Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.
I momentarily hesitated before posting Housel’s “Good Enough” article.
The cause for the pause was the subtitle of the article: “Why rules of thumb beat precision.” That observation is true in many instances, but it is not a universal truism. Applying Rules of Thumb in an uncritical, rote manner can get you into deep water.
I reflected that pause when I wrote that “The trick is to understand the limits on these simplification rules.” In my haste to post, I failed to include two famous quotes that do a terrific job at characterizing my reservations.
The most famous is controversially attributed to Albert Einstein, but others are also credited: “Everything should be made as simple as possible, but no simpler”. I also like one from H. L. Mencken: “For every complex problem there is an answer that is clear, simple, and wrong”.
I believe these cautionary warnings. Oversimplification can lead down a wrong pathway.
It is often difficult to decide when enough is indeed enough. In the investment decision universe, experience is an important factor to inform that decision. Opinions from an independent and diverse group (like the MFO Board members) can also guide to a better “enough” decision.
I also winced and maybe winked a little when Housel observed that “Investing is often taught as if it's something like aeronautical engineering. It's filled with precise equations that give exact answers…”.
That comment pushed the discussion directly into my ballpark. I spent a considerable portion of my working life in a wind tunnel doing high speed aero testing. The purpose of those grueling, comprehensive tests was to collect sufficient real world data to permit approximate solutions of the governing aerodynamic equations.
The evaluation of those equations permitted performance forecasting and design refinements. Outsiders (like Housel) don’t appreciate the approximations embedded in those aerodynamic solutions and their limitations.
It’s several orders of magnitude more approximate, and consequently more unreliable, in the investment world given the modeling shortfalls. The introduction of emotionally motivated market participants adds yet another complexity dimension that is not modeled. Good luck on making reliable predictions under these circumstances.
"It’s several orders of magnitude more approximate, and consequently more unreliable, in the investment world given the modeling shortfalls. The introduction of emotionally motivated market participants adds yet another complexity dimension that is not modeled. Good luck on making reliable predictions under these circumstances."
@MJG: I've made similar observations in response to a number of your previous posts, where I felt that undue weight was being given to the alleged precision of such financial modeling. I came very close to making such a response again, but decided that it would likely only start another contentious exchange. I'm a bit surprised to see that we are in agreement on this, at least. The linguistic construction of your comment is particularly succinct and clear: I certainly would not attempt to improve on it.
Folks who do not commonly make sophisticated calculations (non-linear differential equations) overly trust the precision of the numbers computed. This is especially the case if the calculations include some unfamiliar modeling elements.
Scientists, engineers, and perhaps even mathematicians (attempt at humor) know better. They recognize the simplifications, the many assumptions, and the incomplete data that are often inputs to arrive at an answer. These needed shortcuts often significantly compromise accuracy. That’s why safety factors are introduced.
All models in all fields are simplifications of reality. An infinite number of digits after a decimal point is never taken seriously by an experienced engineer. It is merely an artifact of the calculation itself. He appreciates the accuracy limits of what he knows and rounds-off accordingly. I apply that same discipline in many of my posts.
Unfortunately, we frequently don’t recognize our modeling limitations in the investment world. We fall victim to false myths and deficient models; we overreact to very loose, meaningless correlations. Investment outcomes suffer.
This talk about modeling deficiencies and data inaccuracy prompted a memory of a famous WWII mistake told to engineering students as a lesson. It’s a terrific tale about the first US bomber that was recovered nearly intact by the Germans after being shot-down in an early bombing raid.
The German engineers puzzled over the profile of the wing configuration. It seemed to be inefficiently contoured. Wind Tunnel tests verified that assessment. The Germans could not accept that US designers had blundered so badly.
They postulated that the US designers knew some aerodynamic nuances that escaped them. They committed research resources to discover the suspected subtleties. The research failed. There was nothing to discover. The US designers had simply made an error in interpreting their own wind tunnel data. Mistakes on a grand scale do happen.
Many such mistakes are well documented. The scientific community is not immune to them. That’s why in many research disciplines, a duplication of original results from an independent test source is a mandatory standard before product approval and dissemination.
Given the uncertainties of market forecasting and modeling deficiencies, the decision to dump any precision requirements is easily made when doing market analyses.
"All models in all fields are simplifications of reality. An infinite number of digits after a decimal point is never taken seriously by an experienced engineer. It is merely an artifact of the calculation itself. He appreciates the accuracy limits of what he knows and rounds-off accordingly."
... same applies to these economic forecasts. Is a given report 3.8 or 3.9 or 4.25 percent up or down? WHO CARES (but economists and sell-side pundits)? Give me a good enough rule of thumb that can help me determine a longer-term trend and make my own decision.
I laugh when pundits on TV fret about single-digit-to-the-right-of-the-decimal-point as HUGE areas of contention or concern Puh-lease.
@ MJG, @Old Joe: Greatly enjoyed both of your comments, as well as the article. Reminds me of my own "education" as experimentalist >> modeller.
After a 40-year career in which the first half involved looking at the experimental results of one hard object suddenly encountering another and the second half trying to model these events in collaboration with some of the smartest people I have ever been privileged to work with using the most powerful supercomputers of the moment, one event that always returns to my mind occurred during an informal break at a project review when the young administrative assistant to our lab chief very earnestly asked my boss "Do you modelers ever get the right answer when you don't already know it in advance????"
Compared with economists, financial analysts, climatologists, etc., we at least could design & and perform experiments -- controlled and reproducible, more or less -- and physically and otherwise examine the results thereof. As for economic and financial theories and forecasts, I would be inclined to put them one step from witchcraft and mirrors. Enjoy, but handle with appropriate caution.
(Interestingly, about the time that I was entering this latter phase of my career, it became increasingly difficult for us to recruit rising math PhDs because they were all being recruited by Wall St. at much higher pay. Not sure how many are still there.)
Comments
Yup, I "heart" this one, too.
I momentarily hesitated before posting Housel’s “Good Enough” article.
The cause for the pause was the subtitle of the article: “Why rules of thumb beat precision.” That observation is true in many instances, but it is not a universal truism. Applying Rules of Thumb in an uncritical, rote manner can get you into deep water.
I reflected that pause when I wrote that “The trick is to understand the limits on these simplification rules.” In my haste to post, I failed to include two famous quotes that do a terrific job at characterizing my reservations.
The most famous is controversially attributed to Albert Einstein, but others are also credited: “Everything should be made as simple as possible, but no simpler”. I also like one from H. L. Mencken: “For every complex problem there is an answer that is clear, simple, and wrong”.
I believe these cautionary warnings. Oversimplification can lead down a wrong pathway.
It is often difficult to decide when enough is indeed enough. In the investment decision universe, experience is an important factor to inform that decision. Opinions from an independent and diverse group (like the MFO Board members) can also guide to a better “enough” decision.
I also winced and maybe winked a little when Housel observed that “Investing is often taught as if it's something like aeronautical engineering. It's filled with precise equations that give exact answers…”.
That comment pushed the discussion directly into my ballpark. I spent a considerable portion of my working life in a wind tunnel doing high speed aero testing. The purpose of those grueling, comprehensive tests was to collect sufficient real world data to permit approximate solutions of the governing aerodynamic equations.
The evaluation of those equations permitted performance forecasting and design refinements. Outsiders (like Housel) don’t appreciate the approximations embedded in those aerodynamic solutions and their limitations.
It’s several orders of magnitude more approximate, and consequently more unreliable, in the investment world given the modeling shortfalls. The introduction of emotionally motivated market participants adds yet another complexity dimension that is not modeled. Good luck on making reliable predictions under these circumstances.
Thank you all for your interest.
Best Wishes.
@MJG: I've made similar observations in response to a number of your previous posts, where I felt that undue weight was being given to the alleged precision of such financial modeling. I came very close to making such a response again, but decided that it would likely only start another contentious exchange. I'm a bit surprised to see that we are in agreement on this, at least. The linguistic construction of your comment is particularly succinct and clear: I certainly would not attempt to improve on it.
Folks who do not commonly make sophisticated calculations (non-linear differential equations) overly trust the precision of the numbers computed. This is especially the case if the calculations include some unfamiliar modeling elements.
Scientists, engineers, and perhaps even mathematicians (attempt at humor) know better. They recognize the simplifications, the many assumptions, and the incomplete data that are often inputs to arrive at an answer. These needed shortcuts often significantly compromise accuracy. That’s why safety factors are introduced.
All models in all fields are simplifications of reality. An infinite number of digits after a decimal point is never taken seriously by an experienced engineer. It is merely an artifact of the calculation itself. He appreciates the accuracy limits of what he knows and rounds-off accordingly. I apply that same discipline in many of my posts.
Unfortunately, we frequently don’t recognize our modeling limitations in the investment world. We fall victim to false myths and deficient models; we overreact to very loose, meaningless correlations. Investment outcomes suffer.
This talk about modeling deficiencies and data inaccuracy prompted a memory of a famous WWII mistake told to engineering students as a lesson. It’s a terrific tale about the first US bomber that was recovered nearly intact by the Germans after being shot-down in an early bombing raid.
The German engineers puzzled over the profile of the wing configuration. It seemed to be inefficiently contoured. Wind Tunnel tests verified that assessment. The Germans could not accept that US designers had blundered so badly.
They postulated that the US designers knew some aerodynamic nuances that escaped them. They committed research resources to discover the suspected subtleties. The research failed. There was nothing to discover. The US designers had simply made an error in interpreting their own wind tunnel data. Mistakes on a grand scale do happen.
Many such mistakes are well documented. The scientific community is not immune to them. That’s why in many research disciplines, a duplication of original results from an independent test source is a mandatory standard before product approval and dissemination.
Given the uncertainties of market forecasting and modeling deficiencies, the decision to dump any precision requirements is easily made when doing market analyses.
Thank you for your comments and your patience.
Best Wishes.
"All models in all fields are simplifications of reality. An infinite number of digits after a decimal point is never taken seriously by an experienced engineer. It is merely an artifact of the calculation itself. He appreciates the accuracy limits of what he knows and rounds-off accordingly."
... same applies to these economic forecasts. Is a given report 3.8 or 3.9 or 4.25 percent up or down? WHO CARES (but economists and sell-side pundits)? Give me a good enough rule of thumb that can help me determine a longer-term trend and make my own decision.
I laugh when pundits on TV fret about single-digit-to-the-right-of-the-decimal-point as HUGE areas of contention or concern Puh-lease.
Excellent article, I must say.
After a 40-year career in which the first half involved looking at the
experimental results of one hard object suddenly encountering another
and the second half trying to model these events in collaboration with
some of the smartest people I have ever been privileged to work with
using the most powerful supercomputers of the moment, one event that
always returns to my mind occurred during an informal break at a
project review when the young administrative assistant to our lab chief
very earnestly asked my boss "Do you modelers ever get the right answer
when you don't already know it in advance????"
Compared with economists, financial analysts, climatologists, etc., we
at least could design & and perform experiments -- controlled and
reproducible, more or less -- and physically and otherwise examine
the results thereof. As for economic and financial theories and forecasts,
I would be inclined to put them one step from witchcraft and mirrors.
Enjoy, but handle with appropriate caution.
(Interestingly, about the time that I was entering this latter phase of my
career, it became increasingly difficult for us to recruit rising math
PhDs because they were all being recruited by Wall St. at much higher pay.
Not sure how many are still there.)