Rugby World Cup Week 2 (11/09 - 17/09)
A/B Team predictions
Slight deviations from A vs A predictions above due to updated model being used
Predictions for this week
This week is relatively straightforward as far as predictions are concerned. While last week there were 3 games that were going to be very close (4 if you count South Africa vs Scotland), this week there is only 1 that is likely be close - Australia Fiji.
The model has 80% confidence that Australia will win - that still means that 1 in 5 games it expects Fiji will win
For the rest of the games the model is 90%+ confident in the winner
Detailed Stats
For a lot of the matches this week there are uneven match-ups, therefore it is hard for the model to predict the stats. As a result unfortunately you will see a lot of N/A this week. This is because there is not a lot of evidence of how France plays against a team like Uruguay (the difference in rating is so large)
My Model vs the bookies
Appraisal of least week
Overall I am pretty happy with the preformance of the model last week
7 out of 8 games correctly predicted
Only incorrect prediction was Wales Fiji and that was a very un predictable game
88% overall accuracy - same as bookies
Beat the bookies handicap in 50% of the games
One thing that is interesting is that the model underestimated the scores for a lot of the imbalanced games (e.g. Ireland vs Romania)
Very slight tweak for this week to account for that
