World Championship forecast: accounting for the rapid tiebreaks

The forecast for the 12 game match is:

If it ends in a draw, there will be 4 rapid tiebreak games. While rapid tournaments are becoming increasingly common, they are still very much in the minority. So the rapid forecast is based on a much smaller sample. With an ordered logit, I estimated the draw rate to be about 44%. That is in line with the draw rate in the Paris and Leuven rapid events in the Grand Chess Tour.

In rapids, Carlsen is well ahead of Caruana. The champion's rapid rating is 2880; the challenger stands at 2789. Many people have stated that Carlsen is the huge favorite in the tiebreaks. They're right, but Carlsen's chances might not be as great as they think:

Rapid Tiebreaker

Carlsen's chances of winning the rapids are only about 2 in 3! I expected that his chances would be higher than that. Here's why an upset is possible: (1) There are only 4 games in the tiebreaker. Upsets are more likely in shorter matches. (2) The draw rate is lower than in classical. Upsets are more likely to happen when the draw rate is small. Both of these reasons were explained in my recent ChessBase article.

If the match is still tied after the 4 rapid games have been played, it goes to blitz. If the first 2-game blitz match fails to crown a champion, another pair of blitz games follows. This can continue until 5 blitz matches have been played. Though Caruana has a real chance of surviving the rapids, blitz is another story. Carlsen's blitz rating is 172 points higher. 

Blitz: 2-game tiebreak match

Putting it all together:

The percentages in the total column don't sum to 100%. That's because there is a tiny chance that the players will still be tied after 5 blitz matches.

--------------------------------------------------------------------------------------------------------------------------
Follow me on Twitter