Useless Statistics vs a Weight-of-the-Evidence Approach
You want to weigh all of the evidence available in order to make the best evidence-based decision. You don't care that what happened the last 3 days hasn't happened in 5 years, or 10 years or 50 years. You guys don't give a damn that "on average" the market has does X whenever something else did Y. How does knowing that the S&P500 rallies 1.3% "on average" help you at all? I know. It doesn't. These are just useless statistics.
From the Happiness Advantage:
One of the very first things students in intro psychology, statistics, or economics courses learn is how to "clean up the data". If you are interested in observing the general trend of what you are researching, then outliers mess up your findings. That's why there exist countless formulas and statistics packages to help enterprising researchers eliminate these "problems." And to be clear, this is not cheating; these are statistically valid procedures - if, that is, one is interested only in general trend. I am not.
The typical approach to understanding human behavior has always been to look for the average behavior or outcome. But in my view this misguided approach has created what I call the "cult of the average" in behavior sciences. If someone asks a question such as "How fast can a child learn how to read in a classroom?" science changes that question to "How fast does the average child learn to read in the classroom?" We then ignore the children who read faster or slower, and tailor the classroom towards the "average" child. This is what Tal Ben-Shahar calls "the error of the average." That's the first mistake traditional psychology makes.
If we study merely what is average, we will remain merely average.
Conventional psychology consciously ignores outliers because they don't fit the pattern. I've sought to do the opposite: Instead of deleting these outliers, I want to learn from them.
If I’m standing up and I have one foot on hot coals and one foot in a bucket of ice water, my body temperature is about “average”. You can’t argue with that. It’s the volatility in that temperature that is not being accounted for with that statement. Let’s look at the S&P500 because it’s such an easy example that so many of us can relate to. We know that “on average” it returns 8-10% a year, but it rarely actually returns 8-10% a year. When you look at actual returns, we see years where you’re up 43% and up 54% and other years where you’re down 37% and down 43%. In fact, these sorts of volatile returns are actually much more common that “the average” return. Why? Because we don’t live in an average world. Never have. Likely never will.
So when someone tells you that “on average” the market does X on the week of thanksgiving, or “on average” XYZ goes up X after earnings, ask them, “Who cares and why are you wasting my time?”. Don’t let airtime fillers and clickbaiters waste your work day with irrelevant statistics describing what may happen “on average”. It’s embarrassing.
These are funny:
Barry Ritholtz on "Bad Data Mining":
If a gold mine is a hole in the ground with a liar standing on top of it, a successful data miner is a quant with a data set lying to himself.
Here are two statements that I think are a great example of the useless facts preached by irresponsible data miners. These are then promoted by financial media types whose best interest are in clear conflict with the 99.9% of us who are only in the market to try and make a profit. This comes from Charlie Bilello:
Statement 1: the average year has between 3 and 4 corrections greater than 5%.
Statement 2: every year has between 3 and 4 corrections greater than 5%.
Statement 1 is a true statistic regarding the S&P 500 going back to 1928. It can serve as a helpful reminder that there is no reward without risk and that equity securities are inherently volatile (the current period notwithstanding). Corrections happen, even in good markets.
Problems start to arise when investors assume that because Statement 1 is true that Statement 2 must be true as well. Markets, unfortunately, don’t operate that way. The actual environment rarely looks like the average environment.
Charlie goes on in another post:
Are these stats interesting? Yes.
Are they actionable? No. Certainly not in their current form and perhaps not in any form.
Ok, so let's take the other side of that. Why would data mining help you? Well, if you are on the sell side you know by now that the buy side is obsessed with quantifying everything. You know they're going to ask. They love the fun facts. But guess what? They're less and less interested with every day that goes by. You know why? Because their primary goal is making money in the market, not deferring to whatever data mining you've done to back up your story. This is especially the case once they realize that you're only using these "fun facts" when they fit your narrative, and keeping it to yourself when they don't. You want to get the business after all don't you? They know that.....
What about the media? They freakin' love the data mining. You know why? Because they don't care about your portfolio. That's not their problem. The financial media is not in the business of making money in the market, nor are they in the business of helping you achieve your goals. The media is in the business of selling ads to their precious sponsors. You know what gets a lot of clicks? Some irrelevant data mined fact about what happens the average amount of time or how long it's been since something completely arbitrary took place.
The media goes out of their way to look for these data miners. Some of them even create partnerships with data mining companies so they can get even more irresponsibly mined data to stuff down your throats. Holy crap! Again, because their interests are in complete conflict with ours as market participants. The Financial Media sometimes makes it seem like they are there to help you but their actions clearly speak much louder. Their business model is not a secret: Maximize clicks and views to sell more ads. As market participants our only business model is to make money in the market. You see the difference?
So which one are you?
In all likelihood you are part of the 99.99% of us who get paid if we make money in the market. If you fall in that tiny population of data miners or those interested in people who data mine, then go have a ball. I hope you live a wonderful life of spreadsheet building and click counting. I got no beef with you. You're just going about your business. That's fine.
For the rest of us, however, do we care about what happens on average when we know the market almost never ever actually does what it has done on average? Do we care that the market is up 6 days in a row ending on a Tuesday in March under a full moon for only the 5th time this century? No. It's a bunch of useless bullshit. I know this because you guys tell me. This is especially true from the Financial Advisors. I get it. Your "Research" staff pumps you with useless facts all day, but ignores the primary trends and does not provide you with what you need to grow your business and service your clients properly. That's probably why you're reading in this in the first place and coming here for help.
The data miners come to me and say, "JC but why do you need to say these things. Why do you have to call us out like that"
Dude, half the financial advisors at your firm buy my research. I don't know what to tell you.
I want to thank all of you for the constant feedback and questions. This is how I know what you guys are thinking out there, not just in the U.S. but all over the world. Sometimes I think I might be the only one who recognizes something like this, and then 100 of you tell me that you agree completely and are pissed about it.
That's the type of stuff that inspires blog posts like this. A former intern of mine used to always tell me, "JC give the people what they want". She was right.
You guys want actionable ideas. You guys want to identify the direction of the primary trend and invest accordingly. You want to know what's happening all over the world, not just in your country. You want a weight-of-the-evidence-based conclusion. You guys want well-defined risk vs reward opportunities. You're sick of useless facts built off irresponsibly small sample sizes. It's horse shit and we all know it. Thank you for confirming this. All of you!
Cheers,
JC
Are you still not a Member of Allstarcharts? Why not? What are you waiting for? Please let us know here. I'd love to hear from you!
Start a 30-day risk FREE trial today and see why Premium Members have such incredibly nice things to say about their experience with the platform.
Are you sick of garbage data mining? Do you want actionable information rather than useless fun facts? Check out the video from this month's Members Conference Call for FREE so you can get a better idea of how we do things around here!
Sources:
Beware of Data Mining (Ritholtz)
Interesting vs Actionable (Pension Partners)
We Don't Live In An Average World (Allstarcharts 11/21/15)
The Pursuit of Average (Guiness Atkinson)
The Difference Between Statistics and Strategy (Pension Partners)