4 ms·
Yep, Dixon-Coles is way out now. Generally, Poisson-based models don't work that well. They model some aspects of the game correctly but not others. And when it
by hogFeast 6y ago
Yep, Dixon-Coles is way out now. Generally, Poisson-based models don't work that well. They model some aspects of the game correctly but not others. And when it is bad, it is terrible (their paper was genius though, and changed the industry). You also wouldn't do the one-hot/regression stuff on teams, as there is so much player-level data...some kind of online, off/def model is useful though.
You also have the issue of modelling a bivariate as two independent univariates. The adjustments made by the authors were intended to get around this problem but there has been a lot more work on this aspect of the problem since then, and a few new solutions.
And you are right, expected goals was cutting-edge back in ~2011-12 but even some clubs have caught on now. Most gambling syndicates, afaik, actually produce their own datasets (the largest ones employ a huge number of people, I know one that started up five years ago and already has ~10 PHds on staff + traders + data people, etc.).
I will say though, it is surprising how inefficient betting markets were until recently. In the early part of the 2010s, I imagine that large syndicates were making, conservatively, 300-400%/year in the largest markets that were, apparently, "unbeatable". I am aware of very simple models, similar to DC, that could beat retail bookies on Brier for EPL as recently as 2017. And, ofc, smaller markets are far softer (or other sports). At the cutting edge, yes...it probably is complex (I have no idea). But a surprising amount is actually just doing the simple things well.
- fastbluewave 6y agoAny pointers or papers for someone who's just getting into this?
- thom 6y agoThere's probably some interesting work to be done more from a survival analysis angle. I know I haven't seen much work based on stuff like: https://www.academia.edu/37585525/Flexible_Regression_Models_for_Count_Data_Based_on_Renewal_Processes_The_Countr_Package https://www.academia.edu/37585525/Flexible_Regression_Models... I do generally find these top-down team-strength models quite primitive though. Teams are made up of players, plans, all sorts of under-studied interactions. I would also say that there are enough quant jobs in football now that you don't really need to gamble to make a living. :)
- hogFeast 6y agoDixon-Coles, Ntzoufras, McHale, there are a few papers on Bivariate models from German authors (Groll...so google "Groll bivaraite Poisson"). I would also understand ranking algorithms (there is a book called Who's #1?). Bear in mind though, most papers are fictional/p-hacked/just bad. I would also think carefully about what you are doing and why. Most people fail because they try to bet on markets that are, for them, unbeatable (for example, football data is expensive). It is far easier to pick off obscure markets (I did not do this because I had a junior high school maths education and needed some guidance).
- data_spy 6y agoThe thing is, the size of the wager and the payouts are just as important. I was never a sports gambler, but spent two years counting cards at blackjack in rural casinos as my job. This plays out with the Kelly Criterion: https://en.wikipedia.org/wiki/Kelly_criterion https://en.wikipedia.org/wiki/Kelly_criterion So you can have inefficient odds as the house, and still win. Did the models also provide optimal bet size given the probability of winning?
- thom 6y agoLarger syndicates also have the problem that it's not always easy to wager as much money as they'd like on a small handful of leagues. So it becomes about how you can accumulate data and insight into a broader set of competitions.
- hogFeast 6y agoA lot of the newer syndicates spread themselves quite thin afaik. They maybe don't have the contacts to get liquidity/early prices so they do lots of sports (Tennis and Cricket being two that appear to be growing...again, afaik). I have heard that the largest syndicate has groups that only cover one football team. I have no idea whether this is true or how/where you get the liquidity to justify this.
- marktangotango 6y agoThis is the reality of how the bettors move the lines (ie odds on offer), large bets or a lot of small bets will move the line. In effect the final line at kickoff (or whatever) is kind of a distilled, crowd sourced expectation for the result of the match. The key is to identify advantageous lines when the major books publish them and place bets before the "public" moves the line.