so i create this thread to talk about Elo raings
i recently heard of this term and i m curious how it can apply on making profits on sport betting
for now i know nothing about it so this thread will have the form of diary mostly with me posting my knew knowledge
and questions and hopefully answers from people who are familiar with it
what about Elo ratings?
here is the definition from Wikipedia
https://en.wikipedia.org/wiki/Elo_rating_system
https://en.wikipedia.org/wiki/Elo_rating_system
My brother's a statistician and although he uses more sophisticated software models for betting on football, I believe he has used ELO for premiership football in the past and said they are quite useful. I didn't realise it actually predicts probabilities of outcomes, it would be really cool if they are near accurate.
I've been using ELO Ratings for many years in other sports spheres than betting and, as long as the two parameters (s-curve & k factor) are set correctly, and the Initial Rating are reasonably accurate, then ELO is a good indicator of future performance. If these are not set correctly, they are a waste of time. You have to be fairly confident that the "administrator" who is operating the Ratings knows what they are doing.
At the risk of appearing to be an Anorak, these Rating were devised by Pofessor Arpad Elo (a Hungarian born, American based, mathematician) in the 1960's for the Interenational Chess Federation to rate the "Grand Masters".
Since then other, more sophisticated Rating systems have been devised, but Elo forms the basis of most of them.
At the risk of appearing to be an Anorak, these Rating were devised by Pofessor Arpad Elo (a Hungarian born, American based, mathematician) in the 1960's for the Interenational Chess Federation to rate the "Grand Masters".
Since then other, more sophisticated Rating systems have been devised, but Elo forms the basis of most of them.
I've gathered that the ratings are self-correcting, i.e. inaccurate ratings will gradually become more accurate, but what figure do you use if you're starting from scratch? Presumably every competitor starts with the same rating?crispi wrote: ↑Mon Jul 22, 2019 9:47 pmI've been using ELO Ratings for many years in other sports spheres than betting and, as long as the two parameters (s-curve & k factor) are set correctly, and the Initial Rating are reasonably accurate, then ELO is a good indicator of future performance. If these are not set correctly, they are a waste of time. You have to be fairly confident that the "administrator" who is operating the Ratings knows what they are doing.
At the risk of appearing to be an Anorak, these Rating were devised by Pofessor Arpad Elo (a Hungarian born, American based, mathematician) in the 1960's for the Interenational Chess Federation to rate the "Grand Masters".
Since then other, more sophisticated Rating systems have been devised, but Elo forms the basis of most of them.
Yes, thats right but if you start of with all competitors on the same initial rating it will take some time before they become accurate enough to use. If you have enough information/knowledge to assign a reasonably accurate IR then they become accurate more quickly - as long as you have the parameters right!
- wearthefoxhat
- Posts: 3220
- Joined: Sun Feb 18, 2018 9:55 am
Wouldn't surprise me if the Official Handicappers use something similar-ish. (if not why not)
They have enough data to establish a starting point for any horse (weight for age) and with the formulae tweaked here and there, can adjust it over time when compared to other similar types.
It would then be easy enough to predict ones that have lbs in hand and ones that are deteriorating and can be dropped.
They have enough data to establish a starting point for any horse (weight for age) and with the formulae tweaked here and there, can adjust it over time when compared to other similar types.
It would then be easy enough to predict ones that have lbs in hand and ones that are deteriorating and can be dropped.
Problem is that ELO was devised for a 2 competitor sport and when you have multiple competitors the logarithm becomes very complicated. There are ELO calculators on the net that can do this, but I've never used one so have no idea how accurate they are.
Too many variables in horse racing (going, distance, pace & positioning, etc) that you don't get in football, tennis or chess, makes any rating system based purely on matching horses against each other or position or distance beaten futile.wearthefoxhat wrote: ↑Mon Jul 22, 2019 11:39 pmWouldn't surprise me if the Official Handicappers use something similar-ish. (if not why not)
They have enough data to establish a starting point for any horse (weight for age) and with the formulae tweaked here and there, can adjust it over time when compared to other similar types.
It would then be easy enough to predict ones that have lbs in hand and ones that are deteriorating and can be dropped.
Yes, I'd agree with that. A way around that may be to Rate every variable seperately and then bring them together in a "Master" Rating. I'd be inclined to then analyse the seperate ratings to see if any are more influential than others so that they could be weighted,
Elo isn't a perfect system but millions of people around the world (mostly kids) obsess about their Elo ratings every day in competitive online video games.
https://dotesports.com/general/news/elo ... ined-20565
https://dotesports.com/general/news/elo ... ined-20565
Since when do you care about forum rules, your last thread was straight up deleted. Maybe you would learn more if you stopped with this simpleton charade, you're just fishing for ideas and quizzing people about Poisson and ELO so that you can sell more subscription based services. You already sell enough of them.
-
- Posts: 3140
- Joined: Sun Jan 31, 2010 8:06 pm
Stop feeding him then. It's quite blatant what he's doing no need to encourage him.Kai wrote: ↑Tue Jul 23, 2019 12:33 pmSince when do you care about forum rules, your last thread was straight up deleted. Maybe you would learn more if you stopped with this simpleton charade, you're just fishing for ideas and quizzing people about Poisson and ELO so that you can sell more subscription based services. You already sell enough of them.