Using adjusted opportunity, gather, try, and conversion rates I am able simulate the score of matches. Then simulating each pool’s games 1000 times, I am able to predict each team’s chances of winning their pool or finishing in the top two. This method of simulating accommodates all of the different ways a pool can finish.
The simulations only use data from Dubai and Cape Town so I believe there is some overfitting happening. More than likely, each team should be closer to 50/50 than they are shown. Also, injuries are not accounted for. I don’t expect South Africa to be as dominant without Kyle Brown and Cecil Afrika. Lastly, I do not have data for Papua New Guinea so they are using data from Uganda. But considering PNG’s performance in previous tournaments, Uganda’s stats are only helping PNG’s chances.
Pool A
| Team | Pool Win % | Quarter-Final % |
|---|---|---|
| England | 75.7 % | 94.5 % |
| Argentina | 16.8 % | 67.2 % |
| Kenya | 7.5 % | 37.4 % |
| PNG | 0.0 % | 0.9 % |
Pool B
| Team | Pool Win % | Quarter-Final % |
|---|---|---|
| South Africa | 87.2 % | 98.8 % |
| Fiji | 11.2 % | 68.8 % |
| Australia | 1.6 % | 32.4 % |
| Japan | 0.0 % | 0.0 % |
Pool C
| Team | Pool Win % | Quarter-Final % |
|---|---|---|
| USA | 50.8 % | 82.3 % |
| New Zealand | 36.7 % | 71.5 % |
| France | 11.4 % | 37.6 % |
| Samoa | 1.1 % | 8.6 % |
Pool D
| Team | Pool Win % | Quarter-Final % |
|---|---|---|
| Scotland | 65.8 % | 92.7 % |
| Wales | 29.3 % | 81.5 % |
| Canada | 4.9 % | 23.0 % |
| Russia | 0.0 % | 2.8 % |