The Scientific Method

[PlayinwithFIRE]

In the career world I lived in...the scientific method was part and parcel to everything we did.

A wiki description of the process :

"The scientific method is a rigorous, iterative process used to acquire knowledge through observation, hypothesis testing, and experimentation. It is designed to minimize human bias and error, allowing researchers to develop reliable, cause-and-effect explanations of natural phenomena."

In general, you observe something , try to theorize why that thing happened, then design tests to determine if your theory holds water.

I see alot of observing and theorizing on here, but very little followup on whether the theory has repeatably held water.

Maybe this needs to go in the pet-peeve thread.


Arguably, the markets are hard to predict beyond broad swath but its always interesting to me that we all continue searching for theories and keep coming up
with "whys" ..even if they don't match last weeks "why" which we were so sure of

pwf

 

[KMPC]

Each retired person on this forum is testing their theory and observations in real time. We are the "study group ".

 

[planuntilthefisthits]

I sometimes do post-hoc analyses, e.g., comparing what would have happened had I chosen SPY vs QQQ 10 years ago.

QQQ vs. SPY — ETF Comparison Tool

Compare QQQ and SPY side by side, including charts, performance, statistics, average return, top holdings, and more.

stockanalysis.com

Then I know that I made the right decision.

 

[Fermion]

I sometimes do post-hoc analyses, e.g., comparing what would have happened had I chosen SPY vs QQQ 10 years ago.

QQQ vs. SPY — ETF Comparison Tool

Compare QQQ and SPY side by side, including charts, performance, statistics, average return, top holdings, and more.

stockanalysis.com

Then I know that I made the right decision.

I don't think the qualifications for the "right decision" should be the end gain, otherwise the right decision would have been neither SPY or QQQ but rather bitcoin.

 

[planuntilthefisthits]

I don't think the qualifications for the "right decision" should be the end gain, otherwise the right decision would have been neither SPY or QQQ but rather bitcoin.

But I am testing my hypothesis. You can test yours.

 

[NW-Bound]

I sometimes do post-hoc analyses, e.g., comparing what would have happened had I chosen SPY vs QQQ 10 years ago.

QQQ vs. SPY — ETF Comparison Tool

Compare QQQ and SPY side by side, including charts, performance, statistics, average return, top holdings, and more.

stockanalysis.com

Then I know that I made the right decision.

Look at the period of 2000-2010.

 

[planuntilthefisthits]

Look at the period of 2000-2010.

But I am testing my hypothesis. You can test yours.

 

[OldShooter]

... Arguably, the markets are hard to predict beyond broad swath but its always interesting to me that we all continue searching for theories and keep coming up
with "whys" ..even if they don't match last weeks "why" which we were so sure of

"Random" is hard to predict and we have a half-century or so of failed theories to prove it.

 

[winyaz]

I generally do test my applications first, but as in science, I build on the work of others. Mostly I use Trend Following and Momentum, but there are other well established edges. And to be clear these don't "predict" the market, but they can identify situations where the odds are in your favor. I view it as comparable to gambling, only being the house rather than the customer. No one knows where the ball will land on any given spin in roulette but over many spins the odds clearly favor the house. A comparable example in the stock market is that returns over the next month have been stronger when the market is above it's 10 month moving average, than when below it. There are a number of papers documenting this effect going back 100 years or more.

Some examples of relatively well-established effects:

• Momentum — assets that have outperformed recently tend to continue outperforming for a while. This is one of the strongest documented anomalies.
• Value — cheaper stocks (by earnings, book value, cash flow, etc.) have historically outperformed expensive ones over long horizons.
• Quality / profitability — companies with strong balance sheets and high profitability tend to do better risk-adjusted.
• Low volatility anomaly — lower-volatility stocks have often produced surprisingly competitive returns.
• Trend following — medium/long-term trends across asset classes have shown persistence for decades.
• Mean reversion — at longer horizons, extremely overextended valuations often revert. This is often true in the short term as well.

 

[COcheesehead]

I see some theories on here that hold water, but a bunch will tell you are wrong. So you sorta lose the motivation over time.

 

[PatrickW]

Scientific Method has largely been replaced by "Seeking Consensus".

Only half joking....

 

[Bdarr]

the best approach is the understand Ben Grahams analogy of Mr. Market…the markets are a creation of man and man is somewhat predictable, unless they are bi polar or fall into a mania…. What makes a person buy up the price of an investment? Sociology, psychology and the madness of the masses…following the herd is not as advantageous and predicting where the herd will go…or where the herd ignored and will see their error and herd there…again the best way is the buy a business well run with above 15% growth and high ROIC/ROE also above 15% and little to no debt < 2 yrs FCF over 5 yrs and preferably 10 yrs…especially in areas the herds don't gather but when the news come the stampede begins…lastly determine the business intrinsic value with say a DCF, and validate with other valuations, PE, PEG, PEGY, Multiples etc... and throw in a MOS of 30-50% to protect you from error and ya might be able to follow Buffetts two rules 1. Don't lose money, 2. Don't forget rule 1…it is simple, but not easy…you are often your own worst enemy…it is as much a psychology problem as it is a math (accounting) problem! Just my 2 cents

 

[sengsational]

Ben Graham's short term voting machine, long term weighing machine is where I land.

 

[jim584672]

"Past performance is not indicative of future results", yet the 4% rule is set in stone like it means something.

 

[TheWizard]

"Past performance is not indicative of future results", yet the 4% rule is set in stone like it means something.

Thing is, past performance sort of IS indicative of future results, over the long term and for various index funds, not individual stocks...

 

[jim584672]

Thing is, past performance sort of IS indicative of future results, over the long term and for various index funds, not individual stocks...

I would not make that assumption.

 

[Original Wally]

I keep thinking of the poor guy who, in the summer of 1929, said to wife, I think right after the first of the year we will sell it all, pay the taxes and retire. Oops.

 

[GrayHare]

Scientific Method has largely been replaced by "Seeking Consensus".

It's the Family Feud approach to research. "Name a US state that borders Mexico. The top 8 answers are on the board."

 

[TheWizard]

I would not make that assumption.

My assumption is based on broad stock market performance since I started investing around 1973.
I have no problem if others ignore that...

 

[jim584672]

My assumption is based on broad stock market performance since I started investing around 1973.
I have no problem if others ignore that...

"Keep in mind that the Fund’s past performance (before and after taxes) does not indicate how the Fund will perform in the future." - Source Vanguard VTI prospectus.

 

[OldShooter]

My assumption is based on broad stock market performance since I started investing around 1973.
I have no problem if others ignore that...

Thing is, past performance sort of IS indicative of future results, over the long term and for various index funds, not individual stocks...

Actually, research has shown this idea to be false. S&P publishes a report on what they call "Manager Persistence," where persistence refers to an active manager's performance persisting from period to period. They are pretty much all the same --- little or no persistence. Here is a slide I made a while back for my AdultEd investing class:

I'm sure a quick search will find other, probably more recent, data.​

 

[TheWizard]

Actually, research has shown this idea to be false. S&P publishes a report on what they call "Manager Persistence," where persistence refers to an active manager's performance persisting from period to period. They are pretty much all the same --- little or no persistence. Here is a slide I made a while back for my AdultEd investing class:

View attachment 64281
I'm sure a quick search will find other, probably more recent, data.​

Looks like you're confusing actively managed funds with index funds.
There's a difference...

 

[Diogenes]

I don't think the qualifications for the "right decision" should be the end gain, otherwise the right decision would have been neither SPY or QQQ but rather bitcoin.

Some investments have enough history and sheer mass behind them, that we have ample cause to view their progress as being sustainable. Some do not. An example of the former is residential real estate. It rises and occasionally falls, but in aggregate there's ample cause to believe that it will rise with inflation, ever so slightly exceeding it. But "greater fool" investments like Bitcoin have no plausible reckoning of where they'll go. In the language of the scientific method, our sample size is so small, and noise in our experiment so large, that there's not enough of a trend to support any hypothesis.

Thing is, that if we do have cause to say, within reason, that past performance IS some predictor of future returns, then an end-state-based metric would be sensible. And if not, not. Which brings us to....

"Keep in mind that the Fund’s past performance (before and after taxes) does not indicate how the Fund will perform in the future." - Source Vanguard VTI prospectus.

Strictly speaking this is always true. If we have a nuclear war tomorrow night, Monday's opener on the NYSE, if there even is one, will be disappointing.

We can't control for universal risk... meteor, plague and so on. Remember Covid-19? But we can control for idiosyncratic risk, and that's where Vanguard's advice is both anodyne pablum and sheer genius. It's pablum because sure, OK, don't count on the future until it's already arrived. It's genius because however Intel or Space-X or whatever does in the future, the global market-cap-weighted stock market will likely continue to be a redoubtable engine of growth... next year, next decade, next century and onwards. By removing idiosyncratic risk, we gain confidence that the future really will resemble the past.

 

[TheWizard]

...By removing idiosyncratic risk, we gain confidence that the future really will resemble the past...

Correct; that was my basic point.

If someone was numerically inclined, one could compute the Total Return of the S&P 500, in nominal dollars, over rolling five-year periods since I started in 1973. That would be periods from 1973-1977 (inclusive), 1974-1978, etc.
If someone did that, you would find a significant majority of five-year periods had a gain.

That's what I expect to continue for the next fifty years...

 

[OldShooter]

Looks like you're confusing actively managed funds with index funds.
There's a difference...

Yeah. I didn't understand your reference to index funds, since a broad index fund is the market. So you just mean that you expect the market history trend, upwards by a few percent a year, to continue? I can certainly buy that, less of course the small number of basis points in index fund fees.