The Bayesian vs the Frequentist and the Fundamental Flaw in averages

Using "big data" averages in sports betting has it's pros and cons. It's great for setting markets but it's also great for smart people to exploit playing against these markets. Back in school there was this nerdy guy that sat up front in class who everyone regarded the super smart geeky guy. He was intelligent in his own right but largely in terms of having a vast academic knowledge from studying. Looking back I realise he was the classic frequentist, with me down the back of the classroom thinking you are not really smart at all, Bryson. I fancied I could do everything he could and a lot more besides and in a much more timely and efficient manner but here is the catch ........ only if I was bothered. As it happens I was right, I could. It turns out I was the Bayesian.

The Bayesian vs Frequentist Sports bettor

I gathered this guy for our team one night in the table quiz. I wanted him for the academic questions. A question came up - who sang Breakfast at Tiffanys? "Bryson" had the answer sheet. "Deep Blue Something" I told him. For as smart as he was he couldn't discern that "something" was part of the name of the band. He wrote deep blue ....... (blank) on the answer sheet, thinking I was saying I had forgotten what the third word of the band was. 

I always remember this vividly as it was a real eye-opener to the potential limitations of what I now know as the frequentists - often black and white numbers and facts guys. This is the perfect analogy of what we are up against as Bayesians vs the sports betting markets - the universal algorithm with no common sense (logial reasoning) applied to the numbers.

The Fundamental Flaw with Averages in Sports Betting

What are Box Plots? 

A box and whisker plot is defined as a graphical method of displaying variation in a set of data. Box plots visually show the distribution of numerical data and skewness through displaying the data percentiles and averages. We can use a histogram for analysis

Let's look at a golf example taken from my book "Hypnotised By Numbers" where Collin Morikawa is renowned for his approach and iron play. What happens if he has a mediocre 6 months in that department and is ranked 75th for strokes gained approach? Does this make him an average iron player relative to his peers? Some people based on a modelling approach alone would come to the yes conclusion (the frequentists). These people can be destroyed in the betting world. The pure math guys. There's a big difference between being an expert in math and an expert in sports betting.

Golf variance is a very long term game. We are literally not interested in stats that were derived from when a player was out of form when measuring true ability. The true ability of a player will generally be somewhere around the mean of performance at their baseline and at their ceiling. If you are using bare strokes gained stats to try and get an edge on the golf betting markets, think again.

What is the Clustering illusion in Sports Betting?

The clustering illusion is the tendency to erroneously consider the inevitable "streaks" or "clusters" arising in small samples from random distributions to be non-random. It doesn't actually apply to sports betting at all. Why? Because humans are not robots or random number generators. Streaks in small sample data in sports are usually not coincidence. The narrative is real. Form and confidence are the two main protagonists. In fact those two are very much related as well.

"statistically you can have an algorithm and say it doesn't affect anything but believe me as a golfer it's all about the psyche at this level" - excellent from @mcginleygolf. 

This is a quote from Paul McGinley about momentum and the psyche in high level sport. The very same applies to betting on sport in many respects and focusing on these qualitative variables is how you can garner an edge in the sports betting markets instead of fighting fire vs fire using averages and heuristic type frequentist mathematics.

In the tweet above the guy in question was on about shot tracker not realising shotlink only shows 2 points, A & B - where the ball was struck from and where it ended up. There will be a straight line on the tracker from point 'A' to point 'B'. I thought there was a great analogy here between the actual zig-zag journey of the golf ball and what can happen behind the start and end result of what I call the black box of sports betting markets.

I'll leave you with a quote from Charles Bukowski:

" The problem with the world is that the intelligent people are full of doubts, while the stupid ones are full of confidence "

 

---

 Hypnotised By Numbers - the 2nd sports betting book by Bryan Nicholson is coming soon