"Lies, Damned Lies, and Statistics."

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
Taylor Larimore
Advisory Board
Posts: 29021
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

"Lies, Damned Lies, and Statistics."

Post by Taylor Larimore » Sun Mar 17, 2019 2:26 pm

There are three kinds of lies: Lies, damned lies, and statistics. -- Benjamin Disraeli
Bogleheads:

I know of no better examples of contradicting statistics than in this recent Boglehead topic:

I wrote using Morningstar statistics:
Style ................... 1 Year.....3 Years.....5 Years

US Market..............2.16%......13.7%......10.21%
US Large Cap ......... 2.46.......14.05.......10.83 (Best 5-year return)
US Mid Cap ............0.98.......11.56........ 8.66
US Small Cap ........ -0.92.......12.48 ........6.91
US Small Cap Value. -3.71.........8.81........5.25
vineviz replied using Portfolio Visualizer statistics:
Style..............................1 year......3 years......5 years

Total Stock Market..............7.74%.......8.07%.......7.98%.
S&P 500 Large Cap..............7.28%.......7.63%.......7.58%
S&P 400 Mid Cap................9.79%......10.01%.......9.77%
S&P 600 Small Cap............10.53%.......10.62%.....10.10%
S&P 600 Small Cap Value....10.21%.......10.20%.......9.58% (13.92% 1-year difference)
I am not sure which returns are the most realistic, but the above figures (over a similar period) offer a good illustration of the dangers of relying on academic statistics.

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle

User avatar
arcticpineapplecorp.
Posts: 4674
Joined: Tue Mar 06, 2012 9:22 pm

Re: "Lies, Damned Lies, and Statistics."

Post by arcticpineapplecorp. » Sun Mar 17, 2019 2:57 pm

this is why critical thinking matters. When I read your post Taylor, the first question I had was, "Are these two data sources using the same dates or not?" In other words, is one using 1, 3, 5 years ending 12/31/2018 and the other using 1,3,5 years...based on YTD? You'd likely get different results if they are using different ending dates.

Anyone know what is accounting for the differences?

Wait a second, the link only shows 1985-2019 (feb) so where is the portfolio visualizer link showing any 1, 3 or 5 year returns at all for these indices?

In addition portfolio visualizer is going through Feb 2019...I suspect morningstar is using yearly info, as in through 12/31/18 (though I could be wrong).

I think vineviz needs to show the actual 1, 3, 5 year returns with portfolio visualizer to dispute morningstars. the link vineviz provided only shows 1985-2019 (feb): https://www.portfoliovisualizer.com/bac ... ion4_3=100

so where did the 1, 3, 5 year portfoliovisualizer results come from vineviz?
Last edited by arcticpineapplecorp. on Sun Mar 17, 2019 3:03 pm, edited 1 time in total.
"May you live as long as you want and never want as long as you live" -- Irish Blessing | "Invest we must" -- Jack Bogle

CFM300
Posts: 1674
Joined: Sat Oct 27, 2007 5:13 am

Re: "Lies, Damned Lies, and Statistics."

Post by CFM300 » Sun Mar 17, 2019 2:59 pm

[removed because I was wrong. :D ]
Last edited by CFM300 on Sun Mar 17, 2019 7:50 pm, edited 3 times in total.

drk
Posts: 1427
Joined: Mon Jul 24, 2017 10:33 pm
Location: Seattle

Re: "Lies, Damned Lies, and Statistics."

Post by drk » Sun Mar 17, 2019 3:18 pm

As vineviz stated, he used actual funds' performance over the last X years and shared his source:
vineviz wrote:
Sat Mar 16, 2019 5:43 pm
In case anyone is interested in seeing a longer time period as reflected by actual index funds (iShares ETFs, except Vanguard's VTI), as opposed to category averages
Taylor's numbers are via Morningstar's indexes, which do not appear to represent the performance of any actual funds.

CFM300
Posts: 1674
Joined: Sat Oct 27, 2007 5:13 am

Re: "Lies, Damned Lies, and Statistics."

Post by CFM300 » Sun Mar 17, 2019 3:20 pm

[Removed because I was wrong. :D ]
Last edited by CFM300 on Sun Mar 17, 2019 3:36 pm, edited 1 time in total.

drk
Posts: 1427
Joined: Mon Jul 24, 2017 10:33 pm
Location: Seattle

Re: "Lies, Damned Lies, and Statistics."

Post by drk » Sun Mar 17, 2019 3:24 pm

CFM300 wrote:
Sun Mar 17, 2019 3:20 pm
drk wrote:
Sun Mar 17, 2019 3:18 pm
Taylor's numbers are via Morningstar's indexes, which do not appear to represent the performance of any actual funds.
Please correct me if I'm wrong, but Morningstar's numbers reflect the performance of FUNDS in the category, not an index of individual STOCKS in the category.
It's hard to get a straight answer out of Morningstar, but that does not appear to be case based on the glossary definition or the index portfolio. As far as I can tell, these are just indexes.

CFM300
Posts: 1674
Joined: Sat Oct 27, 2007 5:13 am

Re: "Lies, Damned Lies, and Statistics."

Post by CFM300 » Sun Mar 17, 2019 3:36 pm

I think you're right!

(I won't bother explaining how I was mislead, but I'm going to edit each of my previous posts. Feel free to edit or remove yours as well, if you want. Doesn't matter to me either way, but it might be less confusing to others.)

User avatar
Nate79
Posts: 5403
Joined: Thu Aug 11, 2016 6:24 pm
Location: Delaware

Re: "Lies, Damned Lies, and Statistics."

Post by Nate79 » Sun Mar 17, 2019 3:51 pm

According to Vanguard as of 2/28/2019 the return of VTSAX was 1, 3, 5, 10 years: 5.12%, 15.56%, 10.13%, 16.86%.

venkman
Posts: 1115
Joined: Tue Mar 14, 2017 10:33 pm

Re: "Lies, Damned Lies, and Statistics."

Post by venkman » Sun Mar 17, 2019 9:03 pm

Image

(ETF used for SCV because the Admiral shares MF only went back to 2011.)

sambb
Posts: 2726
Joined: Sun Mar 10, 2013 3:31 pm

Re: "Lies, Damned Lies, and Statistics."

Post by sambb » Sun Mar 17, 2019 9:13 pm

dont know why it matters.. past performance does not have any bearing on future results.

User avatar
Topic Author
Taylor Larimore
Advisory Board
Posts: 29021
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: "Lies, Damned Lies, and Statistics."

Post by Taylor Larimore » Sun Mar 17, 2019 9:15 pm

Bogleheads:

I have enjoyed reading your thoughtful Replies.

sambb is right. We must not forget this important rule:

Past performance does not forecast future performance.

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle

User avatar
Elric
Posts: 241
Joined: Sat Dec 08, 2018 12:23 am
Location: Virginia
Contact:

Re: "Lies, Damned Lies, and Statistics."

Post by Elric » Sun Mar 17, 2019 11:18 pm

sambb wrote:
Sun Mar 17, 2019 9:13 pm
dont know why it matters.. past performance does not have any bearing on future results.
I'd say "may not" or even "often." But no bearing? I'd say that's too strong. Any data we have is from what's happened in the past (much as I'd like data from the future). And when, for example, something outperforms on average across multiple multi-duration rolling averages over multiple market cycles and a shock or two, I think that unless you can identify a key factor that has changed, it a better than 50% chance that will continue awhile into the future. Now a one year outperformance, on the other hand, may be just alignment with ideal conditions or just random noise
"No man is free who works for a living." | Illya Kuryakin

retiringwhen
Posts: 1569
Joined: Sat Jul 08, 2017 10:09 am
Location: New Jersey, USA

Re: "Lies, Damned Lies, and Statistics."

Post by retiringwhen » Mon Mar 18, 2019 8:33 am

Elric wrote:
Sun Mar 17, 2019 11:18 pm
sambb wrote:
Sun Mar 17, 2019 9:13 pm
dont know why it matters.. past performance does not have any bearing on future results.
I'd say "may not" or even "often." But no bearing? I'd say that's too strong. Any data we have is from what's happened in the past (much as I'd like data from the future). And when, for example, something outperforms on average across multiple multi-duration rolling averages over multiple market cycles and a shock or two, I think that unless you can identify a key factor that has changed, it a better than 50% chance that will continue awhile into the future. Now a one year outperformance, on the other hand, may be just alignment with ideal conditions or just random noise
The underlying debate (beside the proper use and attribution of data and statistics) here is really is about the reality (or lack thereof) of the Value/Size Premium. The problem is that depending upon time-frame, country/market and the very definition of the Value/Size Factor(s) you get different answers from past-performance.

So what do you do? Making decisions based upon historical record is nearly guaranteed to be wrong. My observations is that 90% of all financial research and 99.9% of financial journalism is effectively less valuable than toilet paper because most of it is based on a misguided idea that you divine the future based upon the past. But it seldom if ever does seem to work out that way... In war and financial markets and investing, always be wary of fighting the last battle, focus on the next battle.

I will leave it to others to state what we can learn from the past and apply to the future (there are many good things we have learned, but they seldom if every come from CAGR return charts.)

Random Walker
Posts: 4289
Joined: Fri Feb 23, 2007 8:21 pm

Re: "Lies, Damned Lies, and Statistics."

Post by Random Walker » Mon Mar 18, 2019 9:49 am

In Larry Swedroe’s factor book he outlines 5 criteria for evaluating a factor. I believe those 5 criteria really apply to any potential source of return. The criteria are persistent, pervasive, robust, intuitive, investable. We can only invest looking forward, so for me intuitive has a special significance. It is the intuitive risk based or behavioral based explanations for a premium that I think allow us to stick to a plan after periods of poor past performance.

Dave

User avatar
willthrill81
Posts: 15891
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "Lies, Damned Lies, and Statistics."

Post by willthrill81 » Mon Mar 18, 2019 9:55 am

Taylor Larimore wrote:
Sun Mar 17, 2019 9:15 pm
Bogleheads:

I have enjoyed reading your thoughtful Replies.

sambb is right. We must not forget this important rule:

Past performance does not forecast future performance.

Best wishes.
Taylor
Does that mean that the past is irrelevant in forming our plans for the future?

How far do we take this logic?
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

Random Walker
Posts: 4289
Joined: Fri Feb 23, 2007 8:21 pm

Re: "Lies, Damned Lies, and Statistics."

Post by Random Walker » Mon Mar 18, 2019 10:00 am

willthrill81 wrote:
Mon Mar 18, 2019 9:55 am
Taylor Larimore wrote:
Sun Mar 17, 2019 9:15 pm
Bogleheads:

I have enjoyed reading your thoughtful Replies.

sambb is right. We must not forget this important rule:

Past performance does not forecast future performance.

Best wishes.
Taylor
Does that mean that the past is irrelevant in forming our plans for the future?

How far do we take this logic?
I would use the past data on persistence and pervasiveness to buttress one’s conviction to stick with a plan.

Dave

User avatar
siamond
Posts: 5189
Joined: Mon May 28, 2012 5:50 am

Re: "Lies, Damned Lies, and Statistics."

Post by siamond » Mon Mar 18, 2019 10:07 am

drk wrote:
Sun Mar 17, 2019 3:24 pm
CFM300 wrote:
Sun Mar 17, 2019 3:20 pm
drk wrote:
Sun Mar 17, 2019 3:18 pm
Taylor's numbers are via Morningstar's indexes, which do not appear to represent the performance of any actual funds.
Please correct me if I'm wrong, but Morningstar's numbers reflect the performance of FUNDS in the category, not an index of individual STOCKS in the category.
It's hard to get a straight answer out of Morningstar, but that does not appear to be case based on the glossary definition or the index portfolio. As far as I can tell, these are just indexes.
Morningstar is a wonderful resource for raw data, but whenever they derive some metric via computational analysis, they have a knack for producing utterly misleading numbers, and this includes their so-called indices, their 3x3 matrices, their rating system, etc. If one gets curious about their 'index' (?!) methodology, it is described here.

Why does Taylor keep quoting such Morningstar stats instead of reputable index providers (or corresponding Vanguard funds), I have no idea.

DaufuskieNate
Posts: 387
Joined: Wed May 28, 2014 11:53 am

Re: "Lies, Damned Lies, and Statistics."

Post by DaufuskieNate » Mon Mar 18, 2019 10:11 am

siamond wrote:
Mon Mar 18, 2019 10:07 am
drk wrote:
Sun Mar 17, 2019 3:24 pm
CFM300 wrote:
Sun Mar 17, 2019 3:20 pm
drk wrote:
Sun Mar 17, 2019 3:18 pm
Taylor's numbers are via Morningstar's indexes, which do not appear to represent the performance of any actual funds.
Please correct me if I'm wrong, but Morningstar's numbers reflect the performance of FUNDS in the category, not an index of individual STOCKS in the category.
It's hard to get a straight answer out of Morningstar, but that does not appear to be case based on the glossary definition or the index portfolio. As far as I can tell, these are just indexes.
Morningstar is a wonderful resource for raw data, but whenever they derive some metric via computational analysis, they have a knack for producing utterly misleading numbers, and this includes their so-called indices, their 3x3 matrices, their rating system, etc. If one gets curious about their 'index' (?!) methodology, it is described here.

Why does Taylor keep quoting such Morningstar stats instead of reputable index providers (or corresponding Vanguard funds), I have no idea.
Confirmation bias...

User avatar
willthrill81
Posts: 15891
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "Lies, Damned Lies, and Statistics."

Post by willthrill81 » Mon Mar 18, 2019 10:17 am

siamond wrote:
Mon Mar 18, 2019 10:07 am
Why does Taylor keep quoting such Morningstar stats instead of reputable index providers (or corresponding Vanguard funds), I have no idea.
I still cannot figure out why he has often repeated the "past performance does not predict the future" mantra but then proceeded to compare the track record of the 3-fund portfolio to a variety of other portfolios.

To be honest, I don't believe that I've ever met or seen a single investor who truly believes the "past performance" idea. It's just a question of how far people take it. If it weren't for past performance, how on earth would we even begin to determine an AA, for instance?
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

retiringwhen
Posts: 1569
Joined: Sat Jul 08, 2017 10:09 am
Location: New Jersey, USA

Re: "Lies, Damned Lies, and Statistics."

Post by retiringwhen » Mon Mar 18, 2019 11:03 am

Random Walker wrote:
Mon Mar 18, 2019 9:49 am
In Larry Swedroe’s factor book he outlines 5 criteria for evaluating a factor. I believe those 5 criteria really apply to any potential source of return. The criteria are persistent, pervasive, robust, intuitive, investable. We can only invest looking forward, so for me intuitive has a special significance. It is the intuitive risk based or behavioral based explanations for a premium that I think allow us to stick to a plan after periods of poor past performance.

Dave
I will say, I feel very differently about intuitive. I have found that many things in life and especially in investing, intuition is mostly wrong. For example, most peope are trained to believe (and have observed in many other spheres of life) that a smart, motivated person should be able to beat the market. But the hard, repeatable, results are that being smart and motivated are not sufficient to achieve a market beating result, Luck appears to be the deciding factor....

On the flip side, if you have long-term performance data it is very seducing to tell a story that feels like it is right (intuition) to justify it. That is probably as dangerous as straight trusting... I think this board and passive investors in general understand the first problem, but are often not so clear on the form fitting risk of the second.

Random Walker
Posts: 4289
Joined: Fri Feb 23, 2007 8:21 pm

Re: "Lies, Damned Lies, and Statistics."

Post by Random Walker » Mon Mar 18, 2019 11:18 am

retiringwhen wrote:
Mon Mar 18, 2019 11:03 am
Random Walker wrote:
Mon Mar 18, 2019 9:49 am
In Larry Swedroe’s factor book he outlines 5 criteria for evaluating a factor. I believe those 5 criteria really apply to any potential source of return. The criteria are persistent, pervasive, robust, intuitive, investable. We can only invest looking forward, so for me intuitive has a special significance. It is the intuitive risk based or behavioral based explanations for a premium that I think allow us to stick to a plan after periods of poor past performance.

Dave
I will say, I feel very differently about intuitive. I have found that many things in life and especially in investing, intuition is mostly wrong. For example, most peope are trained to believe (and have observed in many other spheres of life) that a smart, motivated person should be able to beat the market. But the hard, repeatable, results are that being smart and motivated are not sufficient to achieve a market beating result, Luck appears to be the deciding factor....

On the flip side, if you have long-term performance data it is very seducing to tell a story that feels like it is right (intuition) to justify it. That is probably as dangerous as straight trusting... I think this board and passive investors in general understand the first problem, but are often not so clear on the form fitting risk of the second.
I can agree mostly with that. Firstly, intuitive sounds more wish washy than it actually is. It refers to sound risk based and behavioral based explanations. That is certainly subject to potential “form fitting”, but that being said, if it all makes sense then it all makes sense. Also, this is where the other 4 criteria come into play. If a source of return is persistent over time and economic regimes, that adds strength to the explanation. If the source of return is pervasive across markets, geographies, asset classes, that further buttresses the intuitive.

Dave

2015
Posts: 2906
Joined: Mon Feb 10, 2014 2:32 pm

Re: "Lies, Damned Lies, and Statistics."

Post by 2015 » Mon Mar 18, 2019 12:56 pm

retiringwhen wrote:
Mon Mar 18, 2019 11:03 am
Random Walker wrote:
Mon Mar 18, 2019 9:49 am
In Larry Swedroe’s factor book he outlines 5 criteria for evaluating a factor. I believe those 5 criteria really apply to any potential source of return. The criteria are persistent, pervasive, robust, intuitive, investable. We can only invest looking forward, so for me intuitive has a special significance. It is the intuitive risk based or behavioral based explanations for a premium that I think allow us to stick to a plan after periods of poor past performance.

Dave
I will say, I feel very differently about intuitive. I have found that many things in life and especially in investing, intuition is mostly wrong. For example, most peope are trained to believe (and have observed in many other spheres of life) that a smart, motivated person should be able to beat the market. But the hard, repeatable, results are that being smart and motivated are not sufficient to achieve a market beating result, Luck appears to be the deciding factor....

On the flip side, if you have long-term performance data it is very seducing to tell a story that feels like it is right (intuition) to justify it. That is probably as dangerous as straight trusting... I think this board and passive investors in general understand the first problem, but are often not so clear on the form fitting risk of the second.
In the end, what is and always has been "robust, pervasive, intuitive, etc., etc." has always been and will always be based on narrative and humans loves them some good stories, historically speaking. Some on this board may "know" past performance does not predict the future but they sure don't act like it. Whole lotta witchcraft, voodoo, and Law of Attraction-type stuff going on. And you are quite correct, luck is always the deciding factor ultimately. Applies to both good and bad luck. It is only when the up market tsunami goes out that we get mooned by the overconfident bending over desperately to pull up their yanked down suits.

OTOH, if everyone stopped paying attention to this tripe financial writers would not be able to pay their mortgages because no one would be monetizing their "work" (if it may be called that).
Last edited by 2015 on Mon Mar 18, 2019 1:07 pm, edited 1 time in total.

2015
Posts: 2906
Joined: Mon Feb 10, 2014 2:32 pm

Re: "Lies, Damned Lies, and Statistics."

Post by 2015 » Mon Mar 18, 2019 1:06 pm

Random Walker wrote:
Mon Mar 18, 2019 11:18 am
retiringwhen wrote:
Mon Mar 18, 2019 11:03 am
Random Walker wrote:
Mon Mar 18, 2019 9:49 am
In Larry Swedroe’s factor book he outlines 5 criteria for evaluating a factor. I believe those 5 criteria really apply to any potential source of return. The criteria are persistent, pervasive, robust, intuitive, investable. We can only invest looking forward, so for me intuitive has a special significance. It is the intuitive risk based or behavioral based explanations for a premium that I think allow us to stick to a plan after periods of poor past performance.

Dave
I will say, I feel very differently about intuitive. I have found that many things in life and especially in investing, intuition is mostly wrong. For example, most peope are trained to believe (and have observed in many other spheres of life) that a smart, motivated person should be able to beat the market. But the hard, repeatable, results are that being smart and motivated are not sufficient to achieve a market beating result, Luck appears to be the deciding factor....

On the flip side, if you have long-term performance data it is very seducing to tell a story that feels like it is right (intuition) to justify it. That is probably as dangerous as straight trusting... I think this board and passive investors in general understand the first problem, but are often not so clear on the form fitting risk of the second.
I can agree mostly with that. Firstly, intuitive sounds more wish washy than it actually is. It refers to sound risk based and behavioral based explanations. That is certainly subject to potential “form fitting”, but that being said, if it all makes sense then it all makes sense. Also, this is where the other 4 criteria come into play. If a source of return is persistent over time and economic regimes, that adds strength to the explanation. If the source of return is pervasive across markets, geographies, asset classes, that further buttresses the intuitive.

Dave

User avatar
siamond
Posts: 5189
Joined: Mon May 28, 2012 5:50 am

Re: "Lies, Damned Lies, and Statistics."

Post by siamond » Mon Mar 18, 2019 3:07 pm

willthrill81 wrote:
Mon Mar 18, 2019 10:17 am
To be honest, I don't believe that I've ever met or seen a single investor who truly believes the "past performance" idea. It's just a question of how far people take it. If it weren't for past performance, how on earth would we even begin to determine an AA, for instance?
This 'past performance' bit has been taken completely out of context and generalized way too hastily. I don't have the pointer handy, but somebody posted a link last year explaining the genesis of this official warning. The point was to warn individual investors to not trust the usual 1-year, 3-years, 5-years, 10-years stats of active funds as predictive of the near future. Active, short-term past, short-term future. In such specific context, this makes complete sense, no question.

Generalizing it to 'nobody knows nothing', passive funds and any arbitrary timeframe is plainly non-sensical. One has to look at past data with care and healthy skepticism (and pick proper metrics, e.g. NOT the Morningstar ones!), but there is a lot to learn from history. As the saying goes, "you ignore the lessons of the past at your own peril".

User avatar
Topic Author
Taylor Larimore
Advisory Board
Posts: 29021
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Past Performance

Post by Taylor Larimore » Mon Mar 18, 2019 6:42 pm

willthrill81 wrote: I don't believe that I've ever met or seen a single investor who truly believes the "past performance" idea.
willthrill81:

You may not have met them but they are out there:
American Association of Individual Investors: "Top Performance lists are dangerous."

Frank Armstrong, financial author: "Rating services such as Morningstar's 'Star Awards' or the 'Forbes Honor Roll' attest to the futility of applying past performance to tomorrow."

Arnott and Bernstein (2002, p. 64): “The investment management industry thrives on the expedient of forecasting the future by extrapolating the past."

Barra Research: "There is no persistence of equity fund performance."

Christine Benz, Morningstar Director of Personal Finance: "When we look at our data, at the factors that are most predictive of good performance going forward, low costs are a much better predictor than is great past performance."

Wm. Bernstein, author of The Four Pillars of Investing: "For the 20 years from 1970 to 1989, the best performing stock assets were Japanese stocks, U.S. small stocks, and gold stocks. These turned out to be the worst performing assets over the next decade."

Jack Bogle: "The biggest mistake investors make is looking backward at performance and thinking it’ll recur in the future."

Bogleheads' Guide to Investing: "Using past performance to pick tomorrow's winning mutual funds is such a bad idea that the government requires a statement similar to this: "Past performance is no guarantee of future performance." Believe it!"

Jack Brennan, former Vanguard CEO: "Fund ranking is meaningless when based primarily on past performance, as most are."

Burns Advisory tracked the performance of Morningstar's five-star rated stock funds beginning January 1, 1999. Of the 248 stock funds, just four still kept that rank after ten years.

Ben Carlson, author of A Wealth of Common Sense : "Dow Jones looked at nearly 2,900 active mutual funds. Only 2 funds in the top quartile stayed in the top quartile of performance over the next four 1-year periods."

Andrew Clarke, author: "By the time an investment reaches the top of the performance tables, there's a good chance that its run is over. The past is not prologue."

Jonathan Clements, author & former Wall Street Journal columnist: "Suppose you picked stock funds that ranked in their category's top 25% over the past five years. A regular updated study suggests that less than a quarter of these funds will remain in the top 25% over the next five years--even worse than the result you would expect based purely on chance."

Prof. John Cochrane, author: "Past performance has almost no information about future performance."

S.T.Coleridge: "History is a lantern over the stern. It shows where you've been but not where you're going"

Dow Jones Indices Report, June 2015: "The data shows a stronger likelihood for the best-performing funds to become the worst performing funds than vice versa." -- June 2016: "Only 3.7% of large-cap funds maintained top-half performance over five-consecutive 12-month periods. For midcap funds, the comparable figure was 5.79%, and for small-cap funds, it was 7.82%."

Charles D. Ellis, author of 16 financial books: "Sadly, investors who rely on performance records are relying on useless data."

Eugene Fama, Nobel Laureate: "Our research on individual mutual funds says that it's impossible to identify true winners on a reliable basis, even if one ignores the costs that active funds impose on investors."

Forbes (2/2/04 issue): "Over the past decade, Morningstar's five-star equity funds have earned an average 5.7% against a 10.3% return for the Wilshire 5000 (Total Stock Market)."

Gensler & Bear, co-authors of The Great Mutual Fund Trap: "Of the fifty top-performing funds in 2000, not a single one appeared on the list in either 1999 or 1998."

Ken Hebner's CGM Focus Fund was the top U.S. equity fund in 2007. In November 2009, it ranked in the bottom 1% of its category.

Mark Hulbert (12-31-2014): "Consider a hypothetical portfolio that each year followed the investment newsletter portfolio that, among the more than 500 tracked by The Hulbert Financial Digest, had the best record during the previous calendar year. Over the past 20 years, that portfolio would have been a disaster, producing an annualized loss of more than -17%."

Mark Hebner, President, Index Fund Advisors: "From 1998 through 2013 only about 8 funds remained in the top 100 the following year."

JPMorgan Chase claimed that 97% of their alternate-asset mutual funds beat their benchmark during the 10-year period ending December, 2013. Morningstar reported that only 33% beat their benchmark during the same period (past-performance calculations differ).

Arthur Levitt, SEC Commissioner: "A mutual fund's past performance, which is the first feature that investors consider when choosing a fund, doesn't predict future performance."

Peter Lynch's Fidelity Magellan Fund (FMAGX), once the world's largest and most successful mutual fund, is now (Feb. 9, 2018) in the bottom 11% of its category for 15-year annualized return

Burton Malkiel, author of the classic Random Walk Down Wall Street: "I have examined the lack of persistency in fund returns over periods from the 1960s through the early 2000s.--There is no persistency to good performance. It is as random as the market."

Mercer Investment Consulting from a study of over 12,000 institutional managers: "Excellent recent performance not only doesn't guarantee future results but generally leads to under-performance in the subsequent period."

Bill Miller, former manager of Legg Mason Value Trust (LMVTX), was the only manager to outperform The S&P 500 Index for 15 consecutive years. On 9/7/2016 Miller’s fund is in the bottom 1% for 15 year returns.

Mark Miller, financial author and journalist: "Only 7.33% of domestic equity funds that were in the top quartile of performance in March 2014 were still there two years later."

Morningstar: "Over the long term, there is no meaningful relationship between past and future fund performance."

Ron Ross, author of The Unbeatable Market: "Extensive studies by Davis, Brown & Groetzman, Ibbotson, Elton et al, all confirmed there is no significant persistence in mutual fund performance. -- Wall Street’s favorite scam is pretending that luck is skill.”

Bill Schultheis, adviser and author of The Coffeehouse Investor: "Using past performance numbers as a method for choosing mutual funds is such a lousy idea that mutual fund companies are required by law to tell you it is a lousy idea."

Sequoia Fund was the top performing large-cap growth fund at the end of 2015 according to Morningstar. On 4/21/2017 it ranked in the bottom 1% for five year returns.

Standard & Poor's Persistence Scorecard (Dec-2014): "The data show a stronger likelihood for the best-performing funds to become the worst-performing funds than vice versa. Of 421 funds that were in the bottom quartile, 14.45% moved to the top quartile over the five year horizon, while 27.08% of the 421 funds that were in the top quartile moved into the bottom quartile during the same period."

Larry Swedroe, author of many finance books: "The 44 Wall Street Fund was the top performing fund over the decade of the 1970s. It ranked as the single worst performing fund of the 1980's losing 73%. -- If you are going to use past performance to predict the future winners, the evidence is strong that your approach is highly likely to fail."

David Swensen, Yale's Chief Investment Officer: "Chasing performance is the biggest mistake investors make. If anything, it is a perverse indicator."

Tweddell & Pierce, co-authors of Winning With Mutual Funds: "Numerous studies have shown that using superior past performance is no better than random selection."

Eric Tyson, author of Mutual Funds for Dummies (2010 edition): "Of the number one top-performing stock and bond funds in each of the last 20 years, a whopping 80% of them subsequently performed worse than the average fund in their peer group over the next 5 to 10 years! Some of these former #1 funds actually went on to become the worst-performing funds in their particular category."

Value Line selected Garret Van Wagoner "Mutual fund Manager of the Year" in 1999. In August 2009, Van Wagoner's Emerging Growth Fund was the worst performing U.S. stock fund over the past 10 years.

Vanguard Study: "Persistence of performance among past winners is no more predictable than a flip of a coin."

Jason Zweig, author and Wall Street Journal columnist: "Buying funds based purely on their past performance is one of the stupidest things an investor can do."
Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle

User avatar
willthrill81
Posts: 15891
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Past Performance

Post by willthrill81 » Mon Mar 18, 2019 7:05 pm

Taylor Larimore wrote:
Mon Mar 18, 2019 6:42 pm
willthrill81 wrote: I don't believe that I've ever met or seen a single investor who truly believes the "past performance" idea.
willthrill81:

You may not have met them but they are out there:
Taylor, if we should ignore past performance, then how should we determine our AA?

If we should ignore past performance, why have you repeatedly displayed the past performance of the 3-fund portfolio compared to other portfolios?
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

User avatar
aspirit
Posts: 478
Joined: Wed Jan 03, 2018 12:52 am
Location: SMA.SFL

Re: "Lies, Damned Lies, and Statistics."

Post by aspirit » Mon Mar 18, 2019 7:26 pm

I certainly do not want to speak for Anyone. However this thread suggests to me to always question statistics, from any and all sources! :mrgreen:
Time & tides wait for no one. A man has to know his limitations. | "Give me control of a nation's money and I care not who makes it's laws" | — Mayer Amschel Bauer Rothschild ~

2015
Posts: 2906
Joined: Mon Feb 10, 2014 2:32 pm

Re: "Lies, Damned Lies, and Statistics."

Post by 2015 » Mon Mar 18, 2019 7:30 pm

siamond wrote:
Mon Mar 18, 2019 3:07 pm
willthrill81 wrote:
Mon Mar 18, 2019 10:17 am
To be honest, I don't believe that I've ever met or seen a single investor who truly believes the "past performance" idea. It's just a question of how far people take it. If it weren't for past performance, how on earth would we even begin to determine an AA, for instance?
This 'past performance' bit has been taken completely out of context and generalized way too hastily. I don't have the pointer handy, but somebody posted a link last year explaining the genesis of this official warning. The point was to warn individual investors to not trust the usual 1-year, 3-years, 5-years, 10-years stats of active funds as predictive of the near future. Active, short-term past, short-term future. In such specific context, this makes complete sense, no question.

Generalizing it to 'nobody knows nothing', passive funds and any arbitrary timeframe is plainly non-sensical. One has to look at past data with care and healthy skepticism (and pick proper metrics, e.g. NOT the Morningstar ones!), but there is a lot to learn from history. As the saying goes, "you ignore the lessons of the past at your own peril".
I've seen nothing that demonstrates consistent success with any form of active investing over time for the average, no-nothing investor. (that's just about everybody). Of course you have to use some kind of past behavior if you're gonna step into the investing casino (e.g., the stock market goes up over time), but to look at "past data" and extrapolate anything using "proper metrics" smacks of oh here we go again with overconfident active Boglehead speak. Yea, there's a whole lot to learn from history, like the only certainty is there is no certainty, how foolish investors can be and how readily they can fool themselves with their pattern making. This happens repeatedly throughout history, and is probably the only constant when it comes to investing. My prediction is the next kick-in-the-crotch to investors will be something they could never have thought of.

It's all narrative to me as it all exists in a complex adaptive system and in such systems risk mitigation is superior to wondering if the gods mad because it's thundering. I ain't bettin' the house on anything like that.

User avatar
willthrill81
Posts: 15891
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "Lies, Damned Lies, and Statistics."

Post by willthrill81 » Mon Mar 18, 2019 8:33 pm

2015 wrote:
Mon Mar 18, 2019 7:30 pm
siamond wrote:
Mon Mar 18, 2019 3:07 pm
willthrill81 wrote:
Mon Mar 18, 2019 10:17 am
To be honest, I don't believe that I've ever met or seen a single investor who truly believes the "past performance" idea. It's just a question of how far people take it. If it weren't for past performance, how on earth would we even begin to determine an AA, for instance?
This 'past performance' bit has been taken completely out of context and generalized way too hastily. I don't have the pointer handy, but somebody posted a link last year explaining the genesis of this official warning. The point was to warn individual investors to not trust the usual 1-year, 3-years, 5-years, 10-years stats of active funds as predictive of the near future. Active, short-term past, short-term future. In such specific context, this makes complete sense, no question.

Generalizing it to 'nobody knows nothing', passive funds and any arbitrary timeframe is plainly non-sensical. One has to look at past data with care and healthy skepticism (and pick proper metrics, e.g. NOT the Morningstar ones!), but there is a lot to learn from history. As the saying goes, "you ignore the lessons of the past at your own peril".
I've seen nothing that demonstrates consistent success with any form of active investing over time for the average, no-nothing investor. (that's just about everybody). Of course you have to use some kind of past behavior if you're gonna step into the investing casino (e.g., the stock market goes up over time), but to look at "past data" and extrapolate anything using "proper metrics" smacks of oh here we go again with overconfident active Boglehead speak. Yea, there's a whole lot to learn from history, like the only certainty is there is no certainty, how foolish investors can be and how readily they can fool themselves with their pattern making. This happens repeatedly throughout history, and is probably the only constant when it comes to investing. My prediction is the next kick-in-the-crotch to investors will be something they could never have thought of.

It's all narrative to me as it all exists in a complex adaptive system and in such systems risk mitigation is superior to wondering if the gods mad because it's thundering. I ain't bettin' the house on anything like that.
I reiterate the question I posed to Taylor: if we take the "past performance is not indicative of future performance" argument at face value, how do we determine an appropriate asset allocation?
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

User avatar
Dialectical Investor
Posts: 527
Joined: Mon Dec 03, 2018 11:41 pm

Re: "Lies, Damned Lies, and Statistics."

Post by Dialectical Investor » Mon Mar 18, 2019 9:23 pm

willthrill81 wrote:
Mon Mar 18, 2019 8:33 pm

I reiterate the question I posed to Taylor: if we take the "past performance is not indicative of future performance" argument at face value, how do we determine an appropriate asset allocation?
You're not supposed to take it at face value, which of course would be absurd. You're supposed to realize the phrase is invoked when someone is displeased with an assertion, and disregarded when otherwise convenient, much like the phrase "lies, damned lies, and statistics." Such displeasure may result when someone isn't able or doesn't care to understand something, perhaps because they are disinterested or lack the cognitive abilities or have legitimate memories of being deceived through deliberate use of such methods. Upon realizing this state of affairs, there's nothing left to do but move along, or ask the question you have asked, and proceed to stand around holding your rain coat in the downpour, refusing to put it on.

Wakefield1
Posts: 1059
Joined: Mon Nov 14, 2016 10:10 pm

Re: "Lies, Damned Lies, and Statistics."

Post by Wakefield1 » Mon Mar 18, 2019 9:35 pm

I have a vague recollection that an authority held that there were some absolutely bottom of the barrel mutual funds that do tend to stay near the bottom (persistent performance) but that a top (recent) performance fund is just as likely to move down to average or even below average during the next period that performance is measured as would be likely for a middle of the pack fund (which might move up towards the top) in other words the bottom 10% showing more persistence than the rest (90%?)
the impression is that an investor could jump around between,say,5 different funds chasing performance,always ending up late to the party,and ending up with results worse than if the investor had stayed with any one of the 5 funds for the duration

I believe that Mr. Bogle held that over time expensive funds tend to trail less expensive funds in return
actually I know that

User avatar
DG99999
Posts: 203
Joined: Fri May 30, 2014 12:30 pm

Re: Past Performance

Post by DG99999 » Tue Mar 19, 2019 8:52 am

willthrill81 wrote:
Mon Mar 18, 2019 7:05 pm
Taylor Larimore wrote:
Mon Mar 18, 2019 6:42 pm
willthrill81 wrote: I don't believe that I've ever met or seen a single investor who truly believes the "past performance" idea.
willthrill81:

You may not have met them but they are out there:
Taylor, if we should ignore past performance, then how should we determine our AA?

If we should ignore past performance, why have you repeatedly displayed the past performance of the 3-fund portfolio compared to other portfolios?
WT8/Taylor

Notwithstandng any of the data and assertions above; my interpretation is that the warnings/cautions about past performance apply to securities, mutual funds and ETFs(etc) that represent some subset of their broad market. I choose to make the assumption that I can use the historical performance of broad markets to set my asset allocation (i.e. not just their average yearly return though!!). Using broad market index performance is not perfect, but may be the best data available to those of us as average investors. I am not defining broad markets, but feel the concept is fairly intuitive and I defer to the logic of Potter Stewart.

Note that using this simplistic thinking - I am even cautious of index funds that represent only a partial segment of broad markets. Also, it could validate Taylor's reference to the returns of his three funds since I would characterize these as approximating the returns of the broad markets they represent (here I am commenting on the data, not necessarily supporting or attacking the argument).
I am not a financial professional. My posts are only my opinion on the topic. You need to do your own due diligence and consult with a professional when addressing your financial questions.

2015
Posts: 2906
Joined: Mon Feb 10, 2014 2:32 pm

Re: "Lies, Damned Lies, and Statistics."

Post by 2015 » Tue Mar 19, 2019 11:56 am

willthrill81 wrote:
Mon Mar 18, 2019 8:33 pm
2015 wrote:
Mon Mar 18, 2019 7:30 pm
siamond wrote:
Mon Mar 18, 2019 3:07 pm
willthrill81 wrote:
Mon Mar 18, 2019 10:17 am
To be honest, I don't believe that I've ever met or seen a single investor who truly believes the "past performance" idea. It's just a question of how far people take it. If it weren't for past performance, how on earth would we even begin to determine an AA, for instance?
This 'past performance' bit has been taken completely out of context and generalized way too hastily. I don't have the pointer handy, but somebody posted a link last year explaining the genesis of this official warning. The point was to warn individual investors to not trust the usual 1-year, 3-years, 5-years, 10-years stats of active funds as predictive of the near future. Active, short-term past, short-term future. In such specific context, this makes complete sense, no question.

Generalizing it to 'nobody knows nothing', passive funds and any arbitrary timeframe is plainly non-sensical. One has to look at past data with care and healthy skepticism (and pick proper metrics, e.g. NOT the Morningstar ones!), but there is a lot to learn from history. As the saying goes, "you ignore the lessons of the past at your own peril".
I've seen nothing that demonstrates consistent success with any form of active investing over time for the average, no-nothing investor. (that's just about everybody). Of course you have to use some kind of past behavior if you're gonna step into the investing casino (e.g., the stock market goes up over time), but to look at "past data" and extrapolate anything using "proper metrics" smacks of oh here we go again with overconfident active Boglehead speak. Yea, there's a whole lot to learn from history, like the only certainty is there is no certainty, how foolish investors can be and how readily they can fool themselves with their pattern making. This happens repeatedly throughout history, and is probably the only constant when it comes to investing. My prediction is the next kick-in-the-crotch to investors will be something they could never have thought of.

It's all narrative to me as it all exists in a complex adaptive system and in such systems risk mitigation is superior to wondering if the gods mad because it's thundering. I ain't bettin' the house on anything like that.
I reiterate the question I posed to Taylor: if we take the "past performance is not indicative of future performance" argument at face value, how do we determine an appropriate asset allocation?
Personally, I didn't use past performance. I don't believe there are any guarantees related to the market. My personal prediction is the next market kick-in-the-crotch will be something that "wasn't supposed to happen" that will knock the hot air out of even the most experienced economists, "experts", academics, "most respected", etc. Why? Because this is how reality has always worked. It really is true. Nobody knows nothin', particularly when it comes to the nature of what reality in any complex system can and will do.

I decided my 50/50 AA based a number of factors:

1) The totality of the system my life has operated in and is operating in at this time. This includes my past, present, and preferred future;
2) Knowledge of my historical psychological and emotional reaction to risk and disruptive situations of all kinds;
3) My choice to use liability matching and a risky portfolio; and
4) My acknowledgement that in the act of even stepping just inside the doors of the casino and sitting down at the 3 Fund Portfolio Table (let alone going deeper into the casino with its cigar smoked back rooms of tilting, valuations, factoring and forever fumbling), I am still gambling.

I believe success can be ours when we focus on building competencies within ourselves beyond investing, personal finance, and microeconomics. Far more valuable than precision in investing is mastering risk analysis and mitigation, creating the ability to be agile, to pivot, to transform, to build effective decision making capabilities, as well as the ability to create options, resilience, persistence, self-mastery, self-discipline and self-control.

User avatar
siamond
Posts: 5189
Joined: Mon May 28, 2012 5:50 am

Re: "Lies, Damned Lies, and Statistics."

Post by siamond » Tue Mar 19, 2019 12:04 pm

2015 wrote:
Mon Mar 18, 2019 7:30 pm
I've seen nothing that demonstrates consistent success with any form of active investing over time for the average, no-nothing investor. (that's just about everybody). Of course you have to use some kind of past behavior if you're gonna step into the investing casino (e.g., the stock market goes up over time), but to look at "past data" and extrapolate anything using "proper metrics" smacks of oh here we go again with overconfident active Boglehead speak. [...]
Huh? Where in my post did I speak of successful active investing? I suspect you misconstrued my use of the word 'metric' as some kind of valuation metric driving an active/timing process of sorts. I didn't mean anything of that sort.

As the poster above hinted at, this is about gathering solid and long-term historical data about broad markets and making long-term decisions accordingly. Such decisions might be inclusive (use a given asset class in your AA) or exclusive (decide to NOT use a given asset class), but this is a fundamental decision-making process where history is our guide. Sure enough, we have to take such history with a significant grain of salt, then we all add our own layers of personal circumstances, beliefs and biases, and we end up in different places, but this is all about informed passive investing. As to the 'nobody knows nothing' folks claiming they made AA/retirement decisions without any such knowledge, they are just agitating a screen of smoke (or worse, solely acting based on their own biases).

Back to the OP, what really annoys me is the direct inference that ALL sorts of statistics and historical data are misleading and dangerous. This is a VERY fundamental misunderstanding of what the sarcastic "Lies, Damned Lies, and Statistics" wording means. On the contrary, this wording (and corresponding books) is about encouraging critical thinking, identify and avoid bad statistics (e.g. Morningstar ratings, 'indices', short-term historical stats, and so on) and about encouraging the use of more carefully chosen data in the right context.

User avatar
willthrill81
Posts: 15891
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "Lies, Damned Lies, and Statistics."

Post by willthrill81 » Tue Mar 19, 2019 12:08 pm

2015 wrote:
Tue Mar 19, 2019 11:56 am
willthrill81 wrote:
Mon Mar 18, 2019 8:33 pm
2015 wrote:
Mon Mar 18, 2019 7:30 pm
siamond wrote:
Mon Mar 18, 2019 3:07 pm
willthrill81 wrote:
Mon Mar 18, 2019 10:17 am
To be honest, I don't believe that I've ever met or seen a single investor who truly believes the "past performance" idea. It's just a question of how far people take it. If it weren't for past performance, how on earth would we even begin to determine an AA, for instance?
This 'past performance' bit has been taken completely out of context and generalized way too hastily. I don't have the pointer handy, but somebody posted a link last year explaining the genesis of this official warning. The point was to warn individual investors to not trust the usual 1-year, 3-years, 5-years, 10-years stats of active funds as predictive of the near future. Active, short-term past, short-term future. In such specific context, this makes complete sense, no question.

Generalizing it to 'nobody knows nothing', passive funds and any arbitrary timeframe is plainly non-sensical. One has to look at past data with care and healthy skepticism (and pick proper metrics, e.g. NOT the Morningstar ones!), but there is a lot to learn from history. As the saying goes, "you ignore the lessons of the past at your own peril".
I've seen nothing that demonstrates consistent success with any form of active investing over time for the average, no-nothing investor. (that's just about everybody). Of course you have to use some kind of past behavior if you're gonna step into the investing casino (e.g., the stock market goes up over time), but to look at "past data" and extrapolate anything using "proper metrics" smacks of oh here we go again with overconfident active Boglehead speak. Yea, there's a whole lot to learn from history, like the only certainty is there is no certainty, how foolish investors can be and how readily they can fool themselves with their pattern making. This happens repeatedly throughout history, and is probably the only constant when it comes to investing. My prediction is the next kick-in-the-crotch to investors will be something they could never have thought of.

It's all narrative to me as it all exists in a complex adaptive system and in such systems risk mitigation is superior to wondering if the gods mad because it's thundering. I ain't bettin' the house on anything like that.
I reiterate the question I posed to Taylor: if we take the "past performance is not indicative of future performance" argument at face value, how do we determine an appropriate asset allocation?
Personally, I didn't use past performance. I don't believe there are any guarantees related to the market. My personal prediction is the next market kick-in-the-crotch will be something that "wasn't supposed to happen" that will knock the hot air out of even the most experienced economists, "experts", academics, "most respected", etc. Why? Because this is how reality has always worked. It really is true. Nobody knows nothin', particularly when it comes to the nature of what reality in any complex system can and will do.
And you formed that belief on the basis of...past performance.

I completely acknowledge that the "past performance is not indicative of future performance" argument, as it was originally put forth in the context of comparing mutual funds' 1, 3, and 5 year performance and selecting the one with the highest performance, appears to have a lot of validity.

But that's a very far cry from the wholesale ignoring of history (i.e. past performance) in developing an AA, an investment strategy, a withdrawal strategy, etc. We cannot even begin to assess risk without examining history.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

User avatar
steve roy
Posts: 1742
Joined: Thu May 13, 2010 5:16 pm

Re: "Lies, Damned Lies, and Statistics."

Post by steve roy » Tue Mar 19, 2019 12:14 pm

Lots of people take past performance into account. If Large Cap index funds have a P/E of 20, and a P/B of 4, those numbers derive from past performance, and analysts estimate future earnings based on those numbers.

So let's stipulate that you can't accurately predict future results, but you can make an informed guess.

User avatar
Elric
Posts: 241
Joined: Sat Dec 08, 2018 12:23 am
Location: Virginia
Contact:

Re: "Lies, Damned Lies, and Statistics."

Post by Elric » Tue Mar 19, 2019 12:15 pm

2015 wrote:
Tue Mar 19, 2019 11:56 am
4) My acknowledgement that in the act of even stepping just inside the doors of the casino and sitting down at the 3 Fund Portfolio Table (let alone going deeper into the casino with its cigar smoked back rooms of tilting, valuations, factoring and forever fumbling), I am still gambling.
Only thing I'd comment on is that a key difference is that in the casino, the odds are always with the house, by design, and if you keep playing, your expected value is negative. I'd argue this is NOT true for many investment strategies, including the 3 Fund Portfolio. It may be gambling in that there's randomness and risk, but unlike any casino in the world, the odds are actually with you.
"No man is free who works for a living." | Illya Kuryakin

User avatar
willthrill81
Posts: 15891
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "Lies, Damned Lies, and Statistics."

Post by willthrill81 » Tue Mar 19, 2019 12:19 pm

Elric wrote:
Tue Mar 19, 2019 12:15 pm
2015 wrote:
Tue Mar 19, 2019 11:56 am
4) My acknowledgement that in the act of even stepping just inside the doors of the casino and sitting down at the 3 Fund Portfolio Table (let alone going deeper into the casino with its cigar smoked back rooms of tilting, valuations, factoring and forever fumbling), I am still gambling.
Only thing I'd comment on is that a key difference is that in the casino, the odds are always with the house, by design, and if you keep playing, your expected value is negative. I'd argue this is NOT true for many investment strategies, including the 3 Fund Portfolio. It may be gambling in that there's randomness and risk, but unlike any casino in the world, the odds are actually with you.
Indeed. This is one of the reasons I can't stand the Bernstein quote that strongly insinuates that the stock market is roughly akin to Las Vegas.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

acegolfer
Posts: 1655
Joined: Tue Aug 25, 2009 9:40 am

Re: "Lies, Damned Lies, and Statistics."

Post by acegolfer » Tue Mar 19, 2019 12:28 pm

This thread confirms that our statistics education is broken. We need to teach high school students so that they can design/solve problems rather than learn how to calculate stats.

Apparently, many BHs have taken stats before, but don't know why, when, how to properly use stats.

retiringwhen
Posts: 1569
Joined: Sat Jul 08, 2017 10:09 am
Location: New Jersey, USA

Re: "Lies, Damned Lies, and Statistics."

Post by retiringwhen » Tue Mar 19, 2019 12:40 pm

willthrill81 wrote:
Tue Mar 19, 2019 12:19 pm
Indeed. This is one of the reasons I can't stand the Bernstein quote that strongly insinuates that the stock market is roughly akin to Las Vegas.
I would argue it is worse than Vegas. In Vegas, you can download a document that describes the legally agreed to odds for each game and the expected return and be assured in normal circumstances that your odds are exactly that.

Financial researchers have been trying to do that for nearly 100 years without success. In fact, the data is so time dependent that one of two things is true: either it is truly a random walk (see discussion of momentum in coin-flips.... :wink: ) or the system is itself dynamic and therefore the odds change over time.

If it is the first, then collecting long-term data may actually allow us to form an understanding of the expected returns, but remember we only have about 4% of economic/market history documented to any level of precision and that assumes the 19th century data is valid. I am not assuming the data before that it valuable beyond anecdote. We frankly don't have enough data to do justice the random walk yet. Sorry siamond, the wonderful simba spreadsheet is only a tiny set of data points in time.

If it is the second, then maybe in 100 years the financial engineers and researchers will have identified enough of the variables in the system to build a reasonably predictive model to understand what changes in the inputs actually mean on the outputs of the system. The fact that we are still at a 3, 4 or 5 factor model as being the gold standard, I believe we are still in the infancy of that effort. Also, since we can't effective predict the inputs, the model may not still get us anywhere useful, sigh.

So what to do? Mostly invest based upon faith. Faith that the market will give a return on investing capital to create, improve and produce things of value to society and the economy, but that nothing is assured either in the short or long run.

Regardless, it is simply not worthwhile arguing about 1, 2, 3, maybe even 20 year returns between subtle shades of an asset class (e.g., Equities). We can't reasonably discern the source of the differences with any truly predictive certainty.

protagonist
Posts: 6117
Joined: Sun Dec 26, 2010 12:47 pm

Re: "Lies, Damned Lies, and Statistics."

Post by protagonist » Tue Mar 19, 2019 12:52 pm

Taylor Larimore wrote:
Sun Mar 17, 2019 2:26 pm
There are three kinds of lies: Lies, damned lies, and statistics. -- Benjamin Disraeli
I agree with the problem of believing everything one reads, Taylor.
But I would humbly replace Disraeli's quote with "Statistics don't lie. Statisticians lie." I have no clue who first said that.
The moral of the story being: If something you read matters enough to you, make sure you fact check and critique methodology very carefully, rather than just acquiescing to laziness or confirmation bias.
Last edited by protagonist on Tue Mar 19, 2019 1:42 pm, edited 2 times in total.

acegolfer
Posts: 1655
Joined: Tue Aug 25, 2009 9:40 am

Re: "Lies, Damned Lies, and Statistics."

Post by acegolfer » Tue Mar 19, 2019 12:57 pm

retiringwhen wrote:
Tue Mar 19, 2019 12:40 pm
willthrill81 wrote:
Tue Mar 19, 2019 12:19 pm
Indeed. This is one of the reasons I can't stand the Bernstein quote that strongly insinuates that the stock market is roughly akin to Las Vegas.
I would argue it is worse than Vegas. In Vegas, you can download a document that describes the legally agreed to odds for each game and the expected return and be assured in normal circumstances that your odds are exactly that.

Financial researchers have been trying to do that for nearly 100 years without success. In fact, the data is so time dependent that one of two things is true: either it is truly a random walk (see discussion of momentum in coin-flips.... :wink: ) or the system is itself dynamic and therefore the odds change over time.

If it is the first, then collecting long-term data may actually allow us to form an understanding of the expected returns, but remember we only have about 4% of economic/market history documented to any level of precision and that assumes the 19th century data is valid. I am not assuming the data before that it valuable beyond anecdote. We frankly don't have enough data to do justice the random walk yet. Sorry siamond, the wonderful simba spreadsheet is only a tiny set of data points in time.

If it is the second, then maybe in 100 years the financial engineers and researchers will have identified enough of the variables in the system to build a reasonably predictive model to understand what changes in the inputs actually mean on the outputs of the system. The fact that we are still at a 3, 4 or 5 factor model as being the gold standard, I believe we are still in the infancy of that effort. Also, since we can't effective predict the inputs, the model may not still get us anywhere useful, sigh.

So what to do? Mostly invest based upon faith. Faith that the market will give a return on investing capital to create, improve and produce things of value to society and the economy, but that nothing is assured either in the short or long run.

Regardless, it is simply not worthwhile arguing about 1, 2, 3, maybe even 20 year returns between subtle shades of an asset class (e.g., Equities). We can't reasonably discern the source of the differences with any truly predictive certainty.
Excellent post. I'll provide my answer to the above specific question. My investment is based on an economic theory, which is derived from a set of simplified assumptions. One may argue that the assumptions are incorrect. True, but if this simplified theory does a good job of explaining the market as a whole, then it's better than some faith.

protagonist
Posts: 6117
Joined: Sun Dec 26, 2010 12:47 pm

Re: "Lies, Damned Lies, and Statistics."

Post by protagonist » Tue Mar 19, 2019 1:24 pm

willthrill81 wrote:
Tue Mar 19, 2019 12:19 pm
Elric wrote:
Tue Mar 19, 2019 12:15 pm
2015 wrote:
Tue Mar 19, 2019 11:56 am
4) My acknowledgement that in the act of even stepping just inside the doors of the casino and sitting down at the 3 Fund Portfolio Table (let alone going deeper into the casino with its cigar smoked back rooms of tilting, valuations, factoring and forever fumbling), I am still gambling.
Only thing I'd comment on is that a key difference is that in the casino, the odds are always with the house, by design, and if you keep playing, your expected value is negative. I'd argue this is NOT true for many investment strategies, including the 3 Fund Portfolio. It may be gambling in that there's randomness and risk, but unlike any casino in the world, the odds are actually with you.
Indeed. This is one of the reasons I can't stand the Bernstein quote that strongly insinuates that the stock market is roughly akin to Las Vegas.
How do we know that "the odds are with us"? By using 90 or so years of past historical data, during what may have represented the greatest century of economic growth in the history of civilization, to predict the outcome of the next thirty or more?

Even if past performance does influence future returns (we have no good evidence that it does or doesn't), that is the statistical equivalent of saying that if the stock market went up at the annualized rate of 8% on every day from Monday through Wednesday, that it will likely go up on Thursday. Or if it didn't rain during the last 3 days, then it won't rain today.

Sounds like faith to me. I'm not arguing against faith when there is no real logic to prove another system better. I have a considerable portion of my assets in the stock market, but I honestly have no idea what will happen to it between now and 2050 (or 2100 or 2200 or ?????). In my case, I don't have faith either- I admit to just guessing. We have to do something with our money. But don't try to call it science.
Last edited by protagonist on Tue Mar 19, 2019 1:43 pm, edited 2 times in total.

User avatar
willthrill81
Posts: 15891
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "Lies, Damned Lies, and Statistics."

Post by willthrill81 » Tue Mar 19, 2019 1:27 pm

protagonist wrote:
Tue Mar 19, 2019 1:24 pm
willthrill81 wrote:
Tue Mar 19, 2019 12:19 pm
Elric wrote:
Tue Mar 19, 2019 12:15 pm
2015 wrote:
Tue Mar 19, 2019 11:56 am
4) My acknowledgement that in the act of even stepping just inside the doors of the casino and sitting down at the 3 Fund Portfolio Table (let alone going deeper into the casino with its cigar smoked back rooms of tilting, valuations, factoring and forever fumbling), I am still gambling.
Only thing I'd comment on is that a key difference is that in the casino, the odds are always with the house, by design, and if you keep playing, your expected value is negative. I'd argue this is NOT true for many investment strategies, including the 3 Fund Portfolio. It may be gambling in that there's randomness and risk, but unlike any casino in the world, the odds are actually with you.
Indeed. This is one of the reasons I can't stand the Bernstein quote that strongly insinuates that the stock market is roughly akin to Las Vegas.
How do we know that "the odds are with us"? By using 90 or so years of past historical data, during what may have represented the greatest century of economic growth in the history of civilization, to predict the outcome of the next thirty or more?

Even if past performance does influence future returns (we have no good evidence that it does or doesn't), that is the statistical equivalent of saying that if the stock market went up at the annualized rate of 8% on every day from Monday through Wednesday, that it will likely go up on Thursday. Or if it didn't rain during the last 3 days, then it won't rain today.

Sounds like faith to me. I'm not arguing against faith when there is no real logic to prove another system better. We have to do something with our money. But don't try to call it science.
Please point out when I did. If you cannot, then please do not put words in my mouth.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

protagonist
Posts: 6117
Joined: Sun Dec 26, 2010 12:47 pm

Re: "Lies, Damned Lies, and Statistics."

Post by protagonist » Tue Mar 19, 2019 1:30 pm

willthrill81 wrote:
Tue Mar 19, 2019 1:27 pm
protagonist wrote:
Tue Mar 19, 2019 1:24 pm
willthrill81 wrote:
Tue Mar 19, 2019 12:19 pm
Elric wrote:
Tue Mar 19, 2019 12:15 pm
2015 wrote:
Tue Mar 19, 2019 11:56 am
4) My acknowledgement that in the act of even stepping just inside the doors of the casino and sitting down at the 3 Fund Portfolio Table (let alone going deeper into the casino with its cigar smoked back rooms of tilting, valuations, factoring and forever fumbling), I am still gambling.
Only thing I'd comment on is that a key difference is that in the casino, the odds are always with the house, by design, and if you keep playing, your expected value is negative. I'd argue this is NOT true for many investment strategies, including the 3 Fund Portfolio. It may be gambling in that there's randomness and risk, but unlike any casino in the world, the odds are actually with you.
Indeed. This is one of the reasons I can't stand the Bernstein quote that strongly insinuates that the stock market is roughly akin to Las Vegas.
How do we know that "the odds are with us"? By using 90 or so years of past historical data, during what may have represented the greatest century of economic growth in the history of civilization, to predict the outcome of the next thirty or more?

Even if past performance does influence future returns (we have no good evidence that it does or doesn't), that is the statistical equivalent of saying that if the stock market went up at the annualized rate of 8% on every day from Monday through Wednesday, that it will likely go up on Thursday. Or if it didn't rain during the last 3 days, then it won't rain today.

Sounds like faith to me. I'm not arguing against faith when there is no real logic to prove another system better. We have to do something with our money. But don't try to call it science.
Please point out when I did. If you cannot, then please do not put words in my mouth.
You didn't, willthrill (smile). That comment was not directed towards you....it was a general comment to many who seem to think there is something scientifically and/or statistically valid about using past market performance to predict the future.

I was responding to the comment (I think Elric's) that "the odds are actually with you". I have no idea what the odds are. (Which makes it safer than Vegas, though much less predictable).

I agree with 2015 (I believe that is who it was) when he stated " Nobody knows nothin', particularly when it comes to the nature of what reality in any complex system can and will do."
Last edited by protagonist on Tue Mar 19, 2019 3:13 pm, edited 2 times in total.

User avatar
siamond
Posts: 5189
Joined: Mon May 28, 2012 5:50 am

Re: "Lies, Damned Lies, and Statistics."

Post by siamond » Tue Mar 19, 2019 1:48 pm

protagonist wrote:
Tue Mar 19, 2019 1:30 pm
Actually, I completely agree with you, willthrill....particularly when you stated " Nobody knows nothin', particularly when it comes to the nature of what reality in any complex system can and will do."
Oh boy. You did it again. Two data points, WE HAVE A TREND!! :D

It was actually 2015, who issued this self-contradicting statement...
2015 wrote:
Tue Mar 19, 2019 11:56 am
Why? Because this is how reality has always worked. It really is true. Nobody knows nothin', particularly when it comes to the nature of what reality in any complex system can and will do.

protagonist
Posts: 6117
Joined: Sun Dec 26, 2010 12:47 pm

Re: "Lies, Damned Lies, and Statistics."

Post by protagonist » Tue Mar 19, 2019 3:13 pm

siamond wrote:
Tue Mar 19, 2019 1:48 pm
protagonist wrote:
Tue Mar 19, 2019 1:30 pm
Actually, I completely agree with you, willthrill....particularly when you stated " Nobody knows nothin', particularly when it comes to the nature of what reality in any complex system can and will do."
Oh boy. You did it again. Two data points, WE HAVE A TREND!! :D

It was actually 2015, who issued this self-contradicting statement...
2015 wrote:
Tue Mar 19, 2019 11:56 am
Why? Because this is how reality has always worked. It really is true. Nobody knows nothin', particularly when it comes to the nature of what reality in any complex system can and will do.
Corrected. Sleepless night. Sorry.....

protagonist
Posts: 6117
Joined: Sun Dec 26, 2010 12:47 pm

Re: "Lies, Damned Lies, and Statistics."

Post by protagonist » Tue Mar 19, 2019 3:15 pm

deleted duplicate post
Last edited by protagonist on Wed Mar 20, 2019 9:48 am, edited 1 time in total.

User avatar
Elric
Posts: 241
Joined: Sat Dec 08, 2018 12:23 am
Location: Virginia
Contact:

Re: "Lies, Damned Lies, and Statistics."

Post by Elric » Tue Mar 19, 2019 3:43 pm

protagonist wrote:
Tue Mar 19, 2019 1:30 pm
I have no idea what the odds are. (Which makes it safer than Vegas, though much less predictable).
I'd think that would depend on your prior, if using Bayesian analysis. If you start with a prior assumption that both have 50/50 odds, due to lack of any knowledge, then yes, learning that the odds are against you in Vegas makes the market a better bet.
"No man is free who works for a living." | Illya Kuryakin

User avatar
steve roy
Posts: 1742
Joined: Thu May 13, 2010 5:16 pm

Re: "Lies, Damned Lies, and Statistics."

Post by steve roy » Tue Mar 19, 2019 4:02 pm

You can never know with any precision what millions of people/investors will do with their money at any given moment in time.

Locked