Author Topic: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator  (Read 39577 times)

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #50 on: August 16, 2013, 03:41:44 PM »

another feature request (or maybe it's there and I'm not understanding):  in your "investigate" section, is there a way to investigate retirement year?  In other words, find the "first fit" date where there is an X% success rate with the given inputs.

My scripts sort of brute force this (with firecalc) ... but there is bound to be a smarter way to do it.

Yeah, that is one "Investigation" that I haven't implemented yet.  I don't think it would be that hard to figure out X years given a portfolio, savings inputs, and success % criteria.  I'll add it to the list.

My brute force method is just a loop.  Try.  If success, stop, otherwise add a year and start over.  There is probably a decent formula there... but I ain't got it.

Yeah, both the "Spending" and "Portfolio" investigations use what is called the "Binary Search method". 

For the "Years till retirement" example, you'd take a min of 0 years, max of say 100, and find the mid (50). You ask "Is the middle number too high or low?"  It's probably too high, so you find the new mid point (between 0 and the old mid, so 25). Then you ask "is the mid too high or low?" Cut the mid in half and bounce around until you find the right value.

Faster than brute force :)

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #51 on: August 16, 2013, 03:48:55 PM »
Thanks for putting your efforts into this. Is there anything that jumped out that you were not expecting?  Anything, that everyone should understand or know?  Types of environments that cause the most challenges, etc.  I know I have looked at Firecalc and other information over the years and noticed a few things, but I would think that the amount of time that you have spent, some information might be worth sharing, besides the calculator in itself.  These calculators seem like hocus pocus, until you really understand what they are, what they tell us, and how they were derived.  Someone, who just plugs their information into them, may not truly understand what the results are t telling them.   

Thanks for all your effort in this project!

Tom

Interesting question.  A few things I've noticed by figuring out the intricacies of this program:

- Inflation is worse than a bad market. The market loss during the great depression was nothing compared to the inflation of the 1970s.
- If you're saving a large percentage of your income, there is usually a pretty big difference in retiring within 1-2 years of your anticipated date.  The earliest possible date may give you 95% success on a just-above-barebones budget, while 1 year later might be 100% success on +15% of that spending. I can totally understand folks with "On more year syndrome".
- The absolute worst thing for a retirement is a major market drop in years 1-3(ish). If that happens, get yourself back to work for a little bit, or get SOME sort of supplemental income (if you don't have it already), and it will make a HUGE difference in the long run.
- There are some interesting "withdrawal strategies" out there that allow for more withdrawals during the good times, and cutting back during the bad times, which yield a higher success rate AND a higher amount of total withdrawals.  That sort of "cutting back" during bad market years is common sense, but there are actual strict rules to some of these strategies that might help people out.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #52 on: August 17, 2013, 08:00:55 AM »
- There are some interesting "withdrawal strategies" out there that allow for more withdrawals during the good times, and cutting back during the bad times, which yield a higher success rate AND a higher amount of total withdrawals.  That sort of "cutting back" during bad market years is common sense, but there are actual strict rules to some of these strategies that might help people out.

Which of these have you found that are the best, and legitimate?

There's so many of these ideas out there that are just untested theory, and complete B.S. (see "buckets" guy who was recently outed for falsifying data, and got in a bunch of legal trouble because of that).
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #53 on: August 17, 2013, 10:37:09 AM »
- There are some interesting "withdrawal strategies" out there that allow for more withdrawals during the good times, and cutting back during the bad times, which yield a higher success rate AND a higher amount of total withdrawals.  That sort of "cutting back" during bad market years is common sense, but there are actual strict rules to some of these strategies that might help people out.

Which of these have you found that are the best, and legitimate?

There's so many of these ideas out there that are just untested theory, and complete B.S. (see "buckets" guy who was recently outed for falsifying data, and got in a bunch of legal trouble because of that).

Caveat: I'm no economist. This is just from my experience building this tool and messing with other people's "withdrawal strategies".

It's a bit of a loaded question.  It depends on what your goal is.  Some strategies are based on the idea of trying to deplete to zero.  Some want to leave the principal intact. Some are willing to lower and increase withdrawals based on the market.  Some are just WAY too complicated. If you're intensely frugal, and have no need/want to increase withdrawals, maybe some of these strategies aren't for you.

There are 3-4 that are on cfiresim.com now. Some are in the /dev/ version, not the main site.

Hebeler's Autpilot - Is essentially a mix between the 4% rule, and RMD.  Which means that as you age, if the market is doing ok/good, you will be able to take out more money with greater confidence.  It's also a fairly straightforward calculation. I tend to like it.

Guyton-Klinger - 2 fairly smart economists made a very very complicated withdrawal method in the early 2000s.  I implemented (but still testing) a simplified version.  Frankly, I find it too complicated and I see massive cuts to withdrawals during bad years.  It doesn't appeal to me.

"Retire Again & Again" - One of the few users that is actually signed up on my forums is also a member of e-r.org and bogleheads and is fascinated with withdrawal methods.  He created his own that is actually fairly straightforward.  Start with a pretty conservative WR (3-3.5%) and keep your spending equal in constant dollars as the years go by. If your portfolio (in constant dollars) ever gets above a certain threshold (default is 5% higher than the initial portfolio, in constant dollars), then the model has deemed it safe to increase your withdrawals.  This model is pretty good in that it is straightforward, it gives you "raises" ("retire again"!) if the market is good, and starts off conservative so that it will last. (remember what I said earlier in this thread, one of the MAJOR killers is a big decline in years 1-3).  So far, I like this one the most, as there is a certain psychological advantage to never having to cut (but rather starting off conservatively).



Nords

  • Magnum Stache
  • ******
  • Posts: 3421
  • Age: 63
  • Location: Oahu
    • Military Retirement & Financial Independence blog
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #54 on: August 17, 2013, 10:55:52 PM »
- There are some interesting "withdrawal strategies" out there that allow for more withdrawals during the good times, and cutting back during the bad times, which yield a higher success rate AND a higher amount of total withdrawals.  That sort of "cutting back" during bad market years is common sense, but there are actual strict rules to some of these strategies that might help people out.
Which of these have you found that are the best, and legitimate?
Guyton's has a lot of potential, but it's so darn complicated that it's begging for a computer program.  Bud Hebeler's is very simple (he recommends charting it out once a year) and might be easier to code into a simulation.

Bob Clyatt also did a lot of computer simulation for "Work Less, Live More" portfolios, and that's how he was able to test the 4%/95% rule.

There's so many of these ideas out there that are just untested theory, and complete B.S. (see "buckets" guy who was recently outed for falsifying data, and got in a bunch of legal trouble because of that).
The real irony behind Ray Lucia is that his system might actually work, except for the part where he lied about it.  One of the bucket "drawbacks" was that the retiree would eventually end up with a portfolio heavy in equities, yet Wade Pfau is starting to analyze portfolios that start with an annuity purchased from an investor's bond allocation.  The result is also a retirement portfolio heavy in equities, and because the owner has an annuity they're willing to tolerate more equity volatility while they're beating inflation.

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #55 on: August 20, 2013, 06:55:43 PM »
FWIW, just did a big update on the main page. Moved everything over from the development version that I'd been working on.

- Added "Retire Again & Again" withdrawal method (my personal favorite as of now)
- Fixed Guyton-Klinger withdrawal method (still complicated to understand)
- Added "End Year" to social security inputs for some morbid users ;)
- Added "Cycle Start Year" to the hover-over information on the graph, so you can see what year each series started at (without doing mental math)
- Added ability to weight the 2 different strategies that are averaged for "Hebelers Autopilot" withdrawal method.
- Added ability to export to CSV (in debug section)
- Added more to the "Inputs" section of the Output page

Since we were just talking about the withdrawal strategies, it would be fun to just play with them.  I suggest trying "Retire Again & Again" with a 3.5% starting WR (Spending/Portfolio), and see what it shows you. In many cases, it'll double/triple your constant-dollar spending by the end, without failing.

tomsang

  • Handlebar Stache
  • *****
  • Posts: 1085
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #56 on: August 21, 2013, 08:28:29 AM »
Bo_knows can your calculator account for a mortgage?  Most calculators account for the payment as an expense vs. principal and interest. https://forum.mrmoneymustache.com/investor-alley/paying-off-mortgage-early-how-bad-is-it-for-your-fi-date/msg124876/#msg124876

I think it would be very educational if we could see how paying down a 30 year fixed sub 4% mortgage hurts or helps your success rate.

Thanks again for taking on this project.

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #57 on: August 21, 2013, 08:37:10 AM »
Bo_knows can your calculator account for a mortgage?  Most calculators account for the payment as an expense vs. principal and interest. https://forum.mrmoneymustache.com/investor-alley/paying-off-mortgage-early-how-bad-is-it-for-your-fi-date/msg124876/#msg124876

I think it would be very educational if we could see how paying down a 30 year fixed sub 4% mortgage hurts or helps your success rate.

Thanks again for taking on this project.

The only way to currently compensate for paying off a mortgage is to add a "Other Income/Saving" input for the years after you pay it off, to compensate for less spending requirements.   For example: If you have $40k in expenses (with a mortgage) and the mortgage part is $10k, and it's due to be paid off in 2020, you can add a "Other Income/Saving" input of $10k from 2020-2100 to compensate for paying it off.  If you want to play with a payoff date, you can inch that input back year-by-year.

I'm not entirely sure how I'd implement "loans" into the equation, other than as a reduction in expenses (after it's paid off).
« Last Edit: August 21, 2013, 08:39:22 AM by bo_knows »

oldtoyota

  • Magnum Stache
  • ******
  • Posts: 3179
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A Firecalc Clone
« Reply #58 on: September 20, 2013, 08:03:15 PM »
Great job, Bo! Thanks for contributing your work and talent to the community.

Nice work! I like how there's a starting and ending year option for savings. My spouse and I will not stop working at the same time, and this is the first calc I've found that would allow me to factor that into the equation.

Thank you!

Jed

  • 5 O'Clock Shadow
  • *
  • Posts: 12
  • Location: Granada, Spain
    • Bucking the Trend
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #59 on: September 25, 2013, 02:48:16 PM »
I thought you may be interested in a Q&A I did with "bo_knows" - the creator of this awesome tool.  Some info has already been covered in this thread, but it also gives you a glimpse into what inspired him to create cFIREsim.

http://bucking-the-trend.com/get-to-know-bo-creator-of-cfiresim/

SnackDog

  • Handlebar Stache
  • *****
  • Posts: 1260
  • Location: Latin America
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #60 on: September 28, 2013, 02:17:48 PM »
Thanks for putting your efforts into this. Is there anything that jumped out that you were not expecting?  Anything, that everyone should understand or know?  Types of environments that cause the most challenges, etc.  I know I have looked at Firecalc and other information over the years and noticed a few things, but I would think that the amount of time that you have spent, some information might be worth sharing, besides the calculator in itself.  These calculators seem like hocus pocus, until you really understand what they are, what they tell us, and how they were derived.  Someone, who just plugs their information into them, may not truly understand what the results are t telling them.   

Thanks for all your effort in this project!

Tom

Interesting question.  A few things I've noticed by figuring out the intricacies of this program:

- Inflation is worse than a bad market. The market loss during the great depression was nothing compared to the inflation of the 1970s.
- If you're saving a large percentage of your income, there is usually a pretty big difference in retiring within 1-2 years of your anticipated date.  The earliest possible date may give you 95% success on a just-above-barebones budget, while 1 year later might be 100% success on +15% of that spending. I can totally understand folks with "On more year syndrome".
- The absolute worst thing for a retirement is a major market drop in years 1-3(ish). If that happens, get yourself back to work for a little bit, or get SOME sort of supplemental income (if you don't have it already), and it will make a HUGE difference in the long run.
- There are some interesting "withdrawal strategies" out there that allow for more withdrawals during the good times, and cutting back during the bad times, which yield a higher success rate AND a higher amount of total withdrawals.  That sort of "cutting back" during bad market years is common sense, but there are actual strict rules to some of these strategies that might help people out.

Inflation is only a problem if it occurs when market conditions are poor - stagflation.  That being said, my grandfather retired in the worst year on record -1966- on a Colorado teaching pension and he lived 34 more years pretty well with plenty left over. But, he was frugal, for the most part.

If the market drops in your first several years, reset your withdrawal rate to 4% of the current balance. If you can't live on that, it is back to work or sell a property.

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #61 on: October 11, 2013, 06:15:12 AM »
Update: http://www.cfiresim.com/phpbb/viewtopic.php?f=3&t=102

Currently in the /dev/ section of the site (www.cfiresim.com/dev/input.php) is the ability to Login, Save Simulations, and Load simulations (even ones other people save). To use the login feature, you use the same credentials as the cFIREsim forums. Therefore, you need an account there to do so.

There are an incredible amount of different combinations of Load/Save/Delete that I haven't tested yet.  Any extra hands would be much appreciated. :)
« Last Edit: October 11, 2013, 06:21:28 AM by bo_knows »

footenote

  • Pencil Stache
  • ****
  • Posts: 801
  • MMMing in MN
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #62 on: October 11, 2013, 07:12:32 AM »
Thanks in advance for testing and reporting any bugs you find - this is an invaluable utility for planning and our feedback can make it even better.

I have provided minor help in the past and found Bo is very gracious to work with. He truly appreciates and responds to our input. Thanks again!

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #63 on: October 23, 2013, 06:59:53 PM »
Pretty large update. I know that a lot of people have complicated inputs.  Now, you can save them and load them another time.

www.cfiresim.com/input.php

- Added "Login" feature. This allows registered users of the cFIREsim forums to do the following:
* Save a set of inputs for a simulation, with a name
* Load a Saved Simulation by name
* Delete Saved Simulations
* Edit the name of Saved Simulations

- Added a feature for registered or non-registered users to load a Saved Simulation by a unique ID number.
- Changed VPW method of spending to "Custom VPW" and revamped it using updated work from the bogleheads forums.
- Removed old "Test String" functionality in lieu of new Save Simulation functionality.
- Changed wording of "Debug" section to reflect it's only current use: "CSV Output".

Random

  • Stubble
  • **
  • Posts: 125
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #64 on: October 24, 2013, 08:36:00 AM »
Wow.  Let me add my voice to the chorus of thanks for your great and continued efforts to build this out in such a cool, functional fashion.

Siamond

  • 5 O'Clock Shadow
  • *
  • Posts: 39
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #65 on: October 26, 2013, 07:02:26 PM »
Hi there.

I am one of the primary troublemakers who keeps giving Bo more stuff to do on cFIREsim... As he indicated in an earlier post, I am quite interested by withdrawal methods. It totally mystifies me that some people are extremely methodical with investments and optimizing spend and tracking budget, but yet just wing it when it comes to planning for retirement withdrawals... So I played with various methods & metrics with my Excel sheet, and have been torturing Bo ever since to improve his wonderful tool to support more and more features... ;-)

After exploring Guyton-Klinger, then VPW (to which I contributed a bit), then creating this RA&A simple approach (which has some drawbacks - just to be clear), playing a bit with some other ideas, I made a big loop, and finally realized that once you factor in the OTHER sources of income (SS/Pension kicking, inheritances, some side/part-time job post-retirement), then a variable method like a simplified form of G-K combined with a spending floor and cap is my preferred method... Buy hey, I'm still learning, so we'll see where this journey will bring me... Notably when I start doing it for real with my early retirement to come (soon!)... ;-)

In any case, kudos to Bo for the best online retirement calculator I've seen so far, and for being so accommodating with feature requests.
« Last Edit: October 26, 2013, 07:04:42 PM by Siamond »

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #66 on: October 27, 2013, 11:56:28 PM »
Hi there.

I am one of the primary troublemakers who keeps giving Bo more stuff to do on cFIREsim... As he indicated in an earlier post, I am quite interested by withdrawal methods. It totally mystifies me that some people are extremely methodical with investments and optimizing spend and tracking budget, but yet just wing it when it comes to planning for retirement withdrawals... So I played with various methods & metrics with my Excel sheet, and have been torturing Bo ever since to improve his wonderful tool to support more and more features... ;-)

After exploring Guyton-Klinger, then VPW (to which I contributed a bit), then creating this RA&A simple approach (which has some drawbacks - just to be clear), playing a bit with some other ideas, I made a big loop, and finally realized that once you factor in the OTHER sources of income (SS/Pension kicking, inheritances, some side/part-time job post-retirement), then a variable method like a simplified form of G-K combined with a spending floor and cap is my preferred method... Buy hey, I'm still learning, so we'll see where this journey will bring me... Notably when I start doing it for real with my early retirement to come (soon!)... ;-)

In any case, kudos to Bo for the best online retirement calculator I've seen so far, and for being so accommodating with feature requests.

Thanks for your contributions in ideas for cFIREsim. It does indeed kick ass, as does Bo_Knows!
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

kkbmustang

  • Handlebar Stache
  • *****
  • Posts: 1285
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #67 on: October 28, 2013, 10:05:26 PM »
Pretty large update. I know that a lot of people have complicated inputs.  Now, you can save them and load them another time.

www.cfiresim.com/input.php

- Added "Login" feature. This allows registered users of the cFIREsim forums to do the following:
* Save a set of inputs for a simulation, with a name
* Load a Saved Simulation by name
* Delete Saved Simulations
* Edit the name of Saved Simulations

- Added a feature for registered or non-registered users to load a Saved Simulation by a unique ID number.
- Changed VPW method of spending to "Custom VPW" and revamped it using updated work from the bogleheads forums.
- Removed old "Test String" functionality in lieu of new Save Simulation functionality.
- Changed wording of "Debug" section to reflect it's only current use: "CSV Output".

You, sir, are a rockstar.

NinetyFour

  • CMTO 2023 Attendees
  • Walrus Stache
  • *
  • Posts: 6876
  • Age: 62
  • Location: Southwestern US
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #68 on: December 01, 2013, 08:05:57 PM »
I have been enjoying playing around with cFIREsim.  One question, or wish, I guess.  Until I retire, the bulk of my funds will be stuck at TIAA, with relatively high fees.  As soon as I quit my full-time gig, I will yank that money out and transfer it to Vanguard where fees are nice and low.  I guess I wish I could tell cFIREsim that my fees for the next 7 years will be x% but thereafter will drop to .1% or so.  Sorry if that question/wish was expressed already.  Many thanks!

Nords

  • Magnum Stache
  • ******
  • Posts: 3421
  • Age: 63
  • Location: Oahu
    • Military Retirement & Financial Independence blog
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #69 on: December 01, 2013, 08:10:06 PM »
I guess I wish I could tell cFIREsim that my fees for the next 7 years will be x% but thereafter will drop to .1% or so.  Sorry if that question/wish was expressed already.  Many thanks!
One workaround would be to show the lower expense by adding the fee difference to your income starting after the seventh year. 

NinetyFour

  • CMTO 2023 Attendees
  • Walrus Stache
  • *
  • Posts: 6876
  • Age: 62
  • Location: Southwestern US
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #70 on: December 01, 2013, 08:13:12 PM »
Ah, thanks, Nords, I will try that!

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #71 on: December 03, 2013, 08:00:43 AM »
I have been enjoying playing around with cFIREsim.  One question, or wish, I guess.  Until I retire, the bulk of my funds will be stuck at TIAA, with relatively high fees.  As soon as I quit my full-time gig, I will yank that money out and transfer it to Vanguard where fees are nice and low.  I guess I wish I could tell cFIREsim that my fees for the next 7 years will be x% but thereafter will drop to .1% or so.  Sorry if that question/wish was expressed already.  Many thanks!

Interesting use-case.  I could add in a drop-down menu to show "Constant Fees %" or "Variable Fees %", and if Variable was chosen, something similar to the "Gradual Allocation Change" menu could pop up.  You could then say "1% fees from 2013-2023, then 0.1% fees from 2023-2063".  I'll add it to the "to-do" list.  In the meantime, stick with Nords' workaround.

NinetyFour

  • CMTO 2023 Attendees
  • Walrus Stache
  • *
  • Posts: 6876
  • Age: 62
  • Location: Southwestern US
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #72 on: December 03, 2013, 10:35:58 AM »
Sounds great.  Thanks very much for all that you do!

Siamond

  • 5 O'Clock Shadow
  • *
  • Posts: 39
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #73 on: December 05, 2013, 07:10:20 AM »
Interesting use-case.  I could add in a drop-down menu to show "Constant Fees %" or "Variable Fees %", and if Variable was chosen, something similar to the "Gradual Allocation Change" menu could pop up.  You could then say "1% fees from 2013-2023, then 0.1% fees from 2023-2063".  I'll add it to the "to-do" list.  In the meantime, stick with Nords' workaround.

Yeah, well, it's probably more complicated than that. Fees may indeed vary in time, but (probably more importantly for most people) they also vary depending on the asset category and the fund(s) you selected for such asset category... And differences could be quite significant. I don't know that there is a good answer without over complicating the user interface...

PS. Somehow averaging out your fees (present and future) would seem like an ok approximation, in absence of a better way.

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Re: The Crowdsourced FIRE Simulator (cFIREsim) - A new Retirement simulator
« Reply #74 on: January 20, 2014, 02:14:39 PM »
I wanted to add that I updated cFIREsim to version 1.25 recently.

This makes a change to how bond returns are calculated.  Through many discussions, including this one on bogleheads: http://www.bogleheads.org/forum/viewtopic.php?f=10&t=130068, one on the cFIREsim forums: http://www.cfiresim.com/phpbb/viewtopic.php?f=2&t=119, and some discussion here... it was stated that the way cFIREsim and Firecalc were calculating bond returns was wrong.

Net result: I think this makes bonds return slightly less than previously, decreasing the SWR for most people (affecting you more if you have more of a bond allocation).

To rehash what I've always stated: If there are any features out there in a retirement calculator that you'd be interested in, don't hesitate to start a discussion. I'm currently working on a "investigate retirement year" option that will be able to tell you what year you can retire based on your inputs.  That's been a long time coming, and I've been lazy on this project as of lately :)

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Mustachians,

I'm about to release a fairly hefty update to cFIREsim, but I wanted to point a few of you in the direction of the "development" version of the site, so you can play with it for a couple of days before I release it... to see if there are any bugs ;)   If you could play around with it more than you usually would, and report any bugs (here or in the cFIREsim forum), it would be much appreciated.

Development Site: http://www.cfiresim.com/dev/input.php
Now, don't bookmark that site, because it's ever-changing as I develop more features. I'll remove the link from this post in a few days.

Here are some things to check out:
Code: [Select]
- Updated Output Statistics Table
- added Median/St.Dev./Highest/Lowest for "Yearly Withdrawals"
- added Median/St.Dev./Highest/Lowest for "Total Withdrawals"
- added "Withdrawal Analysis"
- Shows Avg/Median/St.Dev./Highest/Lowest/Failures for 4 different subsets of your retirement period
- First 5 years
- Beginning Third
- Middle Third
- Final Third
- added "Dip Analysis"
- Shows how many simulation cycles that your Portfolio/Withdrawals dipped >10%, >20%, >40%, or >60% below the initial value at retirement.
- If a simulation cycle did show a dip for one of those categories, it shows the max total # of years it dipped during that specific simulation cycle.
- added Lowest 5 Withdrawal Dips
- Shows dip amount
- Shows Cycle Start / Dip Year
- Updated "Help" section for Output Table with better explanations of the fields
- Added 2 new fields to "Investigate" Section
- Added "Withdrawal Failure Value"
- This defines what a "failure" is to cFIREsim, in regards to Withdrawals (spending).  If spending gets below this value, it will be considered a failure.
- Added "Portfolio Failure Value"
- This defines what a "failure" is to cFIREsim, in regards to your portfolio value. By default, this value was always $0. If your portfolio ever got to $0, it was considered a failure. Now, this field is customizable.
- Changed "Investigate Fees" to "Investigate Drag (Fees/Market Underperformance/Inflation)" to represent alternative ways to use this feature
- Increased the range of investigation from 0-3% to 0-10%
- Added the ability to Save/Load "Investigate" options
- Fixed bug in Original VPW withdrawal method, where inflation was not factoring into results
- Changed spending cap in "Investigate Spending Level for ______ % success" to $1,000,000.
- Added a small checkbox next to "Save Simulation as:" field
- If you save a simulation, you can now check this box to disable the "Save Simulation as:" text field so you can re-run the same simulation without error.

Here is a preview of the new Output Statistics Table, which should have some of you statistics geeks drooling: http://imgur.com/DWep9V8

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
OOohh, shiny!
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

dude

  • Handlebar Stache
  • *****
  • Posts: 2369
Incidentally, on the subject of withdrawal rate adjustments during market downturns, here is another strategy to consider:

http://www.bloomberg.com/news/2013-04-09/maybe-past-performance-does-predict-your-savings-future.html

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Incidentally, on the subject of withdrawal rate adjustments during market downturns, here is another strategy to consider:

http://www.bloomberg.com/news/2013-04-09/maybe-past-performance-does-predict-your-savings-future.html

Huh.  While I don't currently have P/E ratio data in my database, it is available in the Shiller Data set.  Could be interesting to do in the future, thanks.

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
I figured I'd ask this here:

It's hard for me to look at cFIREsim from the prospective of the average user.  Would anyone be interested in detailed explanations of specific things in cFIREsim, outside of the FAQ/Tutorial that already exists?  Is there things that confuse you about the tool? Should there be more graphical representations of the various spending methods?

I'm considering actually starting the "blog" section of the website, but am not sure how much actual content you could write about outside of a few case studies.  Maybe I could help do case studies of specific people?   Just thinking out loud here.

grantmeaname

  • CM*MW 2023 Attendees
  • Walrus Stache
  • *
  • Posts: 5961
  • Age: 31
  • Location: Middle West
  • Cast me away from yesterday's things
Case studies could be a really helpful way to do it! Sometimes it helps to hear things in other words, too, so you could have guest posters case study themselves and then it could be the same concepts covered a little differently.

Dodge

  • Pencil Stache
  • ****
  • Posts: 790
I figured I'd ask this here:

It's hard for me to look at cFIREsim from the prospective of the average user.  Would anyone be interested in detailed explanations of specific things in cFIREsim, outside of the FAQ/Tutorial that already exists?  Is there things that confuse you about the tool? Should there be more graphical representations of the various spending methods?

I'm considering actually starting the "blog" section of the website, but am not sure how much actual content you could write about outside of a few case studies.  Maybe I could help do case studies of specific people?   Just thinking out loud here.

Case studies would be lots of fun!  Your simulator is incredible for planning the future, and realizing that yes indeed we CAN retire if we make a few changes and work part-time for a few years...etc.  Seeing some real-life examples of this would really drive the point home.

In terms of feedback, I thought the simulator was very easy to use.  I haven't checked out the FAQ/Tutorial at all, as the features are intuitive enough for me to figure it out.  One surprising thing I noticed, is anything we put into "Extra Spending" doesn't show up in the "withdrawal" sections of the chart at the end.  Neither does any "Extra Income".  If my extra income exceeds my yearly spending for the first 5 years, it should show the withdrawal at 0, for example.

Thanks for all the incredible work!

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator

In terms of feedback, I thought the simulator was very easy to use.  I haven't checked out the FAQ/Tutorial at all, as the features are intuitive enough for me to figure it out.  One surprising thing I noticed, is anything we put into "Extra Spending" doesn't show up in the "withdrawal" sections of the chart at the end.  Neither does any "Extra Income".  If my extra income exceeds my yearly spending for the first 5 years, it should show the withdrawal at 0, for example.

Thanks for all the incredible work!

Yeah, this is a carry-over from reverse-engineering Firecalc.  Rest assured it IS taking that money out (or putting it back in) of the portfolio, it's just not considered as "Spending". Part of the reasoning is that if I did make it part of the spending, and you had a large enough 1-time expenses, it would completely break the logic being used for any of the "variable" spending plans that depend on keeping your spending within certain limits.

EngineerMum

  • Stubble
  • **
  • Posts: 130
  • Location: Perth, Western Australia
  • Working towards moderate badassity
Can anyone clarify whether the retire again and again function is working? I've tried a bunch of simulations with that option and they all seem to adjust only once the stay at that level. Oh, and never go below my starting withdrawal, even though I selected no floor.

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Can anyone clarify whether the retire again and again function is working? I've tried a bunch of simulations with that option and they all seem to adjust only once the stay at that level. Oh, and never go below my starting withdrawal, even though I selected no floor.

This is probably more suited for the cFIREsim forums, but I do see a weird inconsistency in that Spending Model. I'll check it out.

shuffler

  • Pencil Stache
  • ****
  • Posts: 572
Is there things that confuse you about the tool?  Should there be more graphical representations of the various spending methods?
I was playing with the "Variable Spending (new)" method.  I set bounds of $70k - $90k, yet the table at the end of the simulation tells me that the average yearly withdrawal was $93k and the median was $94k.

The average & median being higher than the upper-bound is confusing to me.
Maybe it's a bug?  Or maybe (for people like me) it needs a bit more explanation?

Thanks.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Is there things that confuse you about the tool?  Should there be more graphical representations of the various spending methods?
I was playing with the "Variable Spending (new)" method.  I set bounds of $70k - $90k, yet the table at the end of the simulation tells me that the average yearly withdrawal was $93k and the median was $94k.

The average & median being higher than the upper-bound is confusing to me.
Maybe it's a bug?  Or maybe (for people like me) it needs a bit more explanation?

Thanks.

No idea, but a thought: Maybe you set 90k real, and 93k nominal was the average?  (I.e. it went above 90k due to inflation, but it never crossed 90k in real dollars?)
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

shuffler

  • Pencil Stache
  • ****
  • Posts: 572
 
No idea, but a thought: Maybe you set 90k real, and 93k nominal was the average?  (I.e. it went above 90k due to inflation, but it never crossed 90k in real dollars?)
I thought that, but:
    (All amounts are listed in Today's Dollars)

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Everything in the ending chart is in Today's Dollars.   I'm betting that you set your Yearly Spending above the Spending Ceiling (which I apparently don't check).... You should be forced to set the initial spending in between the Ceiling and Floor values.

I just did a test of $2M portfolio, $80k spending, 70k/90k floor/ceiling, and the average was ~$85k

http://www.cfiresim.com/input.php?id=28417

shuffler

  • Pencil Stache
  • ****
  • Posts: 572
I'm betting that you set your Yearly Spending above the Spending Ceiling (which I apparently don't check)....
Nope.

But sorry, I should have given repro steps from the start.
Listing only what I changed from the default values:
    Retirement year: 2020
    Years to model:  60
    Equities:        100% (Bonds 0%)
    Spending:        $80,000
    Method:          Variable Spending (new)
    Floor:           Defined Value / $70,000
    Ceiling:         Defined Value / $90,000


This produces Yearly Withdrawals data of:
    Average:  $91,493
    Median:   $94,321
    Max:      $90,000

shuffler

  • Pencil Stache
  • ****
  • Posts: 572
I'm betting that you set your Yearly Spending above the Spending Ceiling (which I apparently don't check)....
    Retirement year: 2020
I think the retirement-year is the key.
I can repro the problem using the link you gave, and then adjusting only the retirement year to 2020.

I think you were on the right track about setting the spending to be above the ceiling.  I'm not doing that directly, but I think that setting the retirement-year into the future is causing the spending to get scaled-up by inflation/CPI for those 6 years before the spending simulation starts in 2020.

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Ahhh, you know what? I haven't even looked at the code, but after reading your comment about the retirement year and looking at the example, I think I figured it out.

See how the 2nd chart (the spending chart) very obviously shows the ceiling ($90,000) and the floor ($70,000) never get crossed?  Yet, the averages are above that?  I almost guarantee that the average calculation is getting screwed up somehow by the number of years, since it should only be counting the years of retirement. 

I'll let you know what I find. Thanks for the additional analysis.

shuffler

  • Pencil Stache
  • ****
  • Posts: 572
I almost guarantee that the average calculation is getting screwed up somehow by the number of years, since it should only be counting the years of retirement.
Hmm.  But I wouldn't expect that to also affect the median, and cause the median to also be above the ceiling ...

Anyhow, thanks for looking.  I'll be interested to hear what you find.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Plus, if that was the bug, wouldn't the average be LOWER (because it's counting year it shouldn't, before retirement, and dividing by more years = a lower average).  I'd think that could be the bug if avg spending was below the floor, but not above...
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Figured it out (and fixed it on the main site).

So, I have code that tracks the inflation-adjusted spending level for each year, and a piece of code that effectively sets that number to $0 if it's not yet the retirement year (so it doesnt subtract expenses).  Turns out that this "Reset to $0" code was getting overruled at some point before it submitted information to my Stats functions.  So, while it wasn't deducting money from your portfolio in non-retirement years, the stats function was still counting it as spending.  So, you were getting 60 years worth of spending numbers, dividing by 54 years of retirement to find an "average" well over $90k.


shuffler

  • Pencil Stache
  • ****
  • Posts: 572
Figured it out (and fixed it on the main site).
Thanks for the fix.  The numbers look good.

FWIW, I like this spending model b/c I set my ceiling equal to my desired retirement income, and set my floor ~15% below that.  This gives me zero growth in my spending (which I like), but some downward flexibility for down markets.  I'm less interested in how much (more) I can spend, and more interested in how a degree of (downward) flexibility increases the success-rate.

In real life, I'm not sure I'd make these sorts of adjustments based on last year's market results though.  I'd probably try to evaluate the overall health of the portfolio.  Whether my yearly expenses are still <= my SWR, etc.  Ex:  if late in life I have 2x-3x what I need, but happen to hit a 10% down market ... I may not feel too compelled to reduce my spending by 5% (w/ z=0.5).

(BTW, a very minor cosmetic thing ... the Average Ending Portfolio is displayed without commas for thousands-separators.)

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator

Thanks for the fix.  The numbers look good.

FWIW, I like this spending model b/c I set my ceiling equal to my desired retirement income, and set my floor ~15% below that.  This gives me zero growth in my spending (which I like), but some downward flexibility for down markets.  I'm less interested in how much (more) I can spend, and more interested in how a degree of (downward) flexibility increases the success-rate.

In real life, I'm not sure I'd make these sorts of adjustments based on last year's market results though.  I'd probably try to evaluate the overall health of the portfolio.  Whether my yearly expenses are still <= my SWR, etc.  Ex:  if late in life I have 2x-3x what I need, but happen to hit a 10% down market ... I may not feel too compelled to reduce my spending by 5% (w/ z=0.5).

Yeah, I'm personally a big fan of that method of modeling as well. I feel like it's a pretty easy exercise to determine "comfortable" spending and "hard times" spending on a personal level.  I hear you on whether or not you'd exactly follow this in retirement or not.  Fact of the matter is, no matter what modeling you do, it'll be nigh impossible to follow that model exactly.  I mostly gain value from the knowledge that "if I do roughly this in retirement, I'm in good shape".  Even the straight-up 4% rule can't be exactly followed in retirement.  Who's spending the exact same amount of money from year to year for decades?

Quote
(BTW, a very minor cosmetic thing ... the Average Ending Portfolio is displayed without commas for thousands-separators.)

Nice catch... thanks.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Can you make it so clicking the (?) icon next to each section opens the FAQ in a new window?  As is, it loses all your inputs.  :)
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

lauren_knows

  • Pencil Stache
  • ****
  • Posts: 846
  • Age: 42
  • Location: Annandale, VA, USA
  • Happiness is a choice
    • The Crowdsourced FIRE simulator
Can you make it so clicking the (?) icon next to each section opens the FAQ in a new window?  As is, it loses all your inputs.  :)

Ha! I never thought about that implication.  I'll make it so (very easy fix).

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
I think your paypal link is broken.  It loads paypal fine, but trying to donate gives an error "An error occurred during processing. Please try again." - I don't think it's on my end, as I used paypal successfully very recently...
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

 

Wow, a phone plan for fifteen bucks!