Due to the quickly following rounds in the Premier League, I was running behind with the results of the Premier League Prediction model. I will correct this now. In this post you will find the result of round 18, 19, 20 and 21.
I was looking for a suitable roller coaster picture but didn't find any. The problem with the model is that it just can't seem to find any consistency. The results will prove this!
Round 18
A profit for one model which is great on its own. But the results of the other model did temper the mood. If I do extract the correct score prediction from it, it still would be in the red. Also the hit rate of 30% for the 1X2 market based on the correct score is below par.
Round 19
Due to a rush to post the model, I did made some mistakes that why only 8 bets were recorded.
Again the model based on the current season results did better than the model based on the last 4 season results. Only a small loss of 1,01 units, which does qualify it as a good losing week.
The model based on the last 4 season was way off. Only 3 out of 8 for the both the 1X2 markets. The over/under markets did do well, but it was not good enough the cover the losses.
Round 20
Let's see if this round had some more luck!
Third round in a row that the model on the left beats the model on the right. The model based on the current season results had a splendid week. Only the BTTS market based on % did fail miserably, dragging the complete profit down. But hey a profit is a profit.
The left model, well what can I say about it. Looks like it is performance worse as the season progresses. Here the biggest losses where made based on the correct score predictions.
Round 21
And when you do think that everything is lost for the model based on the last 4 season it fights back. For sure the 2 correct score predictions did help a lot. Not is this the market where the higher odds are, but also because it means that all predictions based on the correct score are correct! Even without the correct score predictions it would have made a profit.
The model based on the current season result did gave away the small profit it made during the past 3 rounds. It was saved by also 2 correct score predictions. Without these it would be have a very pity ans shameful result.
Overall
But like I always write, we have to look at the bigger picture. Not that this is more green, in contrast but that what we should do!
Well red all over. Okay not all over. The most important markets (1X2) based on the % prediction are in the green. Both do have a hit rate over 55% which makes the model solid.
If we would extract the correct score market, then both models would still be in the red but it would be in the range of an acceptable loss.
A negative ROI of 10% isn't good but probably is a lot better than lots of bettors out there. Remember that only approximately 5% of all sports bettors are long term winners.
I do think that it is safe to say, that I don't expect a recovery for the model based on the last 4 seasons. Overall it only had 5 green weeks. The model based on the current season result even had 6 green weeks and is running 5 weeks less than the left model.
But it seems that both models are prediction more correct scores. Curious if this trend can continue!
Cheers,
Peter
Comments