Football Musings

[wearthefoxhat]

This is a musing prompted by the Matthew Trenhaile article. It`s based upon Pinnacle odds and uses the Home Win odds as a comparator to actual full time scores. Simply, Home win odds = 2.0, 305 fixtures, 144 Home Wins, 83 Draws, 78 Away wins. The initial odds are =(FLOOR) in Excel to group such that each odd value has numerous entries. This is the Premier League from 2004 to 2022 (games played) and plots the Home Win Odds on the bottom axis and the % of games that ended Home Win, Draw or Away Win.


Pinn1.jpg


So by taking the home win odds from the Pinnacle site you can find the corresponding value on the bottom axis and read off how the Pinnacle estimations evolved into actual %`s. The home win odds on the lower axis are NOT the full extent of the values available but the graph loses clarity as more are shown.


Pinn2.jpg

Thanks for posting, nice graphics.

I should be using the Pinnacle site for odds comparison.

[wearthefoxhat]

Thought I'd put up some musings for one of todays league 2 games. (Kick Off 2pm today)


2.94 (touched 3.00) for a 1.84 chance seems to be good value.

Northampton.png

[wearthefoxhat]

The market was against Northampton as the price ended up around 3.10 just before the off. They weren't far wrong as Stockport came out fast and furious and deservedly took an early lead, and maybe should have scored again.

It turned out to be a game of 2 halves, as key subs were made at HT, making all the difference, with Northampton winning 2-1.

HT.png

[wearthefoxhat]

Got this one wrong, but there were other signs that it wasn't that clear cut.

Prediction was a clear Home Win, the result suggests L.Orient "should" have got something from the game, but whoscored picked up on Newport being a bogey team and their outline was more accurate on the game. Also, there was no clear value with the algo, more or less matching the Pinnacle odds.

1.png

2.png

3.png

[wearthefoxhat]

Based on Xg stats, using recent Home/Away records and recent game comparison, Spurs look value today at Home.

I haven't included either teams' recent european games (maybe I should), just the Premier games. (maybe the market's over-hyped Liverpool's win v Napoli)

Pic1.png

Pic2.png

[Archangel]

I find it hard to believe that Liverpool are so short away to Spurs

Third in the league vs Ninth !

[Kai]

Liverpool tend to show up in big games, while in others they think it's enough to just show up

Transitional season means prices will be all over the place, harder to price them so better for trading

[henbet22]

Could be anything this 90 mins. Every match Liverpool win the media thinks it will kickstart their season. Klopp tilting at every windmill like a sober ol' big head in his pomp. Spurs could start slow offsetting Liverpools poor habit of not getting going from the whistle. This musing thread is great fox. Hope your methodology is still coining it in.

[wearthefoxhat]

Watched one of Peter Webbs Bet Angel academy for football. Most of the content he's posted up on You-tube over the years, but it got me thinking on how to create my own Xg figure.

As Peter outlined, the Xg stats world has expanded over the years and there are various sites that display the information, but they do vary and calculate the figure somewhat differently. Peter himself says he has 19 different metrics to help him decide on goals/no goals in a game, so, how hard can it be to set one up myself...

Fortunately, there are some good football websites out there that show data and this can be converted into a spreadsheet and crunched together. The mission, if you wish to accept it, is to focus in on the key metrics that make up game flow and influence goals, or not.

During the world cup, I've tested out a few metrics and weighted them to produce a FoXg, comparing it to a reliable Xg on infogol. (Understat are good for the top leagues)

Here's a couple of outcomes.

FoXg.png

One thing I don't want to do is replicate an existing Xg model, may as well just use infogol and save time. I did want to use my own influence on the rating, based on what I perceive to be important. The website I find useful for data is Sofascore and the information displayed in their statistics section.

Metrics.png

One other "influence" that is worth including was from an old book - Profitable Football Betting - Paul Steele ISBN 978-1-84344-017-8

Paul Steele.png

Page 88 - System Thirteen, can be set up easy enough on a spreadsheet and incorporated.


Going forward, for league games, once I'm satisfied with the rating output, I can then create Home/Away Foxg tables and compare Home/Away records when teams play each other, in turn predicting the outcome of a match as outlined already in this thread.

[wearthefoxhat]

Query.png

As already said, it's work in progress.

The reason for posting it at HT/FT is to use it as a guide on how the game is developing or if it was a fair result after taking into account the actual score. Ultimately, I could use them for an obsure league game that has stats on sofa-score or flash-score, but no Xg on infogol.

Yep, a goal is a goal if headed for sure, my observation is that the real Xg may/may not calculate headed goals/chances differently.

If I'm watching a game, I can input key metrics into the spreadsheet and see how the game flow develops. I can use it to reinforce pre-match predictions or change my mind if there's evidence to do so. (eventually I could scrape the data directly into the sheet and assess in real time)

I could also build up data on teams that are due to play against each other and produce my own prediction on the probable outcome and create my own match odds.

[wearthefoxhat]

Just asked Chat GPT...

AIChat.png

[wearthefoxhat]

There's a YT post from Ben @ Sportstradinglife that recently showed how to use basic Xg in a sequence of IP trades.

https://www.youtube.com/watch?v=9mZNAek ... &index=186

It adds to what I previously posted during the World Cup and can be used on the bigger leagues. A couple of ideas that can be considered going forward not mentioned in the video.

1. Total Xg on both teams compared to the basic momentum stats on sofascore.
2. Impact of substitutions anytime in the second half.

Of course Xg on Infogol can be used on other leagues if necessary.

[Kai]

Just asked Chat GPT...



AIChat.png

They probably named it xG to avoid confusion there, I'm sure Chatty would recognize the xG metric correctly

[wearthefoxhat]

Just asked Chat GPT...



AIChat.png

They probably named it xG to avoid confusion there, I'm sure Chatty would recognize the xG metric correctly

Just asked it again with a more specific question.

XGchat.png

[wearthefoxhat]

As there are no official Xg stats around, it's a good job I can create my own for tonights FA Cup clash.

ManCvArse.png

On balance, Arsenal are performing better and could of been ahead in the first half.