by @TheSchachter

An introduction and explanation of the Simple Rating System can be found at the start of my GW17 article.

Lets start by going over GW20 predictions and results:

4/10 correct predictions. Really hard gameweek for the model. And for my team as well (only 39 points, Viduka captain, dropped from 31 to 51 OR).
In a week like this, first we must review why was the model so far from reality. First thing to recognize, is we have 3 draws, the model will almost never predict a draw, but I would have liked to see the draws with a low margin prediction. We didn’t have that with Liverpool and Leeds. Most importantly, and that’s the major disadvantage of a simple model like this, it will not consider or predict rotation. In a heavily rotated (or when Spurs play 2 amateur GKs) gameweek like this a lot of the predictions go out the window. Moreover, the surprising results from this GW will mislead the model in the future to some extent.

0/3 correct recommendations. Amazing results 😊 Especially in a week that I was very sure. I recommended doubling Leeds attack, Chelsea to win and keep CS, and both teams to score at Highbury. WRONG. Not a lot to add here.

For our 5th week doing this analysis, it is worth to look at what have changed since we started. There have been a lot of movements in the different ratings and this could help recognize some trends.
Man U have kept the 1st place in the SRS rating since we started. They started with a 1.17 rating and they have now amazingly 1.17. Between those GWs their ratings fluctuated quite a lot, with a low of 0.97.
Spurs are the team that had the biggest decline from 0.79 to 0.35. In the home\away ratings they completely crushed. They are now the 17th rated home team and their away rating was cut by more than half since we started.
Go back to the previous results and see how the ratings change every week.

Now for GW21 ratings and predictions:
SRS Ratings:

Home & Away Ratings:

Bayes Rating:

GW21 Predictions:

I have Leeds double up in attack with Viduka and Keane, so I have to back my boys here against a shattered Spurs defense and an amateur GK. The fact that we have a negative margin, when a relatively strong home team plays, is very reassuring. (if you look at the home\away ratings, you’ll see that even Spurs the 16th ranked home team, is still higher than most away results, which I think captures nicely the home advantage of the game)

The model backs heavily both Man U and Mid (+2). I will try to avoid recommending United assets (you should always have 3 of those). So, I predict a Mid bounce back here, with the perfect opportunity against the hammers at home. Boksic will hurt the managers who sold him, including me.

For the last, differential recommendation, I will go for the 2nd time with Aston Villa attackers. With 3 losses in a row, coming after their famous 6-1 win over West Ham, I predict they will be at it again against Charlton. With a predicted margin of almost 1. I predicted a 2-1, 3-1 win here.