The poisson models based on the correct score do like draws a little bit too much, but round 17 in the Premier League didn't had any draws at all. So, curious how the results will be!

Round 17

Well both models made a loss. It looks like a fully decorated Christmas tree, where all the red decoration do cover up almost the green tree!

Both models did fail to predict a correct score. But even if I would exclude the correct score markets both models would have made a loss. A loss which would be maybe a little less but still a loss.
The regular 1X2 markets in both models made a profit. Also a great indication that odds do matter. If we take a look at the 1X2 markets based on the correct score prediction, we do see that both models did have a hit 60% but failed to make a profit. This is mainly because the model did correctly bet on the favorites!

Season 2018/2019

Let's start with the positive points. Both regular 1X2 markets still do make a profit. Both prediction around 57% of the games correct. But below that red takes over!

If I would exclude the correct score prediction from the model based on the current season results, it would just not be around break even.

Also both models are suffering from a losing streak, the model based on the last 4 seasons now had its fourth losing week in a row, but the model based on the current season results is still going stronger because it now has a losing run of 7 weeks :(

Model based on the last 4 season results
Model based on the current season results

Guess that I will have to keep the faith!

Cheers,
Peter