From Round 6 to Round 12, the Bradley-Terry model correctly tipped 31 of the 56 games played. Is this any better than simply flipping a coin to decide who wins or loses?
It is certainly possible to get 31 (or more) games correct out of the 56 games just by flipping a coin. In fact, the chance of this happening is approximately 25.2%.
In other words, if 56 coins were to be tossed over and over again 100 times, then 25 of these occasions they would be able to correctly predict as many as the 31 winners predicted by the Bradley-Terry model. There is therefore no evidence that the Bradley-Terry model is in fact doing better than just flipping a coin.
The above probability of 25.2% is an example of a p-value which is often used in hypothesis testing to make decisions.
cross ✛ validated 