Hi All
Producing the workbooks for laying the draw in one of my previous threads and the discovery that they are simply amended to other leagues I have decided to produce another set of workbooks that will shortlist games that have a high probability of early goals and plenty of them to optimise trading the overs markets and primarily the Over 2.5 market
A work in progress template is available in the usual place
https://bet-or-not.com/2022/03/30/over- ... -template/
Does anyone have any suggestions as to what relevant statistics can be included on top of those already in the workbook
The following are the stats already in the workbook
Average minute OGS was scored (Home)
Percentage of Home games with Over 2.5 goals
Average Number of Goals scored at Home per Match
Percentage of games with Over 2.5 goals in the last 8 Games
Percentage of games where both teams scored in the last 8 Games
Any input would be helpful
J
Trading Over 2.5 goals ANOTHER of my workbook aproaches
Very late to this topic - I have a machine learning alforithm that uses a deepnet to select games that are under 2.5 goals. The data I use is:
Gap: the distance between the two teams in the league
Home GF: Home golas for
Home GA: Home goals against
Away GF
AwayGA
Home HT goals
Away HT goals
First goal home: First goal scored in gamesfor or against
First goal away
> 3 goals last 4 home
>3 goals last 8 home
>3 goals last 4 away
>3 goals last 8 home
home % above 3 last 8
away % above 3 last 8
There are others I use but these are the ones that have the biggest correlation.
My data is all automated from API's so I dont have to work this out and the last training was done on 12k results.
By and large I have an accuracy of about 63% correct when the Phi Coefficient is above .55 and it makes about 30 points per season on selected games.
I would expect this to get better over time. I now have 17k results and will be retraining again over the summer.
Gap: the distance between the two teams in the league
Home GF: Home golas for
Home GA: Home goals against
Away GF
AwayGA
Home HT goals
Away HT goals
First goal home: First goal scored in gamesfor or against
First goal away
> 3 goals last 4 home
>3 goals last 8 home
>3 goals last 4 away
>3 goals last 8 home
home % above 3 last 8
away % above 3 last 8
There are others I use but these are the ones that have the biggest correlation.
My data is all automated from API's so I dont have to work this out and the last training was done on 12k results.
By and large I have an accuracy of about 63% correct when the Phi Coefficient is above .55 and it makes about 30 points per season on selected games.
I would expect this to get better over time. I now have 17k results and will be retraining again over the summer.