1975. If I set "Only use data between" 1900-2000 I only get data from 1900 to 1955 displayed? I had Length of Retirement set to 45 years, changing this variable changes how much data is displayed... Have I missed something or is it broken?
Hey @Shaz_Au ! Thanks for trying it out.
The behavior that you're seeing is expected. Let me explain it a bit more:
The calculator can only run simulations when there is enough historical data to cover the entire retirement length. What this means is that for a retirement that's 45 years long, the latest start year that you will see is 1975 (2020 - 45 years = 1975) when you're using all of the available data. Similarly, if you restrict the years to be 1900-2000, then the latest start year that you'll see will be 2000 - 45 = 1955.
If you click into one of the years that you are seeing (which, again, represents the start year of a particular retirement simulation), you will see more details, and it will show how the portfolio evolves over the duration of the specified retirement. So clicking into, say, the 1975 year from your first calculation will show a retirement spanning 1975 - 2020.
One other thing that might help you understand what is going on is to set the length of retirement to be 1 years. Then you'll see start dates all of the way up until 2019. But, again, 2020 won't be displayed because if that one-year retirement started in January 2019 then it ended in January 2020.
Does that help clarify things?
Yes, thank you for your reply, a typical "ID Ten T" (IDIOT) error! Exactly as you thought I didn't recognise the year represented the start year.
-
I just messed around with this thing for the better part of an hour.
Things I wist I had:
This is awesome! Thank for you the detailed list!
- Reoccurring withdraws (ie. new car every 10 years)
Makes sense. I also thought about adding this. I'm still open to it; I just want to be mindful that the add/edit dialog for these things doesn't get too complicated/confusing. I'll keep thinking about it and let you know if I come up with a way to add it in.
- Visual of which year is the "best" year and more importantly "worst" year
I definitely want to add more analytics to the results page. This is a cool idea, I'll jot it down. What do you think might factor into what determined the best/worst year?
One idea is end portfolio value. Another is average spending over the retirement. Others might be whether the portfolio dipped or not. Maybe a custom algorithm that factors all of these in, and maybe more, to get some kind of "score"? I'm just spitballing here – did you have something specific in mind with this suggestion?
- Ability to set a floor on the portfolio value which causes a "Fail" or "near failure" - if I start with 1,000,000 and get below 500,000 I am having a panic attack and getting a job.
Agreed. This will be added (although it might take some time)
- Wasn't clear how the income / withdraws were changing the results, had to put in very large values to 'see' the difference. Maybe a less dedicated person wouldn't take the time.
I also noticed that lot of the withdrawal strategies are remarkably resilient to a couple of big purchases here and there, so yeah, I totally hear you on this one. Maybe there's an indicator I can add to the year pills that represents additional spending or income for that year, so that it's more easy to dig into the years when those withdrawals/income occur. (I just realized this suggestion makes no sense lol)
- ability to adjust portfolio
Absolutely. I'm working on this now, and support for bonds is nearly done.
Also noted that if I had a "max/min" set from one withdraw style and switched to another style it would keep those max/min unless I manually went back and unchecked it. Probably should automatically un-check when the combo box value changes.
Agreed, I'll update this.
Looks good - congrats
Thank you!
Yes, thank you for your reply, a typical "ID Ten T" (IDIOT) error! Exactly as you thought I didn't recognise the year represented the start year.
Ha, nah, you're not an idiot. All of these calculators have a steep learning curve, because what they're doing is inherently complicated. This just indicates to me that I have more work to do on my end to make the information more clear.
I'm glad we got it sorted out, though ✌️
-
I definitely want to add more analytics to the results page. This is a cool idea, I'll jot it down. What do you think might factor into what determined the best/worst year?
One idea is end portfolio value. Another is average spending over the retirement. Others might be whether the portfolio dipped or not. Maybe a custom algorithm that factors all of these in, and maybe more, to get some kind of "score"? I'm just spitballing here – did you have something specific in mind with this suggestion?
My initial thought was to have a gradient color scale on the year boxes that would visually express best to worst instead of a static green/yellow/red. That way if they all pass you still have an idea of what the trouble years were, and if you have multiple red you know which is the worst year.
Probably a multi-factor approach is most appropriate as you suggested. Me personally, I wasn't concerned with the ending portfolio amount - I was looking at how many consecutive years I was hitting my minimum withdraw amount, and what the minimum portfolio value was as % of initial value. My thought being that if I have a year or two when I live on a tight budget then that isn't too bad - but if it is 10 years in a row that is a structural problem with my plan. 10 years spaced out is way better than 10 years in a row. - I could tell my wife, "we can't go visit your parents this year" ... but I can't tell her, "we can't go visit your parents this decade!"
ETA - Your average casual user probably only cares about ending portfolio amount.
I like the average/median withdraw stats too - that would be helpful when comparing different withdraw strategies.
-
My initial thought was to have a gradient color scale on the year boxes that would visually express best to worst instead of a static green/yellow/red. That way if they all pass you still have an idea of what the trouble years were, and if you have multiple red you know which is the worst year.
Probably a multi-factor approach is most appropriate as you suggested. Me personally, I wasn't concerned with the ending portfolio amount - I was looking at how many consecutive years I was hitting my minimum withdraw amount, and what the minimum portfolio value was as % of initial value. My thought being that if I have a year or two when I live on a tight budget then that isn't too bad - but if it is 10 years in a row that is a structural problem with my plan. 10 years spaced out is way better than 10 years in a row. - I could tell my wife, "we can't go visit your parents this year" ... but I can't tell her, "we can't go visit your parents this decade!"
ETA - Your average casual user probably only cares about ending portfolio amount.
I like the average/median withdraw stats too - that would be helpful when comparing different withdraw strategies.
This is awesome, thanks for sharing it. The current analysis is definitely leaning toward ending portfolio values, primarily because the Bengen/Trinity study do, as does cFIREsim, but a goal of mine is to allow folks to be able to customize the app to focus on the things that they care about. What you described here seems totally reasonable, so I'm going to note is as a use case that I want the app to support.
Realistically, it'll probably be awhile before I can add this kind of thing as I still have some fundamentals to add, like more portfolio options.
Thanks for writing that up for me!
-
PTF
As a color confused person, a more distinctive difference between the not-so-good and the awful start years would be nice.
Maybe a brighter not-so-good notation. This might also help in mobile usage in brighter ambient light.
-
PTF
As a color confused person, a more distinctive difference between the not-so-good and the awful start years would be nice.
Maybe a brighter not-so-good notation. This might also help in mobile usage in brighter ambient light.
Great feedback. Question: do you have a particular type of color blindness, or do you see all colors and it's just that you find it difficult to distinguish between those two colors in the app right now?
-
An idea that might be of interest but I have noticed that many calculators don't take into account the notion of Means Tested Income/Assets. Basically if the stash is reduced below a certain level the Retiree might have the right to a certain amount of income from the government.
I know this is the case in Australia where an old age pension is provided after age 66 but it decreases the more your income increases. I suppose it might be a bit complex to model though. So maybe something like:
If Assets less than X, additional income increases by Y.
But also
If withdrawal from assets less than X, boost Additional Income by fraction Y of reduced income below the threshold (X) up to maximum Z (where income does not increase further).
-
PTF
As a color confused person, a more distinctive difference between the not-so-good and the awful start years would be nice.
Maybe a brighter not-so-good notation. This might also help in mobile usage in brighter ambient light.
Great feedback. Question: do you have a particular type of color blindness, or do you see all colors and it's just that you find it difficult to distinguish between those two colors in the app right now?
My formal diagnosis is red-green. I can see, identify high saturation colors including fire engine red or Kelly green. Pastels are a big source of confusion, although I can see the color (not just grey).
Mom found out when I was in kindergarten, and the teacher and I had a disagreement on the color of a man's overall I was drawing. I said blue, she said purple.
I still have my last school art from 9th grade. I explained my issues to that teacher, who seemed ok with it.
Since you are using a black background, just turning up the RGB or hex values on the "yellow" would be good enough for me. YMMV
I usually only get to the first page of the Ishihara test. https://www.colour-blindness.com/colour-blindness-tests/ishihara-colour-test-plates/
On well kept, newer paper copies of the test, sometimes I can guess correctly on other pages.
-
My formal diagnosis is red-green. I can see, identify high saturation colors including fire engine red or Kelly green. Pastels are a big source of confusion, although I can see the color (not just grey).
Mom found out when I was in kindergarten, and the teacher and I had a disagreement on the color of a man's overall I was drawing. I said blue, she said purple.
I still have my last school art from 9th grade. I explained my issues to that teacher, who seemed ok with it.
Since you are using a black background, just turning up the RGB or hex values on the "yellow" would be good enough for me. YMMV
I usually only get to the first page of the Ishihara test. https://www.colour-blindness.com/colour-blindness-tests/ishihara-colour-test-plates/
On well kept, newer paper copies of the test, sometimes I can guess correctly on other pages.
Whoa, this is incredibly useful. Thank you so much for writing all of this up for me!
I'd love to get your thoughts on this idea I have: I'm not sure if one set of colors will provide the best experience for everybody, and I want this app to be the best experience for everybody. So rather than pick one set up of colors for everyone, what I can do is create different color schemes, so there will be an option that looks the best for, say, red-green. Once you set that once, the setting will be stored in your browser so you'll never need to set it again.
What do you think of that? If you agree that it could work, I can add it the app and then I'd love to get your feedback on the color choices.
Thanks again for sharing such useful feedback, @markbike528CBX !
An idea that might be of interest but I have noticed that many calculators don't take into account the notion of Means Tested Income/Assets. Basically if the stash is reduced below a certain level the Retiree might have the right to a certain amount of income from the government.
I know this is the case in Australia where an old age pension is provided after age 66 but it decreases the more your income increases. I suppose it might be a bit complex to model though.
This seems doable. One thing to note is I am focusing on the US use-case right now (since the stock market data that I have is US-based, I am in the US, and I also have limited time to add features 😬), but I'm going to add a note to myself to revisit this idea. Thanks for sharing!
-
Beautiful job! Congratulations and thanks for making it available to us!
-
My formal diagnosis is red-green. I can see, identify high saturation colors including fire engine red or Kelly green. Pastels are a big source of confusion, although I can see the color (not just grey).
Mom found out when I was in kindergarten, and the teacher and I had a disagreement on the color of a man's overall I was drawing. I said blue, she said purple.
I still have my last school art from 9th grade. I explained my issues to that teacher, who seemed ok with it.
Since you are using a black background, just turning up the RGB or hex values on the "yellow" would be good enough for me. YMMV
I usually only get to the first page of the Ishihara test. https://www.colour-blindness.com/colour-blindness-tests/ishihara-colour-test-plates/
On well kept, newer paper copies of the test, sometimes I can guess correctly on other pages.
Whoa, this is incredibly useful. Thank you so much for writing all of this up for me!
I'd love to get your thoughts on this idea I have: I'm not sure if one set of colors will provide the best experience for everybody, and I want this app to be the best experience for everybody. So rather than pick one set up of colors for everyone, what I can do is create different color schemes, so there will be an option that looks the best for, say, red-green. Once you set that once, the setting will be stored in your browser so you'll never need to set it again.
What do you think of that? If you agree that it could work, I can add it the app and then I'd love to get your feedback on the color choices.
Thanks again for sharing such useful feedback, @markbike528CBX !
If the colors are right next to each other, it is clearer. The problem is if the are few and scattered various colors (success can be confusing?).
I tried grey scale and then the very dark and the OK (black) blend too much.
If you make the lightest color too light then you blend with the button color.
I checked out the colors
ca1010 (failed) and B36300 (warning/near fail) at https://htmlcolorcodes.com/color-picker/ and they have similar attributes (RGB, CYMK values)
mmmm...... not sure what I can suggest....
Not so relevant to the current discussion but a good thinking book, and a Great coffee table book for geeks is:
The Visual Display of Quantitative Information
https://www.edwardtufte.com/tufte/books_vdqi
-
Lots of updates since my last post here:
- the results page now show some statistics on your portfolio and withdrawals (more to come!)
- bonds are supported now
- you can choose to rebalance your portfolio, as well as specify the frequency of rebalancing
- you can specify a glide path. Right now, only a linear glide path is supported, but I'm working to add more options
- many more withdrawal strategies
- lots of guides (https://ficalc.app/introduction/) have been written that explain how the calculator works. More are on the way
Would love to see real estate investment options added. A bunch of folks have between a few and a whole bunch of properties to fund retirement.
@Blindsquirrel , if you have time, I'm curious to hear how you're imagining real estate properties to work in this calculator. One way is that you could specify how much you expect to earn in cash flow from the properties. If this is all that you need, then the Additional Income feature may support it? Or do you think other features specific to real estate could be useful?
Also, if anyone else has real estate and would like to see it added to this calculator, I'm curious to hear what you like to see as well!
-
Just wanted to say a HUGE thank you, Jamesplease. Your app is AMAZING! Thank you so much for sharing it with us.
-
I wouldn't normally post for a single feature, but I've been working toward today's update for weeks now, so I'm really excited about it and want to share it!
You can now configure your glide path. Choose between "evenly", "quickly," or "slowly" to determine the rate at which your assets move from their start allocation values to their end allocation values. The application shows you the different rates in a chart. I've attached an image showing a portfolio that's configuring to rebalance from 80/20 to 50/50 "slowly".
Just wanted to say a HUGE thank you, Jamesplease. Your app is AMAZING! Thank you so much for sharing it with us.
Thank you @Goatee Joe ! I appreciate your kind words!
-
PTF
-
PTF
Hi Nereo! I hope you’re doing well. Is this an acronym? If so I don’t know what it stands for!
-
The site has been looking great! I am loving the new changes and the glidepath option! Looking forward to trying out all of the new withdrawal options!
By the way PTF means posting to follow!
-
Two more updates. One small, one big:
Small: Hebeler Autopilot II has been introduced. This is distinct from the Autopilot used in cFIREsim (which is Autopilot I).
Big: you can now share/save simulations. I know a lot of people have wanted this, so I'm glad to finally have it released!
The site has been looking great! I am loving the new changes and the glidepath option! Looking forward to trying out all of the new withdrawal options!
Yooo! @themagicman ! Good to hear from you again :) Thanks for the kind words. I'm curious to know: what feature could I add next that would be most impactful for you? Also, if anyone else wants to chime in feel free. I've got a long to-do list but I'm more motivated when I know someone will use a feature.
By the way PTF means posting to follow!
Oooh, good to know. Thanks for clarifying that for me :)
-
Simply wonderful work! Easy to read and input. I especially like the ability to add secondary deposit and withdrawals for set periods of time to account for social security or loans. I can't think of any real suggestions at this point that aren't already on your list. Thank you for building this and sharing!
-
Two more updates. One small, one big:
Small: Hebeler Autopilot II has been introduced. This is distinct from the Autopilot used in cFIREsim (which is Autopilot I).
Big: you can now share/save simulations. I know a lot of people have wanted this, so I'm glad to finally have it released!
The site has been looking great! I am loving the new changes and the glidepath option! Looking forward to trying out all of the new withdrawal options!
Yooo! @themagicman ! Good to hear from you again :) Thanks for the kind words. I'm curious to know: what feature could I add next that would be most impactful for you? Also, if anyone else wants to chime in feel free. I've got a long to-do list but I'm more motivated when I know someone will use a feature.
Yeah I was MIA for a while (work got too busy). Hopefully should be getting back normal now.
The most impactful feature for me right now would definitely be a monte carlo simulation aspect to the calculator! I would find that extremely valuable!
-
Simply wonderful work! Easy to read and input. I especially like the ability to add secondary deposit and withdrawals for set periods of time to account for social security or loans. I can't think of any real suggestions at this point that aren't already on your list. Thank you for building this and sharing!
Thanks for the kind words, @Raeon ! I'm glad it's working well for you! If you come up with any features that you'd like to see, or if you find something that could be improved, don't hesitate to send me a message.
Yeah I was MIA for a while (work got too busy). Hopefully should be getting back normal now.
Cool, cool, I'm glad things are chilling out for you over there :)
The most impactful feature for me right now would definitely be a monte carlo simulation aspect to the calculator! I would find that extremely valuable!
Good to know. Even though it's not usable yet, I've been working toward supporting this under-the-hood, and I'm already most of the way toward having it done. I'll put finishing it up on my shortlist of todo's.
-
First of all, ***thank you*** for making downloadable *.csv files that are import-ready for programs such as R and Matlab. As a data-wrangler that's been my biggest beef with other retirement calculators (it's takes a lot of data-massaging to do any further analysis).
A few things I would like to see
- Highlight/focus on years that resulted in failures in the "Results" section. That's really what people care most about, and it can be enlightening to see under what conditions (e.g. market returns, inflation, SORR) these failures occurred
- In the 'Results' section generate a single graph with the median return as a thick, red line and the CI from that return as a gray area, and the SD within as a blue area. That will quickly show how fast a portfolio will grow/shrink with 'average' conditions, as well as the spread throughout the period.
Just my 2¢.
BTW, as others have said "PTF" means "Posting to Follow". Basically I did that because I wanted to come back and spend more time looking at your calculator and giving constructive feedback, but I didn't have the time to do a proper job when I first came across your thread.
-
Thanks for doing this!
Chuckling that I'm posting right after a discussion of PTF. I haven't had time to read all the details yet, am indeed posting so I can easily catch up later. :)
Fwiw, skimming, I was going to ask for downloadable results. It looks like you're doing that already; thumbs up.
If you want a light income stream to cover possible expenses, maybe you could set up a Patreon account for it?
-
First of all, ***thank you*** for making downloadable *.csv files that are import-ready for programs such as R and Matlab. As a data-wrangler that's been my biggest beef with other retirement calculators (it's takes a lot of data-massaging to do any further analysis).
Oh, nice! Although I admit I didn't intentionally test supporting those programs, I'm glad that I didn't unintentionally break the CSV files for them. This is definitely a use case I want to support, though. Do let me know if there's anything I can do to make your life easier when it comes to exporting the data into these other apps that you use.
A few things I would like to see
- Highlight/focus on years that resulted in failures in the "Results" section. That's really what people care most about, and it can be enlightening to see under what conditions (e.g. market returns, inflation, SORR) these failures occurred
- In the 'Results' section generate a single graph with the median return as a thick, red line and the CI from that return as a gray area, and the SD within as a blue area. That will quickly show how fast a portfolio will grow/shrink with 'average' conditions, as well as the spread throughout the period.
Just my 2¢.
I love these suggestions, and both are things that I want to add. In fact, an earlier version of the app did indicate which years failed (you can see screenshots earlier on in this thread). I temporarily removed because some withdrawal strategies, such as VPW, intentionally bring you to $0 in the final year, so for those strategies it is a good thing when your portfolio hits $0 (because they are operating as intended). I am working on a way to differentiate how results are displayed based on the goals of each individual strategy, but there's still a way to go. Added it back for all strategies!
I also agree that there needs to be an overview graph. It's on the todo list.
BTW, as others have said "PTF" means "Posting to Follow". Basically I did that because I wanted to come back and spend more time looking at your calculator and giving constructive feedback, but I didn't have the time to do a proper job when I first came across your thread.
Good to know! Thank you :D
Thanks for doing this!
Chuckling that I'm posting right after a discussion of PTF. I haven't had time to read all the details yet, am indeed posting so I can easily catch up later. :)
Fwiw, skimming, I was going to ask for downloadable results. It looks like you're doing that already; thumbs up.
If you want a light income stream to cover possible expenses, maybe you could set up a Patreon account for it?
Yo @BicycleB ! Thanks for stopping by. A Patreon is an interesting idea that I hadn't considered. I'll definitely keep it on the table as I think about the different options.
Atm I'm thinking I'll just setting aside enough of my personal money to pay the cost for a couple of years, just to give me time to build the best app possible w.o. thinking about the financial aspect.
Also, @themagicman , Monte Carlo is in beta. There is still a lot of work to do, and it's a secret option in the calculator atm, but you can see a verrrrrrry early version at this link:
https://calculator.ficalc.app/?calculationType=monteCarlo
There is obviously a ton of work that I need to do (way, way more simulation cycles, verifying the results, etc.), but the feature is well underway :)
-
Great this is going to be very beneficial! When I went to look at it, it looks like the the success percentage is gone (and the yearly) for all but the constant withdraw option. I was originally thinking this was because it is still being worked on but then I noticed that the success percentage is going away even in the historical results calculator.
-
Great this is going to be very beneficial!
:D
When I went to look at it, it looks like the the success percentage is gone (and the yearly) for all but the constant withdraw option. I was originally thinking this was because it is still being worked on but then I noticed that the success percentage is going away even in the historical results calculator.
It's been this way for awhile, and you are right actually that it's something that's still being worked on. This just the current state of where the Results section of the calculator is at right now.
I've been pretty focused on the "Configuration" side of the calculator the past few weeks, and the "Results" section hasn't been updated much at all. The Configuration section is maybe 75% where I want it to be, but the Results section is maybe 20% where I want it to be.
I'm realizing now to finish Monte Carlo I'll need to put some more thought into where I'm headed with the Results, so I'll be making some more updates to Results over the next few days/weeks as part of preparing for Monte Carlo. One of the planned updates will be making the results more consistent for all of the different strategies.
-
I like that you've added a color key to the years that failed. Nice and effective visual.
It might be even better if you expanded it to be a true 'heat-map', with Red (years that failed), Orange (years that ended with < 35% starting assets) and then a couple shades of green for portfolios that ended 100-200% of starting assets, 200-400% above starting assets and >400% above starting assets.
...or something like that. It's great to see the failures and how they tend to cluster togerther. It's also nice to see years that were runaway successes where you wound up with far more money than you ever needed.
-
I like that you've added a color key to the years that failed. Nice and effective visual.
It might be even better if you expanded it to be a true 'heat-map', with Red (years that failed), Orange (years that ended with < 35% starting assets) and then a couple shades of green for portfolios that ended 100-200% of starting assets, 200-400% above starting assets and >400% above starting assets.
...or something like that. It's great to see the failures and how they tend to cluster togerther. It's also nice to see years that were runaway successes where you wound up with far more money than you ever needed.
My long-term goal is to allow a degree of customization with the squares and the analysis, so that each individual can specify the kinds of things that they are interested in. I think adding a green square by default might make sense, though. I'll try it out and see how it feels.
Today's update: charts on the Results overview page. They're not perfect, but I'm curious to get any feedback that folks have on the charts. Do these convey all of the information you'd like to see from a visual representation of the data? Is there something else you're interested in seeing that the charts don't show?
-
I would appreciate it if a year-by-year withdrawal graph was presented so that one could verify that additional withdrawals are being accounted for (for example two lines showing baseline per the withdrawal strategy and the actual, modified withdrawal, above that). Looking at the CSV, what I see is an interesting pattern where the "Additional Withdrawal Amount" drops to zero for long stints even when one configures additional expenses.
For example, I added two expenses representing potential excess medical expenses of 30k a year, one starting at +20Y and one starting at +30Y, each set to repeat for 100Y (that is, forever). In the "Additional Withdrawal Amount" this seems not to be doing what it should be.
-
I would appreciate it if a year-by-year withdrawal graph was presented so that one could verify that additional withdrawals are being accounted for (for example two lines showing baseline per the withdrawal strategy and the actual, modified withdrawal, above that). Looking at the CSV, what I see is an interesting pattern where the "Additional Withdrawal Amount" drops to zero for long stints even when one configures additional expenses.
For example, I added two expenses representing potential excess medical expenses of 30k a year, one starting at +20Y and one starting at +30Y, each set to repeat for 100Y (that is, forever). In the "Additional Withdrawal Amount" this seems not to be doing what it should be.
Withdrawals can drop to 0 if the portfolio runs out of money, which may explain what you’re seeing.
Can you share a link to this calculation? There is a “Share Calculation” button on the overview page that will grab a URL you can share.
It's also nice to see years that were runaway successes where you wound up with far more money than you ever needed.
Agreed. I added a single color that represents > 300%. I strayed away from green, as I didn't want folks to think that these years are necessarily a good thing.
In the future, I'd like for folks to be able to customize this number as well as add additional ones.
-
I like that you've added a color key to the years that failed. Nice and effective visual.
It might be even better if you expanded it to be a true 'heat-map', with Red (years that failed), Orange (years that ended with < 35% starting assets) and then a couple shades of green for portfolios that ended 100-200% of starting assets, 200-400% above starting assets and >400% above starting assets.
...or something like that. It's great to see the failures and how they tend to cluster togerther. It's also nice to see years that were runaway successes where you wound up with far more money than you ever needed.
My long-term goal is to allow a degree of customization with the squares and the analysis, so that each individual can specify the kinds of things that they are interested in. I think adding a green square by default might make sense, though. I'll try it out and see how it feels.
Today's update: charts on the Results overview page. They're not perfect, but I'm curious to get any feedback that folks have on the charts. Do these convey all of the information you'd like to see from a visual representation of the data? Is there something else you're interested in seeing that the charts don't show?
Charts look great! I think for me the two big things you have covered with ending value and withdraw amount. I am not really sure on if there is any other chart I would want to see but I will be thinking on that. Two things of feedback on the chart (the application itself not the data)
1. It is not clear to me how I can can view sections on the chart. For example, if I want the chart to show me range and the percentage of 1 standard deviation (I am assuming those lines are standard deviation) from the mean and 2 staandard deviations from the mean, it is difficult to get the system to do this. It has done it for me before but I cannot figure it out. It would be nice to click in there and it select that. When I hover over it, it is showing me the whole 90% range. Does that make sense? If not, I can reword with screen shots
2. It would be nice to be able to change the calculating amounts on the charts. For example, I want to see ending values between X and Y and know what % that is.
-
Also on the Monte Carlo sim, does it take the data averages based on the history and use that for the sim?
I really hope that the monte carlo sim is in the early enough stages that it is just not calculating right and not the fact that my plan is not going to hold up. Getting a 99% chance of success of my plan using historical data and only 35% on monte carlo data
-
Also, I can't remember if I had saw this before or not but love the start year,end year, current year columns on the CSV output!
-
This calculator is great. Thank you.
-
Another feature that I would like is the ability to decrease my withdrawal rate once social security kicks in. I could not figure out how to calculate that in. Otherwise this is a great calculator.
-
1. It is not clear to me how I can can view sections on the chart. For example, if I want the chart to show me range and the percentage of 1 standard deviation (I am assuming those lines are standard deviation) from the mean and 2 staandard deviations from the mean, it is difficult to get the system to do this. It has done it for me before but I cannot figure it out. It would be nice to click in there and it select that. When I hover over it, it is showing me the whole 90% range. Does that make sense? If not, I can reword with screen shots
The chart shows you the section that you're hovered over. So if you move your mouse (or finger) horizontally in the chart, you should see details of whatever specific item you're hovered on. It's the same interaction model used for all of the charts in the app. Let me know if that helps!
2. It would be nice to be able to change the calculating amounts on the charts. For example, I want to see ending values between X and Y and know what % that is.
Interesting idea. I've taken and a note and I'll think about how I can add something like this.
Also on the Monte Carlo sim, does it take the data averages based on the history and use that for the sim?
I really hope that the monte carlo sim is in the early enough stages that it is just not calculating right and not the fact that my plan is not going to hold up. Getting a 99% chance of success of my plan using historical data and only 35% on monte carlo data
I wouldn't recommend using Monte Carlo to test anything right now since there is a lot of work to do. For example, it only runs 0.2% of the target number of calculations that I want it to run, so the success rate metrics etc. will all be inaccurate for that reason. Also, I have no tests to verify that it's working as expected. I just sent over the link as a little demo, but it's very much a work in progress feature.
With that said, I'd say I'm 95% certain that the results that you're seeing are correct. Should that be cause for worry? Maybe, but probably not.
Here's what's going on: FI Calc uses data from 1871 to 2020 by default. This is OK for the backtesting calculations because it always goes chronologically like so: 1871...1872... ...2019...2020. Throughout this process, the market steadily grows, but never too much from year to year.
However, the total difference in stock price between years close to 1871 and years close to 2020 is enormous. Because the Monte Carlo calculation randomly samples data from the full data set, it's regularly flipping back and forth between these two extremes (as random chance would have it). Using the full data set for Monte Carlo (which again, is the default right now) is effectively like testing how your portfolio performs if the worst stock market crash that is mathematically possible occurred every couple of years. Very few portfolios will survive such havoc. The fact that you're getting 35% success rate could even be a cause for celebration.
Most retirement calculators restrict Monte Carlo sims to use data from about 1975 to 2020, which yields considerably better results because the year-to-year changes are relatively small. If you open up the "Historical Data" section of the Configuration and restrict your calculations to those years you should see an improvement in your success rates. But again, take the results with a grain of salt. There are only 500 cycles run with Monte Carlo right now, and it needs to be 100k - 250k.
When I get around to working on Monte Carlo again, the default historical range will likely be shrunken so that the results aren't so grim.
Also, I can't remember if I had saw this before or not but love the start year,end year, current year columns on the CSV output!
Thank you! I actually added some of those columns based on your feedback a couple of months back :)
This calculator is great. Thank you.
🙏 I'm glad you find it useful!
Another feature that I would like is the ability to decrease my withdrawal rate once social security kicks in. I could not figure out how to calculate that in. Otherwise this is a great calculator.
The calculator actually behaves this way, although the charts are misleading right now. It's on my todo list to make this behavior more clear in the charts. You can read more about the behavior of Additional Income in the guides (https://ficalc.app/configuration/additional-income/). Check out the section "Effect on Withdrawals".
Ideally, the guides should be able to answer any and all questions related to how the calculations work. If you have a question, I'd recommend checking out the guides to see if it's answered in there. If you can't find the answer in the guides, let me know and I'll add it!
-
1. It is not clear to me how I can can view sections on the chart. For example, if I want the chart to show me range and the percentage of 1 standard deviation (I am assuming those lines are standard deviation) from the mean and 2 staandard deviations from the mean, it is difficult to get the system to do this. It has done it for me before but I cannot figure it out. It would be nice to click in there and it select that. When I hover over it, it is showing me the whole 90% range. Does that make sense? If not, I can reword with screen shots
The chart shows you the section that you're hovered over. So if you move your mouse (or finger) horizontally in the chart, you should see details of whatever specific item you're hovered on. It's the same interaction model used for all of the charts in the app. Let me know if that helps!
See the attached. I am trying to find the numbers that are in between the red x and the percentage of withdraw that fall in that range. Does that make sense and is that possible?
Also changing the monte carlo do 1975 on changed it significantly. So is how the calculator is working it take a random year in the range and then chooses another random year in the range and so on? Is that why it was a low chance before? Because in reality poor years are generally followed by good years but in the monte carlo it is not? Does it take into account the mean returns or the standard deviation of the history returns or does it just choose a random actual year? Also I guess 1975 on is not as volatile and that is why it gives a higher percentage of success then the whole thing?
-
@jamesplease thanks for putting this together. Just played around with it. Easy to use and had all the various variables I could think of factoring in.
-
See the attached. I am trying to find the numbers that are in between the red x and the percentage of withdraw that fall in that range. Does that make sense and is that possible?
Ah okay, that makes sense! Thanks for taking the time to share that screenshot. I definitely misunderstood you before, but I think I get what you’re asking now.
Unfortunately that’s information is not currently possible from the graphs. I plan to add a tabular version of the data from the graphs, which would then allow you to see the all of the different quantile values numerically.
Also changing the monte carlo do 1975 on changed it significantly. So is how the calculator is working it take a random year in the range and then chooses another random year in the range and so on? Is that why it was a low chance before?
Yup, that’s right.
Because in reality poor years are generally followed by good years but in the monte carlo it is not?
Note: this explanation may actually be inaccurate. Leaving it here for posterity but the subsequent posts clarify it.
The difference is that Monte Carlo with the full historical time range can – and repeatedly does – produce a crash far, far worse than any crash in history.
Does it take into account the mean returns or the standard deviation of the history returns or does it just choose a random actual year?
It picks a random actual year. No means, standard deviations, or any other statistical variations are factored in. There are different kinds of Monte Carlo algorithms, and this is the one used by all of the popular Monte Carlo retirement calculators that I looked at, which is why I went with it.
Also I guess 1975 on is not as volatile and that is why it gives a higher percentage of success then the whole thing?
Update: it turns out this explanation I gave is probably wrong (see the following posts). I’m leaving it up here for posterity’s sake but it’s likely inaccurate!
The reason the success rate is higher is that the difference between the stock market price from 1975-2020 is a lot smaller than the difference between the stock market price from 1871-2020, so simulated crashes are less extreme, and, perhaps, more realistic.
To put some numbers on it:
Shiller’s data has the S&P 500 valued at 4.44 in 1871. In 2020, it is 2930. So using the full data set can simulate situations where your portfolio drops to 0.1% of its value from one year to the next. Will any withdrawal strategy survive changes that are this extreme? No, probably not. Is it realistic? Again, probably not (in my opinion).
The value of the S&P 500 is around the $80ish range, which is ~3% of the value in 2020. This can still lead to some pretty serious drops, but comparatively it’s less extreme, and probably more realistic, than using the full data set. There are also less likely to be fewer drops since the data is less spread out, so the standard deviation is smaller.
Let me know if that helps clarify what’s going on any better!
@jamesplease thanks for putting this together. Just played around with it. Easy to use and had all the various variables I could think of factoring in.
Hey, thanks! I’m glad you found it easy to use :D
-
Shiller’s data has the S&P 500 valued at 4.44 in 1871. In 2020, it is 2930. So using the full data set can simulate situations where your portfolio drops to 0.1% of its value from one year to the next. Will any withdrawal strategy survive changes that are this extreme? No, probably not. Is it realistic? Again, probably not (in my opinion).
If I understand your description accurately, I would be very surprised if your Monte Carlo implementation is similar to others out there.
Let's take your example numbers from above, and further assume the market went up 5% in 1871 and 10% in 2020.
If your Monte Carlo simulation were to randomly pick 2020 and 1871 as the first and second years of the simulation, I think most others out there would take the initial portfolio value and increase it by 10% then by 5%. What you seem to be saying is that you at some point would multiply by something like 4.44 / 2930.
To say the same thing a different way, a typical Monte Carlo simulation will have a table by year of annual investment returns, and it will repeatedly sample randomly from that yearly table. You seem to be sampling from a table by year of S&P levels, which is, uh, waaaay wrong.
What I have seen typically is that Monte Carlo simulations tend to be about 25% to 50% (depending on the exact details of the implementation) more conservative than historical models. So if a historical model says that a 4% WR is safe, then a Monte Carlo model with equivalent inputs will say that a 2% or 3% WR is safe.
-
If I understand your description accurately, I would be very surprised if your Monte Carlo implementation is similar to others out there.
Let's take your example numbers from above, and further assume the market went up 5% in 1871 and 10% in 2020.
If your Monte Carlo simulation were to randomly pick 2020 and 1871 as the first and second years of the simulation, I think most others out there would take the initial portfolio value and increase it by 10% then by 5%. What you seem to be saying is that you at some point would multiply by something like 4.44 / 2930.
To say the same thing a different way, a typical Monte Carlo simulation will have a table by year of annual investment returns, and it will repeatedly sample randomly from that yearly table. You seem to be sampling from a table by year of S&P levels, which is, uh, waaaay wrong.
What I have seen typically is that Monte Carlo simulations tend to be about 25% to 50% (depending on the exact details of the implementation) more conservative than historical models. So if a historical model says that a 4% WR is safe, then a Monte Carlo model with equivalent inputs will say that a 2% or 3% WR is safe.
Now that I think about it, the way you’re describing must be how FI Calc is doing the Monte Carlo calculation, as FI Calc only uses % changes, like you describe, and never uses absolute stock price values. So the explanation given in my last post must be wrong. Sorry about that!
It is odd, though, that the results would be so negative when including data back to 1871. I’ll need to dig more into why that is.
Whatever the reason, I know that there’s a lot of work to do with the Monte Carlo side of things. It’s nowhere near the point that it needs to be for folks to use it for evaluating portfolios, which is why it isn’t a public feature. It’s probably best to just ignore it for now.
It’ll likely be a couple of weeks before I have time to really dig into it to make sure that it’s behaving as expected. And once it’s ready there will be an accompanying page in the guides that explain the algorithm in detail ✌️
-
This is really exceptionally good, well done ! The options for dynamically changing allocation , the floor and ceiling withdrawal amounts and different withdrawal strategies are very useful, and the UI is spot on given it's something that could be over-complicated.
One observation that would make it even more useful for me, the 1/n and VPW withdrawal strategies have a target amount remaining. But the simulation counts a success as ending up with >0 as opposed to > target amount remaining. Using the target amount would make this simulation more meaningful for me as the target amount does not really matter.
Anyway, well done, and thank you, you should be proud of this.
-
This is really exceptionally good, well done ! The options for dynamically changing allocation , the floor and ceiling withdrawal amounts and different withdrawal strategies are very useful, and the UI is spot on given it's something that could be over-complicated.
Thanks so much for the kind words! It was tough to strike a balance between making the app feel approachable while still supporting a lot of different options. I'm glad I was at least somewhat successful :)
One observation that would make it even more useful for me, the 1/n and VPW withdrawal strategies have a target amount remaining. But the simulation counts a success as ending up with >0 as opposed to > target amount remaining. Using the target amount would make this simulation more meaningful for me as the target amount does not really matter.
This is a great idea. One of the upcoming features will allow you to tweak those values in the analysis, so that you could specify exactly that. A little longer ways out is a more complete custom analysis system.
Anyway, well done, and thank you, you should be proud of this.
Thanks so much! I put a ton of effort into it, and I'm glad it's proving useful for some folks :)
-
PTF
As a color confused person, a more distinctive difference between the not-so-good and the awful start years would be nice.
Maybe a brighter not-so-good notation. This might also help in mobile usage in brighter ambient light.
@markbike528CBX , it's been some time, but I've been working on an update to FI Calc to help folks with color vision deficiencies. If you have time, I'd love to know if you find these colors to be an improvement.
To try it out:
1. Navigate to https://staging.ficalc.app (note: this is different from the usual FI Calc URL because this feature isn't deployed to production yet)
2. In the configuration panel, you'll see "User Preferences" at the bottom
3. Select the color deficiency that you have (based on your description I'd guess it's either protanopia or deuteranopia)
4. Click "OK"
5. View the new colors under "Simulations by Start Years"
I haven't updated all of the colors in the app, yet; only the ones under Simulations by Start Year will be different.
-
PTF
As a color confused person, a more distinctive difference between the not-so-good and the awful start years would be nice.
Maybe a brighter not-so-good notation. This might also help in mobile usage in brighter ambient light.
@markbike528CBX , it's been some time, but I've been working on an update to FI Calc to help folks with color vision deficiencies. If you have time, I'd love to know if you find these colors to be an improvement.
To try it out:
1. Navigate to https://staging.ficalc.app (note: this is different from the usual FI Calc URL because this feature isn't deployed to production yet)
2. In the configuration panel, you'll see "User Preferences" at the bottom
3. Select the color deficiency that you have (based on your description I'd guess it's either protanopia or deuteranopia)
4. Click "OK"
5. View the new colors under "Simulations by Start Years"
I haven't updated all of the colors in the app, yet; only the ones under Simulations by Start Year will be different.
Personally, I like the protanopia setting the most.
Setting -- comments
none- orange and reds don't distinguish well. It may be the oranges are green, my color-visioned wife chimed in, but I forgot the exact colors she stated.
protanopia -- OK for me
deuteranopia -- pretty much the same as protanopia
tri-tanopia -- red and blacks too close for ease of distinguishing.
YMMV, in fact Your Mileage Will Vary :-)
Thanks for considering us color confused people!
test conditions to date:
large 40" UHD monitor in dim room.
-
Personally, I like the protanopia setting the most.
Setting -- comments
none- orange and reds don't distinguish well. It may be the oranges are green, my color-visioned wife chimed in, but I forgot the exact colors she stated.
protanopia -- OK for me
deuteranopia -- pretty much the same as protanopia
tri-tanopia -- red and blacks too close for ease of distinguishing.
YMMV, in fact Your Mileage Will Vary :-)
Thanks for considering us color confused people!
test conditions to date:
large 40" UHD monitor in dim room.
Thanks so much for trying this out, and I'm glad the protanopia/deuteranopia settings are an improvement for you.
I'm going to go through and update the rest of the colors in the app so that other things in the app should be more legible, too, and then I'll get this released. Thanks again @markbike528CBX !
-
Personally, I like the protanopia setting the most.
Setting -- comments
none- orange and reds don't distinguish well. It may be the oranges are green, my color-visioned wife chimed in, but I forgot the exact colors she stated.
protanopia -- OK for me
deuteranopia -- pretty much the same as protanopia
tri-tanopia -- red and blacks too close for ease of distinguishing.
YMMV, in fact Your Mileage Will Vary :-)
Thanks for considering us color confused people!
test conditions to date:
large 40" UHD monitor in dim room.
Thanks so much for trying this out, and I'm glad the protanopia/deuteranopia settings are an improvement for you.
I'm going to go through and update the rest of the colors in the app so that other things in the app should be more legible, too, and then I'll get this released. Thanks again @markbike528CBX !
I would have liked to test it on my mobile devices, but my WiFi router died, so only the big screen got tested. Maybe in a week or so when I get the new router.
-
I would have liked to test it on my mobile devices, but my WiFi router died, so only the big screen got tested. Maybe in a week or so when I get the new router.
I'd appreciate that if you have time. I just deployed this update so you can view these settings at the regular https://calculator.ficalc.app (https://calculator.ficalc.app) URL now.
-
I would have liked to test it on my mobile devices, but my WiFi router died, so only the big screen got tested. Maybe in a week or so when I get the new router.
I'd appreciate that if you have time. I just deployed this update so you can view these settings at the regular https://calculator.ficalc.app (https://calculator.ficalc.app) URL now.
Sure, I'll try to remember. I'm FIRed so usually nothing is pressing enough to be remembered. It is now on my paper calendar.
I forgot how much we use the WiFi, so I just had standard (slow) shipping. Dang
-
one thing no calculator I've seen does is Alternative investment such as Real Estate. It would be nice to see equity integrated into asset allocation somehow.
INFO: how RE market is trending vs equity growth...enables making informed decision about when/if to sell.
-
Another PTF so I can find this later. Thanks!
-
one thing no calculator I've seen does is Alternative investment such as Real Estate. It would be nice to see equity integrated into asset allocation somehow.
INFO: how RE market is trending vs equity growth...enables making informed decision about when/if to sell.
Presently, real estate can be integrated using the Additional Income feature of FI Calc. There's more info about how to do that here (https://ficalc.app/other/faq/#how-can-i-enter-real-estate-into-fi-calc%3F).
If additional features other than the ones described there would be useful to you, let me know! I'd love to make the calculator have even more robust support for RE.
Sure, I'll try to remember. I'm FIRed so usually nothing is pressing enough to be remembered. It is now on my paper calendar.
Ha, no problem at all if you don't remember, you've already been a huge help. Also, it's amazing that you're FIRE'd! It'll be awhile longer for me :)
Another PTF so I can find this later. Thanks!
✌️
-
This... this is amazing! Thanks! I'll be spending some time for sure in the next month or so exploring this calculator!
-
This... this is amazing! Thanks! I'll be spending some time for sure in the next month or so exploring this calculator!
Nice! Feel free to post any suggestions / questions / feature requests in this thread if you have any :)
-
Really clean interface, I enjoyed using the calculator.
One suggestion so far that I didn't find: Ability to execute the allocation glidepath faster. All of the functions are pretty slow. Maybe a user input for # of years until the transition is complete? For example, ERN's blog suggests moving from 60/40 to 90/10 over 10 years is better than increasing over 20 or 30 years.
Also a question - I didn't follow carefully all of the monte carlo discussion above. Is that a beta feature? In other words, if I'm using the currently published calculator it should not be monte carlo. Rather it should be similar to cfiresim (will run your scenario for every 30yr period in history). Is that correct?
Keep up the good work!
-
Really clean interface, I enjoyed using the calculator.
Thanks!
One suggestion so far that I didn't find: Ability to execute the allocation glidepath faster. All of the functions are pretty slow. Maybe a user input for # of years until the transition is complete? For example, ERN's blog suggests moving from 60/40 to 90/10 over 10 years is better than increasing over 20 or 30 years.
Thanks for the suggestion! I've taken a note to look more into this ✌️
Also a question - I didn't follow carefully all of the monte carlo discussion above. Is that a beta feature? In other words, if I'm using the currently published calculator it should not be monte carlo. Rather it should be similar to cfiresim (will run your scenario for every 30yr period in history). Is that correct?
Keep up the good work!
Your understanding is correct. Monte Carlo is in development, and is not available in the calculator. So the calculations currently use a similar approach to what you'd find in cFIREsim.
-
Hi, I'm back and now have WiFi in the house
Setting -- large 40" UHD monitor in dim room. Mac Pro(2013, trashcan)
none- orange and reds don't distinguish well. It may be the oranges are green, my color-visioned wife chimed in, but I forgot the exact colors she stated.
protanopia -- OK for me
deuteranopia -- pretty much the same as protanopia
tri-tanopia -- red and blacks too close for ease of distinguishing.
Setting -- iPhone 5s iOS 12.4.7 - dim screen
none- red and blacks pretty close but distinguishable, red and greens are actually better distinguished compared to bright screen
protanopia -- OK for me
deuteranopia -- pretty much the same as protanopia
tri-tanopia -- red and blacks much too close
Setting -- iPhone 5s - pretty bright screen
none- looks pretty good
protanopia -- OK for me
deuteranopia -- pretty much the same as protanopia
tri-tanopia -- red and blacks too close for _ease_ of distinguishing.
ipad2 iOS 9.3.5
Simply doesn't work
Black screen comes up, a small scroll indicator on right side, but that's it.
IPad Pro 13" (2017) bright screen
none- looks pretty good
protanopia -- OK for me
deuteranopia -- pretty much the same as protanopia
tri-tanopia -- red and blacks too close for _ease_ of distinguishing.
IPad Pro 13" (2017) dim screen
none- reds and blacks are really close
protanopia -- OK for me
deuteranopia -- pretty much the same as protanopia
tri-tanopia -- red and blacks too close for _ease_ of distinguishing.
This whole thing brings back a job I did, where I had to physically touch a screen and see what happened and see what colors came up.
I then had to write out in long hand, "white", "red", "blue" on a table in the worksheet paper what I had seen.
So I can confirm that the local system screens at Barakah Units 1,2,3,4 in the United Arab Emirates will be operable by color confused people. :-)
https://en.wikipedia.org/wiki/Barakah_nuclear_power_plant
-
My formal diagnosis is red-green. I can see, identify high saturation colors including fire engine red or Kelly green. Pastels are a big source of confusion, although I can see the color (not just grey).
Mom found out when I was in kindergarten, and the teacher and I had a disagreement on the color of a man's overall I was drawing. I said blue, she said purple.
I still have my last school art from 9th grade. I explained my issues to that teacher, who seemed ok with it.
Since you are using a black background, just turning up the RGB or hex values on the "yellow" would be good enough for me. YMMV
I usually only get to the first page of the Ishihara test. https://www.colour-blindness.com/colour-blindness-tests/ishihara-colour-test-plates/
On well kept, newer paper copies of the test, sometimes I can guess correctly on other pages.
Whoa, this is incredibly useful. Thank you so much for writing all of this up for me!
I'd love to get your thoughts on this idea I have: I'm not sure if one set of colors will provide the best experience for everybody, and I want this app to be the best experience for everybody. So rather than pick one set up of colors for everyone, what I can do is create different color schemes, so there will be an option that looks the best for, say, red-green. Once you set that once, the setting will be stored in your browser so you'll never need to set it again.
What do you think of that? If you agree that it could work, I can add it the app and then I'd love to get your feedback on the color choices.
Thanks again for sharing such useful feedback, @markbike528CBX !
If the colors are right next to each other, it is clearer. The problem is if the are few and scattered various colors (success can be confusing?).
I tried grey scale and then the very dark and the OK (black) blend too much.
If you make the lightest color too light then you blend with the button color.
I checked out the colors
ca1010 (failed) and B36300 (warning/near fail) at https://htmlcolorcodes.com/color-picker/ and they have similar attributes (RGB, CYMK values)
mmmm...... not sure what I can suggest....
Not so relevant to the current discussion but a good thinking book, and a Great coffee table book for geeks is:
The Visual Display of Quantitative Information
https://www.edwardtufte.com/tufte/books_vdqi
Can confirm the colours look fine to the moderately protanomalous.
ETA: Oh, I didn't see the options...
Default - best for me. Orange and blue are are a bit drab, but can distinguish everything well enough.
Protanopia - mauve and pink ok to tell apart, but not as easy to use at a glance. Probably aimed at dichromats or stronger anomalous trichromats that struggle more with the default.
Deuteranopia - identical to protanopia?
Tritanopia - blue and purple too similar for me. Red also duller than default red. Not surprising; just saying for completeness.
-
Hi, I'm back and now have WiFi in the house
This is amazing! Thanks for the thoroughness.
I'm not surprised that the app doesn't work on the iPad 2. It takes more time and effort to support old devices/browsers, and I'm just the one person. I don't have a device with that operating system to even test on, for instance! I'm glad it worked on all of your other devices, though.
This whole thing brings back a job I did, where I had to physically touch a screen and see what happened and see what colors came up.
I then had to write out in long hand, "white", "red", "blue" on a table in the worksheet paper what I had seen.
So I can confirm that the local system screens at Barakah Units 1,2,3,4 in the United Arab Emirates will be operable by color confused people. :-)
https://en.wikipedia.org/wiki/Barakah_nuclear_power_plant
That is seriously cool. I'm going to add "FI Calc has been Quality Assured by a nuclear power plant Quality Assurance Engineer" to the homepage! (just kidding)
Can confirm the colours look fine to the moderately protanomalous.
ETA: Oh, I didn't see the options...
Default - best for me. Orange and blue are are a bit drab, but can distinguish everything well enough.
Protanopia - mauve and pink ok to tell apart, but not as easy to use at a glance. Probably aimed at dichromats or stronger anomalous trichromats that struggle more with the default.
Deuteranopia - identical to protanopia?
Tritanopia - blue and purple too similar for me. Red also duller than default red. Not surprising; just saying for completeness.
Thanks for the review!
I agree that the orange and blue are a bit drab. My primary goal is to ensure that the text remains legible for as many people as possible; aesthetics comes second. There's no way to make the orange brighter without sacrificing legibility, which then has a snowball effect on the rest of the color palette. I'd love for it to be brighter, though. Maybe one of these days I'll find some new colors that can brighten it up without sacrificing the color contrast.
-
.......
This whole thing brings back a job I did, where I had to physically touch a screen and see what happened and see what colors came up.
I then had to write out in long hand, "white", "red", "blue" on a table in the worksheet paper what I had seen.
So I can confirm that the local system screens at Barakah Units 1,2,3,4 in the United Arab Emirates will be operable by color confused people. :-)
https://en.wikipedia.org/wiki/Barakah_nuclear_power_plant
That is seriously cool. I'm going to add "FI Calc has been Quality Assured by a nuclear power plant Quality Assurance Engineer" to the homepage! (just kidding)
My last actual title was "Principal Field Service Engineer" , but I'm a chemist, dammit.
That job was an off-season filler. One other guy there had also been in fieldwork so we would trade true field stories that none of the computer testing people believed.
-
one thing no calculator I've seen does is Alternative investment such as Real Estate. It would be nice to see equity integrated into asset allocation somehow.
INFO: how RE market is trending vs equity growth...enables making informed decision about when/if to sell.
Presently, real estate can be integrated using the Additional Income feature of FI Calc. There's more info about how to do that here (https://ficalc.app/other/faq/#how-can-i-enter-real-estate-into-fi-calc%3F).
If additional features other than the ones described there would be useful to you, let me know! I'd love to make the calculator have even more robust support for RE.
Sure, I'll try to remember. I'm FIRed so usually nothing is pressing enough to be remembered. It is now on my paper calendar.
Ha, no problem at all if you don't remember, you've already been a huge help. Also, it's amazing that you're FIRE'd! It'll be awhile longer for me :)
Another PTF so I can find this later. Thanks!
✌️
Additional income is never the problem...all calculators have that. How do you show equity of real estate in your asset allocation (alternative investment) is the grand question?
make sense? for instance say I have a $1M investment portfolio with $200k in REITs (20%) and then another $500k equity in a physical investment property (appraisal - mortgage); how to add the $500k into the alternatives asset in the allocation so it displays "full" portfolio not just traditional investment. $1.5M with 47% in Alternatives
Now I can see I have too much in RE and need to rebalance. Also nice to trend RE appreciation vs the other markets
-
Additional income is never the problem...all calculators have that. How do you show equity of real estate in your asset allocation (alternative investment) is the grand question?
make sense? for instance say I have a $1M investment portfolio with $200k in REITs (20%) and then another $500k equity in a physical investment property (appraisal - mortgage); how to add the $500k into the alternatives asset in the allocation so it displays "full" portfolio not just traditional investment. $1.5M with 47% in Alternatives
Now I can see I have too much in RE and need to rebalance. Also nice to trend RE appreciation vs the other markets
Ah okay, I understand what you're asking now. Thanks for the clarification.
This is something I've thought about, but so far I've decided against it because I'm not quite sure how it would work within the calculator. Physical real estate is quite different from the asset types listed. For one, it's not as straightforward to "rebalance" a portfolio that includes real estate as it is to rebalance, say, bonds and stocks. For example, there are additional fees like closing costs that are difficult to estimate.
One of the features of FI Calc is that the way that the numbers transform year over year are based on historical data. Historical data for homes in every market doesn't exist as far as I'm aware. Folks could enter their own values for growth and taxes, but that is at odds with the main value prop of a calculator like FI Calc. The cycles in the historical data are where the interesting results come from; analyses with static values are useful, but using historical data can be useful for other reasons. A goal of mine is to minimize the number of "static values" in the calculator, and rely on historical data as much as possible.
So yeah, I'd love to add physical real estate, but I first need to think of a way to include it in a way that makes sense. Do you have a more specific idea of how the algorithm for incorporating physical real estate might work?
Regarding REITs: I'm interested in adding it, and this is easier to add than physical property. I just need the historical data.
-
Is https://ficalc.app (https://ficalc.app) down for everyone or just for me?
-
its down.
So yeah, I'd love to add physical real estate, but I first need to think of a way to include it in a way that makes sense. Do you have a more specific idea of how the algorithm for incorporating physical real estate might work?
Regarding REITs: I'm interested in adding it, and this is easier to add than physical property. I just need the historical data.
to answer the rebalance...if you are too heavy/light in "alternatives" you would sell/buy REITs or some other form of traded alternative; not necessarily physical RE. The physical RE is just an anchor that appreciates and builds equity or falls the other way.
what bugs me with the mints or PC's out there...my equity adds to my networth but doesn't show a percentage of my total portfolio, so all the assumptions about future growth are skewed not accounting for investment properties.
Zillow seems to have an API that looks up Zestimate for any address, most property tax records are public...just a matter of time before someone finds a way to collect them in an automated fashion.
-
I’ve just tried to access the page using the link in the origina post and get Error 526. Anyone else having the same problem?
-
to answer the rebalance...if you are too heavy/light in "alternatives" you would sell/buy REITs or some other form of traded alternative; not necessarily physical RE. The physical RE is just an anchor that appreciates and builds equity or falls the other way.
Makes sense. I've added REITs to my list of things to look into ✌️
Sorry about FI Calc being down! I'm working on getting it running again now. In the meantime, you can access the calculator directly through the URL https://calculator.ficalc.app (https://calculator.ficalc.app). The issue is with the calculator guide website.
Update: this is a problem with GitHub (where the guide site is hosted). They're working on fixing it now.
-
ficalc.app is fixed! Sorry for the downtime.
-
A couple of updates over the past few days:
- Additional income/withdrawals can now be set as lasting indefinitely. This can be useful for, say, a pension or SS.
- The misleading "Target end portfolio value" has been removed from VPW (VPW does not allow you to set a target end portfolio value). It can still be found under CVPW, but it has been renamed to "PMT Future Value," which is a more accurate label.
- There's an update to the definition of success for Maximize Spend strategies like VPW: sims will count as failed if the final year withdrawal is smaller than your target minimum withdrawal. i.e.; if you want to spend $30k but your portfolio only has $20k in that final year, then that would count as a fail.
-
Thanks for amending the definition of success, will be very useful !
Noticed a bug in this release with the Maximum Annual Withdrawal for 1/N and VPW scenarios. If you use the scenario linked below it has a 97.86% success rate, but if you tick the box to limit the maximum withdrawal per year the success rate reduces to 67.14%. Cannot be correct as limiting your withdrawals cannot decrease your success.
https://calculator.ficalc.app?additionalIncome=%5B%5D&additionalWithdrawals=%5B%5D&bondsFees=0.05&bondsFinalRatio=15&bondsInitialRatio=15&cashFees=0&cashFinalRatio=5&cashGrowth=1.5&cashInitialRatio=5&changeAllocationsOverTime=false&equitiesFees=0.04&equitiesFinalRatio=80&equitiesInitialRatio=80&initialPortfolioValue=450000&maxWithdrawalLimit=31000&maxWithdrawalLimitEnabled=false&minWithdrawalLimit=30000&minWithdrawalLimitEnabled=true&numberOfYears=10&oneOverNTargetPortfolio=0&portfolioRebalanceEquation=linear&rebalance=true&rebalanceFrequency=1&retirementStartingAge=60&withdrawalStrategyName=oneOverN (https://calculator.ficalc.app?additionalIncome=%5B%5D&additionalWithdrawals=%5B%5D&bondsFees=0.05&bondsFinalRatio=15&bondsInitialRatio=15&cashFees=0&cashFinalRatio=5&cashGrowth=1.5&cashInitialRatio=5&changeAllocationsOverTime=false&equitiesFees=0.04&equitiesFinalRatio=80&equitiesInitialRatio=80&initialPortfolioValue=450000&maxWithdrawalLimit=31000&maxWithdrawalLimitEnabled=false&minWithdrawalLimit=30000&minWithdrawalLimitEnabled=true&numberOfYears=10&oneOverNTargetPortfolio=0&portfolioRebalanceEquation=linear&rebalance=true&rebalanceFrequency=1&retirementStartingAge=60&withdrawalStrategyName=oneOverN)
-
Thanks for setting this up! I'm enjoying it.
I do have a question on the additional income streams. Could you confirm my understanding of how the "no inflation" choice works?
Let's say I'm retiring today and in 10 years I'll start receiving a pension of $35k/year with no inflation adjustments either between now and when I can draw, or after I start drawing. Fixed dollar amount indefinitely. Is that how "no inflation" calculates things?
-
Thanks for setting this up! I'm enjoying it.
I'm glad to hear it!
I do have a question on the additional income streams. Could you confirm my understanding of how the "no inflation" choice works?
Let's say I'm retiring today and in 10 years I'll start receiving a pension of $35k/year with no inflation adjustments either between now and when I can draw, or after I start drawing. Fixed dollar amount indefinitely. Is that how "no inflation" calculates things?
Your understanding is correct. That is how "no inflation" calculates things, and the pension that you're describing sounds to me like a "no inflation" additional income stream.
Feel free to ask any other Q's if something isn't clear. Cheers!
-
Love the UI.
If this was already asked - my apologies.
TLDR; Can Mixed Portfolios be added?
One aspect I like about FireCalc - is that you can select various percentages of various asset classes in a given portfolio rather than a flat percentage of Equity/Fixed Income.
Conversely - CFireSim does not - but I really enjoy the range of Spending Plans, as you have with FICalc.
I have a mix of asset classes and I would like to see how that mix stacks up against all the Withdrawal Strategies you have. I suspect if my flat 75%/25% is good for 90% of the time - a mixed asset portfolio would easily be at 100%, but I am not too sure.
Any chance of getting this into a future version?
Again thank you - a wonderful mix of FireCalc and CFireSim.
-
I really enjoy the Available Spend feature!!! Thanks for making such a cool product
-
Love the UI.
If this was already asked - my apologies.
TLDR; Can Mixed Portfolios be added?
One aspect I like about FireCalc - is that you can select various percentages of various asset classes in a given portfolio rather than a flat percentage of Equity/Fixed Income.
Conversely - CFireSim does not - but I really enjoy the range of Spending Plans, as you have with FICalc.
I have a mix of asset classes and I would like to see how that mix stacks up against all the Withdrawal Strategies you have. I suspect if my flat 75%/25% is good for 90% of the time - a mixed asset portfolio would easily be at 100%, but I am not too sure.
Any chance of getting this into a future version?
Again thank you - a wonderful mix of FireCalc and CFireSim.
Great question! Unfortunately mixed assets aren’t supported, primarily because of the data source that I use (Shiller’s dataset), which doesn’t include info for those other asset types.
I’d love to expand that to support more sources, but I’m not sure of any free data sources that could be used. I’ve considered switching over to the SBBI Yearbook which likely has more detailed data, but it costs $250 for each new publication (which is annual).
One of the potential pros of Shiller’s data over something like SBBI is that it goes back to 1871 rather than 1926, so more sims are run. There is a question of how applicable that data is to the modern investor, but FI Calc gives you the choice about whether or not you want to include that data or not.
I really enjoy the Available Spend feature!!! Thanks for making such a cool product
Thank you!! I’m glad you’re enjoying it 🙂
-
As a Software Engineer myself, I'd love to just say that it looks great and is extremely responsive.
Are you using some sort of Single-Page-Application for tech?
-
As a Software Engineer myself, I'd love to just say that it looks great and is extremely responsive.
Are you using some sort of Single-Page-Application for tech?
Thanks! I've put a lot of effort into it, and it's definitely the culmination of all of the knowledge I've learned building apps on the web over the past ~10 years.
And you're dead-on that the calculator itself is a SPA, although there aren't very many "pages". It's built in React.
The guides are a separate project, and they are static pages with a sprinkle of JS in there for some of the interactive features, such as the nav.
-
So, one drawback of this approach (actual historical) for data is that if you want a 50* year simulation run, you can only start 50 years ago or further back.
What about an option to start on any year, but when you get to 2020, it "wraps around" either to the first year of the data set, or some other year within the data set? Heck, you could default to the start year, but allow the user to choose a "wrap around" year.
This avoids the biggest issue I have with Monte Carlo (it allows Great Depression immediately followed by the Great Recession immediately followed by the 2020 tech bust) by staying mostly with just following historical patterns - while increasing the number of viable start years.
*Or 30 years or whatever, the point is the same. It's hard to judge viability of a portfolio that started in 2007.