Monday 14 December 2015

We must have a winner

Tie-break systems are fraught with difficulties, and there is often no good way to separate players who finish on the same scores. The latest example of this is the finish to the London Chess Classic, where tie-breaks were used to determine both the playoff seedings, and the final standings to the Grand Chess Tour.
In the final round Magnus Carlsen beat Alexander Grischuk, after Grischuk missed a strong exchange sacrifice and then found a terrible rook sacrifice a few moves later. With the other 4 games drawn, Carlsen caught up to Giri and Vachier-Lagrave, to share first place on 5.5/9. The win for Carlsen was doubly fortuitous as Grishuk had scored more points than the players beaten by Giri and MVL, and so Carlsen finished on top with a better SB score (the first two tie-breaks being equal).
As the regulations required a playoff, Carlsen also got to watch MVL and Giri battle it out, before playing the winner. Vachier-Lagrave beat Giri 2-1 but then lost to Carlsen 1.5-0.5.
Now here's where it gets weird. The Grand Chess Tour regulations also award the finishing points based on tie-break as well, so Carlsen scored the maximum points (12) and won the overall event. Having beaten Giri in the playoff it might be assumed that Vachier-Lagrave would have finished 2nd, but the playoff was only for the London prizes (not placings!) and so Giri score 10 points and Vachier-Lagrave 8. This relegated MVL to 4th place in the overall standings (instead of =2nd if he swapped points with Giri), and also means he doesn't qualify for next years series (only the top 3 automatically re-qualify, with the next 6 places being based on rating).
If on the other hand the finishing places had simply been shared for players who finished on the same scores in each event (which is especially sensible in a RR), then the standings would have been both different, and probably fairer.

1 comment:

Anonymous said...

Everyone waited around too long on a "safe" plus score, and Magnus caught them at the end.