O2.5 filter based on EPL 2021-22 stats

[A Former User]

Hi all,

This is something I've thought about doing for a while now and with the introduction of the back testing on the horse software and with the EPL season just finished, I'd thought I'd give it a go. At present we can't back test on the football stats but we now have the data for the season just finished so I thought I would go and look at the games which finished with over 2.5 goals and see what the similarities between them were with a view to building a filter around them.

So to the data...there were 380 games in the EPL with 205 (53.95%) going O2.5 as shown in the table below:

I focused on the 205 games which were O2.5 and have broken them down as below:

So you can see that BTTS occurred in 72% of the games which ended up O2.5 and a FHG in 89% and a SHG in 97%. So these 3 stats are ones to look for when building the filter. This might all sound like common sense but it's nice to have it confirmed in the data.

Also the shots on target and corners for both the Home & Away teams are of interest and with the BTC stats we can incorporate them into the filter.

Last but not least are the odds. I've used the odds from B365 as the site I downloaded the data from didn't have the exchange odds available but at least they give a consistent baseline. As shown below, the lowest odds for O2.5 were 1.28 and the highest was 2.5 with the average being 1.82:

Probably not a surprise that the Liverpool v Leeds game had the lowest O2.5 odds!

So to the filter... the things I will be including are:
Overall BTTS
Overall H1 Over 0.5
Overall H2 Over 0.5
Overall FT O2.5
Home shots on target per match Average
Away shots on target Average
Home corners per match Average
Away corners per match Average

I'm not going to include an odds filter until I've got some games on this but will look to record them as I go.

I had planned on using the average figures from the EPL data for each of these e.g. 5 Home shots on target, 4 Away shots etc. but I might reduce that slightly just now until I get some matches appearing and then tighten it up if need be. I'm planning on doing a similar exercise on some of the other top leagues around Europe and see how they compare to the EPL stats and will update this post when I get around to doing that. I think that Martin or Ryan said that they have different filters for winter and summer leagues so who knows, maybe I'll end up with a different filter for each league!

Hope this is of interest and I'll update it once I've got a few results on the board.

[Matt Wood]

@ian-wylie Thanks Ian, thats great.

[A Former User]

@matt-wood Hi Matt, I got them from the site below, they've got a good number of leagues on there which are all free to download the data. There's also a text file which explains what the abbreviations are for the column headers on the downloads.

https://football-data.co.uk/downloadm.php

Cheers

[Martin]

@ian-wylie keep it up! great work

[Matt Wood]

More great stuff Ian. You mentioned downloading the data, are you able to say where the data came from? No worries if not.
Keep up the great work though

[A Former User]

I've completed the same exercise for the Serie A season just finished and the results are shown below. They are very similar to the EPL ones so hopefully my filter will be along the right lines once I start getting some games appearing on it.

[A Former User]

I’m loving this,

[Ryan]

Really interesting read this, excellent work.

Cant wait to see this in action.

[Matt Wood]

Excellent stuff Ian, really interesting read

[A Former User]

@ian-wylie great work Ian thanks for sharing mate. I can't wait until we can backtest the football like we do with the racing

[Martin]

great read and very good work!