Fink Tank Test: This weekend’s games

I’m testing the predictive and value betting ability with the Fink Tank ratings. Have a good read of the first post here to make sense of it.

Last week the Fink Tank performed as follows:

Raw 1×2 betting: +2.56
Raw with Asian Handicaps:
-0.64
With team news filter:
-0.54

Too early to say anything concrete with the bulk of the raw profits coming from the Blackpool bet which didn’t make it through the team filter.

This week’s matches:

Quite a few, with a home bias.

That first bet is a Double Chance on Birmingham as there is value in the draw and Birmingham according to Fink Tank.

Fink Tank Test: This weekend's games

I’m testing the predictive and value betting ability with the Fink Tank ratings. Have a good read of the first post here to make sense of it.

Last week the Fink Tank performed as follows:

Raw 1×2 betting: +2.56
Raw with Asian Handicaps:
-0.64
With team news filter:
-0.54

Too early to say anything concrete with the bulk of the raw profits coming from the Blackpool bet which didn’t make it through the team filter.

This week’s matches:

Quite a few, with a home bias.

That first bet is a Double Chance on Birmingham as there is value in the draw and Birmingham according to Fink Tank.

Testing the Fink Tank

Over the world cup we put Castrol rating’s prediction model to the test and were quite impressed. Sadly they aren’t applying their model to make predictions for domestic games, but we’ve another public and well researched model we’re going to be putting to the test this season.

The Fink Tank is a football prediction model developed by Dectech with academic research on football probabilities from the university of Warwick. The Fink Tank was the engine for many an interesting football column in the Times newspaper. Sadly that column appears to be no more, but the ratings are still available.

The ratings are used to create a percentage chance that the home or away team has of winning, with the draw also taken into account.

In a past edition of our magazine we put Fink Tank’s predictions for previous seasons to the test with some interesting results. So this season we thought we’d continue to test the ratings model using live odds.

Value bets:

Bets are taken when Fink Tank’s ratings give a team a better chance than the odds currently imply. To ease the admin burden, we’ll just be tracking the Premier League.

The downside of Fink Tank is that its model does not take into account team news. So if a star player is injured, the ratings don’t adjust. This season we’ll be trying to account for this by running a team news filter.

Castrol ratings provide a score of individual players based on a variety of statistical factors. The team news filter will check each team for injury news and make it a no bet if a star player (as deemed by Castrol) is not playing. We’re also not sure if Castrol accounts for the fact the promoted teams stats will be against weaker opposition so we’ll also be applying the team news filter to promoted teams for the first few weeks.

We’ll be tracking the raw results and the results with the filter applied separately so we can see which works best.

Without further ado, here’s this week’s qualifiers:

The value is in relation to the home or away odds. Not the actual bet selected. In most cases we’ll be using the home & away odds, but from time to time if the underdog is really long odds, we might apply a handicap bet.

The team news filter takes out the newly promoted teams. Villa are without their top defender (going by the latest team news) and Man City (Spurs’ opponents) have added a significant batch of new players recently which may make Fink Tank’s ratings a little obsolete for a few weeks. Everton becomes a no bet if Cahill doesn’t play.

This is just a tracking experiment, so please don’t back these.