Useless Lying Statistics - Part 1
If You Torture the Data Long Enough, It Will Confess to Anything!
The modern world of 2021 seems obsessed with useless lying statistics. While some stats may be technically accurate I find it rare for them to tell a complete story with more detail and context. I think some of the obsession with stats began with the MTV generation of a decrease of attention span. "Just give me the results!" "Can't I just start at the finish line?"
Let's start with useless lying statistics in sports
Some of you already know I'm a big sports fan. Three Cheers for Baseball! I watch all the major American team sports albeit less than I used to in my younger days. It's currently summertime so I'll use a broadcast of that in my example. Baseball really is a simple game at its core and in the future I'll be comparing the simplicity of baseball to the simplicity of stock selection. "See the ball, hit the ball" is what Pete Rose used to say about his hitting philosophy. Now you may think Pete Rose is an asswipe (I do too) and a knucklehead but there is no denying he had more hits than any other baseball player in history. I have to guess that "see the ball, hit the ball" is a simpler approach than the garbage stats jammed down our throats in a modern baseball broadcast. You think ol' Pete gave a rat's ass about a discussion of launch angle and exit velocity of a batted ball? I don't think so!
Can't we just watch the game?
I can see using cruddy stats to help the players improve but how does all this masturbatory information enhance to viewing experience? The statistic that counts is which team won and which one lost? The same goes for the other American major sports of Football and Basketball. Sometimes I put on ESPN in the morning and see some junk like a proprietary "power index" percentage change for one team to win a game. Is any of this crap useful or actionable? We seem to be inundated with useless lying statistics when we don't care about them but they are mysteriously absent. They're a little like cops, never around when you need them. You know the absent stats I would like to see in sports? I would like to see how often those pointy headed nerds making up their power index predictions? How often are they right? You know who is often right? Those Vegas oddsmakers. Those mo-fo's can nail down some predictions of how a game will go and when you consider the "juice" or the vigorish involved in making a wager. It's no wonder it's so hard to beat the house. Not only do you need to win more bets than you lose but also make up for the 10% the house takes.
Does any of this rant apply to finance?
Yes. I'll give you a couple of examples from our own finances and writing about the Smidlap Portfolio. Back in 2018 I bought a small starter position in a fantastic company and stock, Intuit, whom you may know as the maker of Quickbooks and TurboTax. Well, the company continued to perform so well that when it was time to deploy some more cash in the fall of 2020 I bought a much larger position in the same stock, about 4x the number of shares I initially bought. The initial buy was @$228/share. The second larger buy was @$328/share. Today the shares sit at a tidy $580/share.
note: not all our investments are go up in a straight line like this. Sometimes we make bad picks that lose us money.
This is my take on the useless lying statistics around this investment. When I report our results each month for accountability those stats are for the combined investments or all the different buy points. Those are naturally weighted and skew towards the highest number of shares purchased. So when you see INTU on our list you see a gain of around 88%. The gain on the initial small number of shares from 2018 is a very sexy 150%! Meanwhile the gains contributed from the 2020 purchase are only 75%. The part you don't see is the larger purchase from 2020 has contributed almost 3 times the absolute dollar gains towards the total due to buying so many more shares the second time. So...even though the 75% is not nearly as sexy as that 150+% from 3 years ago the objective was not a bunch of "look at me" statistics. The objective in team sports is not a high batting average or spin rate in a curveball the objective is to win the game. You win the game of investing by maximizing the total dollars returned, statistics be damned!
Useless Lying Statistics case study: Missy versus Smidlap
If you are new here in addition to tracking our own stock portfolio I also write a column once a month called the Malevolent Missy Stock Series. I have Missy buy one stock a month in order to beat the major stock indices like VTSAX or the S+P 500. If you just looked at the statistics comparing the two portfolios you would say the Smidlap Portfolio is a much much better performer. But that completely overlooks Missy's objectives and her point in her investing career. She is young and we Smidlap's are relatively near the finish line of accumulating our retirement money. Here are how the performance graphs look against the indexes:
The Smidlap Portfolio over 5 years
Missy's Results over 24 months
Here is where the statistics lie in my opinion. Missy is still accumulating her wealth and buying an equal amount of stock each month. After 24 months as soon as she makes her next buy an automatic zero is averaged into the calculations for both hers the the indexes' results. On day one they haven't yet had time to percolate or matriculate to gains or losses yet. On the flip side we have not added any significant funds to the Smidlap Portfolio since 2017 as we enjoy our glide path into retirement. So not adding significant fresh money and averaging in the lower returns for more recent purchases skews everything in our favor on the graph as does the longer timeline. That 5 year versus 2 year chart for Missy really makes a huge different especially in a market that has been trending steadily upward. Here is my last illustration of time being a huge factor but why Missy should not worry about big sexy statistics on a graphic just for bragging rights. She adds a relatively equal buy each month like any good financial lumberjack ought to. Time and consistency win the investing game. I recently broke down Missy's gains by the year in which she bought the shares to illustrate why she should not worry about the stats and averaging in lower numbers. 2019: 4 stocks averaged 154% gain vs. 48% average for VTSAX 2020: 12 stocks averaged 90% gain vs. 45% average for VTSAX 2021: 8 stocks averaged 7% gain vs. 8% average gain for VTSAX So, would any fool or knucklehead say our gal should have stopped at the end of 2019 just to claim bragging rights to a 154% gain? ! I didn't think so. Much like the Intuit example above Missy made over $6000 so far from those good selections from 2019. If she rested on her laurels and was satisfied with some idiotic idea of Coast FI in early 2020 she would have missed out on over $11,000 gained in 2020 and another $600 for her slow starting 2021 purchases. Missy knows the objective is to grow the total pie to fund her future liberty from any putrid employer or bad situation in general. Our gal is smarter than to be solely concerned with a bunch of useless lying statistics. She is merely content to use stats when called for and appropriately flush them down the bowl when they are not. What other statistics lie? I was thinking of writing one about geoarbitrage and related stats. There are plenty of places I don't want to live even if I could stay there for $11.87/year. p,s. The annual Smidlap-con is coming up around the Labor Day weekend and week. If you are in the Adirondack Park region give a shout and we'll have a hike followed by a beverage.