Author Topic: Portfolio Design: Idiots v. Gurus  (Read 6070 times)

BicycleB

  • Handlebar Stache
  • *****
  • Posts: 1724
  • Location: Live Music Capital of the World
  • Older than the internet, but not wiser... yet
Re: Portfolio Design: Idiots v. Gurus
« Reply #100 on: October 17, 2019, 08:09:44 PM »
^ Good answer, thanks. Still pondering, but good answer.

Part of me instantly goes "Ah, crypto = baking soda". To me it tastes terrible and, by itself, seems likely to fall flat. But since crypto is highly variable, with possibly a tendency to gain value when ordinary currencies and investments tremble, it's conceivable that crypto could be part of a tasty cookie.

I mean, it scares me too much to actually do it. Just thinking out loud.

chevy1956

  • 5 O'Clock Shadow
  • *
  • Posts: 34
Re: Portfolio Design: Idiots v. Gurus
« Reply #101 on: October 18, 2019, 02:34:17 AM »
Is there a selection bias in being on portfoliocharts.com that makes us miss something important?

That's a fair question.  Without sidetracking this discussion, the short story is that one of my goals is to fight the data availability bias very common in investing analysis by offering a wide variety of modern asset options both at home and abroad.  But it's an ongoing project, and the current list of assets is definitely not all-inclusive of everything I think might be useful.  For example, I think TIPS are nice products and would love to include them on the site if I can find an accurate method to simulate their performance prior to their introduction in 1997. It's a difficult problem to solve, and I'm still researching models behind the scenes. So I encourage people to use the data to expand their thinking rather than artificially limit it.

How do we tell which assets really are flaky and useless, and should be excluded from even a very diverse portfolio?

Since there's no single portfolio suitable for all people, to a large extent I think the answer is personal.  If you really hate a specific asset like gold, then choose a portfolio that doesn't have it!  There are lots of good options.  My one overarching piece of advice, however, is to try to stop tasting each portfolio ingredient in isolation and instead think about its important contribution to the overall recipe.  Nobody thinks baking soda tastes good, but there's a reason it's in the cookies.

These are really good points. I also think @ChpBstrd has pointed out a really interesting point on asset allocation.  A lot of those portfolios that look good may under perform over the next 50+ years. Maybe gold will be reduced to an irrelevancy and cryptos will become the most common means of exchange.

Tyler

  • Handlebar Stache
  • *****
  • Posts: 1159
Re: Portfolio Design: Idiots v. Gurus
« Reply #102 on: October 18, 2019, 12:52:10 PM »
A lot of those portfolios that look good may under perform over the next 50+ years. Maybe gold will be reduced to an irrelevancy and cryptos will become the most common means of exchange.

Say what you want about gold, but it's been a valuable means of exchange for thousands of years longer than any other asset on the list.  So I seriously doubt it's going to suddenly vanish in our lifetimes.  ;) 

I hear ya, though.  I appreciate your perspective about the uncertainty of the best performers in the future, and I agree that nobody has a crystal ball.  Personally, that's one reason I value diversification!  Eggs, baskets, and all that.

For the purposes of this discussion, just keep in mind my previous point about how Radagast's SWR analysis accounts for consistency over all timeframes rather than simply high returns over one timeframe (also read his original "no bullshitting" explanation in Note 1). It's an important distinction that implies the order of the list is not as random as you may believe if you're only accustomed to thinking about maximizing returns. 

« Last Edit: October 18, 2019, 01:15:34 PM by Tyler »

BicycleB

  • Handlebar Stache
  • *****
  • Posts: 1724
  • Location: Live Music Capital of the World
  • Older than the internet, but not wiser... yet
Re: Portfolio Design: Idiots v. Gurus
« Reply #103 on: October 18, 2019, 05:05:48 PM »
Re-reading the OP, I finally notice something that's probably important:

"Note 3: Portfolio Charts has 9:10:4:3 HomeStocks:ForeignStocks:Bonds:"Real"Assets, so all naive portfolios follow that basic ratio, which might be significant."

Oh. Ah. Ding. Light bulb. (Man, how did I miss this?) So...all the naive portfolios are about 70-75% stock. Where they differ from a lot of expert portfolios is having more "real assets" (I guess real estate, gold and commodities) than some of the expert portfolios, and possibly having more diversification with the broad categories listed in the quote. They're kind of a consistent allocation of their own, seemingly.

An allocation that worked well, evidently. But... so does that lead us back to questioning whether a good past allocation is a good future one?

I suppose if we tested a "naive" approach with a significantly different weighting of broad categories, it could help us distinguish between the allocation-of-main-categories element and the diversification-within-main-categories effect.



Tyler

  • Handlebar Stache
  • *****
  • Posts: 1159
Re: Portfolio Design: Idiots v. Gurus
« Reply #104 on: October 18, 2019, 09:53:47 PM »
I suppose if we tested a "naive" approach with a significantly different weighting of broad categories, it could help us distinguish between the allocation-of-main-categories element and the diversification-within-main-categories effect.

Now you're talking!  That basic idea is one goal of the Portfolio Finder.  Try this:

https://portfoliocharts.com/portfolio/portfolio-finder/

It will help you model hundreds of naive portfolios simultaneously, and you can explore the effects of diversification breadth vs. depth by controlling the assets under consideration.
« Last Edit: October 19, 2019, 09:11:48 AM by Tyler »

chevy1956

  • 5 O'Clock Shadow
  • *
  • Posts: 34
Re: Portfolio Design: Idiots v. Gurus
« Reply #105 on: October 19, 2019, 03:33:48 AM »
A lot of those portfolios that look good may under perform over the next 50+ years. Maybe gold will be reduced to an irrelevancy and cryptos will become the most common means of exchange.

Say what you want about gold, but it's been a valuable means of exchange for thousands of years longer than any other asset on the list.  So I seriously doubt it's going to suddenly vanish in our lifetimes.  ;) 

I hear ya, though.  I appreciate your perspective about the uncertainty of the best performers in the future, and I agree that nobody has a crystal ball.  Personally, that's one reason I value diversification!  Eggs, baskets, and all that.

For the purposes of this discussion, just keep in mind my previous point about how Radagast's SWR analysis accounts for consistency over all timeframes rather than simply high returns over one timeframe (also read his original "no bullshitting" explanation in Note 1). It's an important distinction that implies the order of the list is not as random as you may believe if you're only accustomed to thinking about maximizing returns.

I understand where you are coming from in relation to portfolio design. I like diversification as well but maybe the message should be buy the most diversified low cost index funds across different sectors. So real estate, bonds, stocks and commodities. Then possibly pick your asset allocation among those four sectors. We can all come to different conclusions based on the data. I think recognizing the principles of portfolio design and WR's is really important to understand but there isn't one solution to fit all.

Your data and portfolio analysis is really good. Your site is unreal. Thank you for that.

Radagast

  • Handlebar Stache
  • *****
  • Posts: 1585
  • Location: West of the Mountains, East of the Sea
  • One does not simply work into Mordor
Re: Portfolio Design: Idiots v. Gurus
« Reply #106 on: October 19, 2019, 11:30:59 PM »
How do we tell which assets really are flaky and useless, and should be excluded from even a very diverse portfolio? Crypto seems like a bad "investment" to me, but... well, where is the line?  "The answer is stocks" seems too glib. Even "the ones Tyler wisely chose because of a bunch of data" just seems...well, probably good, but I feel like we're missing something. Even if the categories on portfoliocharts.com are the "right" ones historically, it seems likely to me that they probably include something (stocks?) that are ready for a permanent fall, and exclude something with a shorter history or other exclusion reason that will perform much better. Is diversifying amoung Things That Did Well Up To Now really our best option?

Sorry if I'm rambling. This part seems really confusing to me.
I was recently reading Taleb, who suggests that the time an idea or concept can be expected to endure in the future is proportional to the time it has already existed. Apparently it originated with Broadway shows, a show that played for a week would be expected (on average) to end in a week, one that lasted a year would continue another year. He extended that to ideas: Socrates has been around 2500 years and will probably be around another 2500, Machiavelli has been around six hundred and will probably be around another 600, Arendt has been around 70 and will still be read in another 7 years. Also possibly for species. The fact that something abstract has been around a long time indicates it has enduring value that will be relevant to the future.

Extrapolate those to investments. "Real" assets like land, gold, wheat, pork bellies would be expected to continue to be traded for thousands of years in the future because they have already been traded for that long. Petroleum "rock oil" has been big for maybe 150 years, and will likely mirror that and disappear in another 150 years. Bitcoin is the oldest and safest of cryptos and might be expected to last another decade, while a crypto introduced a year ago will probably fade by next year.

Of course that is a median expectation, not a perfect explanation. Obviously new things appear, while Cats left Broadway and lead plumbing is a terrible idea even though it was used for millennia. But it seems like a good starting point.

Radagast

  • Handlebar Stache
  • *****
  • Posts: 1585
  • Location: West of the Mountains, East of the Sea
  • One does not simply work into Mordor
Re: Portfolio Design: Idiots v. Gurus
« Reply #107 on: October 19, 2019, 11:48:46 PM »
Re-reading the OP, I finally notice something that's probably important:

"Note 3: Portfolio Charts has 9:10:4:3 HomeStocks:ForeignStocks:Bonds:"Real"Assets, so all naive portfolios follow that basic ratio, which might be significant."

Oh. Ah. Ding. Light bulb. (Man, how did I miss this?) So...all the naive portfolios are about 70-75% stock. Where they differ from a lot of expert portfolios is having more "real assets" (I guess real estate, gold and commodities) than some of the expert portfolios, and possibly having more diversification with the broad categories listed in the quote. They're kind of a consistent allocation of their own, seemingly.

An allocation that worked well, evidently. But... so does that lead us back to questioning whether a good past allocation is a good future one?

I suppose if we tested a "naive" approach with a significantly different weighting of broad categories, it could help us distinguish between the allocation-of-main-categories element and the diversification-within-main-categories effect.
Yeah I probably understated that. "Is very significant" might be more appropriate. Although one of the "Naive" portfolios is 90% stock 10% cash and I don't recall it standing out. I do think that the ratio in Portfolio Charts is a pretty good guess about the future though.

Another point is that I first thought of this idea a few years ago and even did a few tests using Portfolio Visualizer, which uses a totally different set of assets over a different time. It also has very different backtesting tools. PV has a lot more and different types of bonds and would have given more bonds and lower stocks and fewer "real" assets. So this was more of a test which supported the hypothesis.

kenmoremmm

  • Bristles
  • ***
  • Posts: 305
Re: Portfolio Design: Idiots v. Gurus
« Reply #108 on: October 20, 2019, 01:30:33 AM »
How do we tell which assets really are flaky and useless, and should be excluded from even a very diverse portfolio? Crypto seems like a bad "investment" to me, but... well, where is the line?  "The answer is stocks" seems too glib. Even "the ones Tyler wisely chose because of a bunch of data" just seems...well, probably good, but I feel like we're missing something. Even if the categories on portfoliocharts.com are the "right" ones historically, it seems likely to me that they probably include something (stocks?) that are ready for a permanent fall, and exclude something with a shorter history or other exclusion reason that will perform much better. Is diversifying amoung Things That Did Well Up To Now really our best option?

Sorry if I'm rambling. This part seems really confusing to me.
I was recently reading Taleb, who suggests that the time an idea or concept can be expected to endure in the future is proportional to the time it has already existed. Apparently it originated with Broadway shows, a show that played for a week would be expected (on average) to end in a week, one that lasted a year would continue another year. He extended that to ideas: Socrates has been around 2500 years and will probably be around another 2500, Machiavelli has been around six hundred and will probably be around another 600, Arendt has been around 70 and will still be read in another 7 years. Also possibly for species. The fact that something abstract has been around a long time indicates it has enduring value that will be relevant to the future.

Extrapolate those to investments. "Real" assets like land, gold, wheat, pork bellies would be expected to continue to be traded for thousands of years in the future because they have already been traded for that long. Petroleum "rock oil" has been big for maybe 150 years, and will likely mirror that and disappear in another 150 years. Bitcoin is the oldest and safest of cryptos and might be expected to last another decade, while a crypto introduced a year ago will probably fade by next year.

Of course that is a median expectation, not a perfect explanation. Obviously new things appear, while Cats left Broadway and lead plumbing is a terrible idea even though it was used for millennia. But it seems like a good starting point.

hmm. maybe it's late and i'm not comprehending, but this taleb concept makes no sense to me.

why is the expected duration based on the year in which i look at that commodity to estimate how much time it has left? if bitcoin has been around 10 years, and is expected to go kaput in 10 more, how do i logic this out when bitcoin has been around for 15 years? do i say it only has 5 more years to go, or another 15?

it seems like this is an infinity loop.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28058
  • Age: -999
  • Location: Seattle, WA
Re: Portfolio Design: Idiots v. Gurus
« Reply #109 on: October 20, 2019, 11:33:28 AM »
When it's been around 15 years its self-life is expected to be another 15.

However long it has lasted is a rough rule of thumb for how much longer it will last.
We are two former teachers who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and are now settled with two kids.
If you want to know more about us, or how we did that, or see lots of pictures, this Business Insider profile tells our story pretty well.
We (rarely) blog at AdventuringAlong.com. Check out our Now page to see what we're up to currently.

kenmoremmm

  • Bristles
  • ***
  • Posts: 305
Re: Portfolio Design: Idiots v. Gurus
« Reply #110 on: October 20, 2019, 11:55:01 AM »
sorry. that's a garbage loop.

if it's been around 1 days, then i expect tomorrow will be its last day???

but then, when it's day 2, now it'll be around 2 more days?

this makes no sense.

dragoncar

  • Walrus Stache
  • *******
  • Posts: 8798
  • Registered member
Re: Portfolio Design: Idiots v. Gurus
« Reply #111 on: October 20, 2019, 12:09:11 PM »
sorry. that's a garbage loop.

if it's been around 1 days, then i expect tomorrow will be its last day???

but then, when it's day 2, now it'll be around 2 more days?

this makes no sense.

Yeah Iím struggling to figure out how this rule minimizes prediction error.  Letís say a new play opens in year 1 and ends in year 10.  At the end of year 1, the rule predicts a two year run, error 8.  In year two, prediction is 4, error 6.  So each (year; prediction; error):

1; 2; 8
2; 4; 6
3; 6; 4
4; 8; 2
5; 10; 0
6; 12; 2
7; 14; 4
8; 16; 6
9; 18; 8
10; 20; 10

Average  error 5, so presumably average error is half of the life of the play. 

Even a naive ďthe play will last one more yearĒ rule would have an average error of 3.7.  Im sure there are even better rules
« Last Edit: October 20, 2019, 12:10:54 PM by dragoncar »

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28058
  • Age: -999
  • Location: Seattle, WA
Re: Portfolio Design: Idiots v. Gurus
« Reply #112 on: October 20, 2019, 06:44:01 PM »
I think it's more for older things, saying "Socrates has been around a long time, it's a good bet he'll still be around in 500+ years (2k+ according to his rule of thumb" and "this thing is very new, I shouldn't predict it should last forever."

I think it's mostly to remind you that the old stuff isn't going away, and the new stuff that looks great may not be around for as long as you think. Sort of a mental model to help ward against recency bias.
We are two former teachers who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and are now settled with two kids.
If you want to know more about us, or how we did that, or see lots of pictures, this Business Insider profile tells our story pretty well.
We (rarely) blog at AdventuringAlong.com. Check out our Now page to see what we're up to currently.

Radagast

  • Handlebar Stache
  • *****
  • Posts: 1585
  • Location: West of the Mountains, East of the Sea
  • One does not simply work into Mordor
Re: Portfolio Design: Idiots v. Gurus
« Reply #113 on: October 20, 2019, 10:48:42 PM »
Think of it as mean time to failure, not a definite time it will end. If it is relevant after ten years you estimate there is a 50/50 chance it will still be relevant relevant in another 10. Not it will end then, but that on average you would expect something that has endured that long to double itís longevity. Half would fail sooner and half later.

Or you could invert it and estimate the odds it will become worthless next year. If something started ten years ago you estimate there is a 1/10 chance it will be gone next year. If it has been around 5,000 years and is still widely used, you estimate there is just a 1/5000 chance it will suddenly become irrelevant next year.

mrmoonymartian

  • Stubble
  • **
  • Posts: 177
  • Age: 37
  • Location: Brisbane
Re: Portfolio Design: Idiots v. Gurus
« Reply #114 on: October 21, 2019, 02:09:38 AM »
Yeah Iím struggling to figure out how this rule minimizes prediction error.
Probably talking about different errors. I see that if you don't take the absolute value of the error and keep the vector, then the errors cancel out. Meaning for a particular event predicted regularly, the mean of all of predictions was in fact accurate. They just weren't particularly precise, as you demonstrated.

talltexan

  • Magnum Stache
  • ******
  • Posts: 2662
Re: Portfolio Design: Idiots v. Gurus
« Reply #115 on: October 21, 2019, 09:54:37 AM »
sorry. that's a garbage loop.

if it's been around 1 days, then i expect tomorrow will be its last day???

but then, when it's day 2, now it'll be around 2 more days?

this makes no sense.

Yesterday, I thought it was going to be dead at the end of today with 50% probability. Since it's clear that hasn't happened, I need to update my beliefs with the fact that it didn't die and recalculate.

ChpBstrd

  • Handlebar Stache
  • *****
  • Posts: 1566
Re: Portfolio Design: Idiots v. Gurus
« Reply #116 on: October 21, 2019, 12:00:56 PM »
The durability of cars and Thanksgiving turkeys and virtually anything physical doesnít work this way. IDK about ideas/memes though, such as the idea that gold/cryptocurrency has value or that one is supposed to eat turkey on Thanksgiving. Technology, culture, and industries are changing at a faster pace now than at any point in history. I would be wary of claims that people in 20-30 years will behave the same way people do now, or even want the same things (e.g. car ownership is starting to decline, people now spend more money and time on cell phones than TVs, veganism is taking off, stock trade commissions are free, and all this was not even sci-fi a few years ago.).

Radagast

  • Handlebar Stache
  • *****
  • Posts: 1585
  • Location: West of the Mountains, East of the Sea
  • One does not simply work into Mordor
Re: Portfolio Design: Idiots v. Gurus
« Reply #117 on: October 21, 2019, 09:38:51 PM »
sorry. that's a garbage loop.

if it's been around 1 days, then i expect tomorrow will be its last day???

but then, when it's day 2, now it'll be around 2 more days?

this makes no sense.

Yesterday, I thought it was going to be dead at the end of today with 50% probability. Since it's clear that hasn't happened, I need to update my beliefs with the fact that it didn't die and recalculate.
Right, you continually update your expectations of the unknown. In 2010 I expect bitcoin to be forgotten in a year, if anybody even knew of it. In 2011 I guess there is a 50/50 chance bitcoin will disappear by 2013. In 2013 I think there is a 50/50 chance it will be relevant in 2017. In 2017 it is still around so I estimate there is a 50% chance it will exist in 2025. Right now in 2019 we guess it has 50/50 of making it to 2029.

The durability of cars and Thanksgiving turkeys and virtually anything physical doesnít work this way. IDK about ideas/memes though, such as the idea that gold/cryptocurrency has value or that one is supposed to eat turkey on Thanksgiving. Technology, culture, and industries are changing at a faster pace now than at any point in history. I would be wary of claims that people in 20-30 years will behave the same way people do now, or even want the same things (e.g. car ownership is starting to decline, people now spend more money and time on cell phones than TVs, veganism is taking off, stock trade commissions are free, and all this was not even sci-fi a few years ago.).
Easy come, easy go.

It is supposed to apply to things without a defined expiration date. A mosquito, rat, or human would individually die on a fairly predictable schedule. Mosquitoes, rats, and humans can be expected to last another 200 million, 50 million, and 100,000 years.

Anyhow getting off topic.
« Last Edit: October 21, 2019, 09:44:14 PM by Radagast »