General Betting
There is currently 1 person viewing this thread.
13 Feb 20
17:30
Joined:
Date Joined:
02 May 08
Is anyone willing to share any insights/ thoughts on the key parameters which should be factored into a model/ regression analysis to determine xG for a team in any particular game?
In my view, accurately being able to determine the xG for any team in a given game is the 'holy grail' of football betting as it would theoretically enable you to accurately price up all markets and in play (and thus readily identify any market inefficiencies/ value prices).
For those who have built models, I am also interested in any views people area willing to share as to how you factor 'non tangibles' into your model (e.g. news that a "key player" may have a slight injury, extreme weather conditions, unrest in the squad etc).
I have built a regression model to calculate xG for League 1 and League 2 which outputs value bets which does appear to generate a (very small) profit. However, I feel like I am looking at far too few variables and need to expand this to get a much more robust xG prediction.
I've not posted on this forum in a long long time...good to see it still going strong(ish)!
In my view, accurately being able to determine the xG for any team in a given game is the 'holy grail' of football betting as it would theoretically enable you to accurately price up all markets and in play (and thus readily identify any market inefficiencies/ value prices).
For those who have built models, I am also interested in any views people area willing to share as to how you factor 'non tangibles' into your model (e.g. news that a "key player" may have a slight injury, extreme weather conditions, unrest in the squad etc).
I have built a regression model to calculate xG for League 1 and League 2 which outputs value bets which does appear to generate a (very small) profit. However, I feel like I am looking at far too few variables and need to expand this to get a much more robust xG prediction.
I've not posted on this forum in a long long time...good to see it still going strong(ish)!
sort by:
Show
per page
Replies: 1
|
By:
have a look in the Football forum - Premier League Predictions using XG
I have invented a simple model based on xG averages to predict correct scores. The model is over 50pts in profit since its inception I can share exactly how I am doing it and from which site but it would have to be in a PM as I don't want to give away my secrets to everybody ;) If i was you what I would do with your League 1 and 2 data is something similar, work out a weighting system to work out an average expected xG and xGA in home away matches and make the data come somewhere close to an expected goals figure based on that and that gives you your correct score. then trial it against the market to see how well it performs For e.g. Rochdale vs Swindon - i'm just making up teams here Rochdale average 1.65 xG at home, Swindon concedes 1.80 xG in away matches, so Rochdale would be expected to score 1.73 goals in the game (rounding it up to 2). Swindon 1.20 xG away Rochdale xGA at home 1.11 Swindon would score 1. Correct score = 2-1. You get the idea, it's pretty simplistic. It's worked for me so far but i could have just been lucky. time will tell |