Here is an update on how my Poisson
Distribution is going so far this season. This covers the whole of July &
August
If you missed it, I talk in an earlier blog post about my
motive behind this but below are the 19 separate bets across the 23 different
leagues that I have analysed
For each game I predicted:
- Which team would win or if it would be a draw
- Double Chance (Win/Draw for each team)
- Score Prediction
- Both Teams To Score (or Not)
- Over and Under 0.5, 1.5, 2.5, 3.5 & 4.5 Goals per game (10 separate Bets)
- Corners – Under 10, 10-12, 12+ (English Prem – L2 only)
- Cards Predicted – Under 4, 4-6, + (English Prem – L2 only)
- Red Card – Yes/No (English Prem – L2 only)
- Home & Away Cards Predicted (Under 2, 2-3, Over 3) (English Prem – L2 Only) – 2 Separate Bets
I will update the scores again at the end of September and so forth on a monthly basis.
Hi Dave, I am actually doing something similar to this but focussing only on the home/draw/away market. Here's a few observations from me:
ReplyDelete1) Rather than quoting the proportion of bets which were successful (as this does not necessarily mean profitable), the true success of the model can only be judged by assessing what the profit/loss would be if the model had been used to place bets on. To do this therefore, you need actual bookmakers's odds (www.oddschecker.com is a good place to get the best odds across all bookmakers). You can then compare your implied probabilities with the bookmakers odds to decide whether to bet.
2) The model can be improved by taking into account who the opposition is when looking back at a particular team's statistics.
3) The goals model could be improved by taking into account information such as shots on/off target, possession etc
I appreciate the latter 2 points require a greater degree of statistical expertise but from working in the industry I know it is the type of analysis people are doing to find that (very small) edge on the bookmakers. Anything less sophisticated is highly unlikely to be profitable.
Hi
DeleteThanks for the comments, always good to see how other people work on similar things.
1) I take your point on how many are successful/unsuccessful rather than profitable but using your method would only bet when there is a clear difference between the bookmakers odds and what my model thinks they should be. I’ve been using it for a few different things rather than actually placing 11k bets, but mainly found it useful for divisions I know little about and as with all betting the model isn’t infallible and I’m always working on ways to improve it.
2) The model does look at opposition as it takes into account their attacking strength & weakness.
3) I have been looking at ways to add things into the model to make it more robust. The problem with the things you mention is how much actual influence do they have over the result of the game? You would think something like shots on target is clear but look at PSG’s game against Ajaccio, this finished 1-1 yet the shot count was 37-1 in PSG’s favour. Same with possession I think Stoke ceded possession in something like all but 3 games last season and yet finished the season with 9 wins and 15 draws. I accept that these are extreme cases and every model should be stable enough to not be influenced too much by outliers.
As I said it is something I’m working on, one thing that is throwing it out at the minute is Monaco. I have accounted for teams getting promoted/relegated but with Monaco spending well beyond the normal capabilities of a team promoted from Ligue 2 they aren’t following the normal curve. Same applies for teams which have gone into administration and been relegated.
Thanks again for the feedback
Cheers
Dave
Right away the positioning may possibly irrefutably have renowned among the most with writing individuals, because hardworking write-ups or significant assessments. 메이저사이트
ReplyDeleteThank you for very usefull information.. บอลออนไลน์UFA800
ReplyDelete