Candidates 2016 - What Are Your Predictions?

Cover photo from Agon Limited

“To play a match for the World Championship is the cherished dream of every chess player.” – David Bronstein

It’s 2016, it’s World Championship year, and on March 10^th, 8 of the world’s top players will meet in Moscow, to play for the right to play Magnus Carlsen, the reigning World Champion, for the ultimate title in the game of chess.

Now, there of course will be predictions bouncing all around the place, so why shouldn’t I get in on the action! However, I will base some predictions using some more statistical methods, and see if they can stack up against my intuition.

Here are the 8 players, with their March 2016 ratings:

So, I’m going to use a variety of different methods to get some predictions of how the Candidates will go.

Elo Ratings

The Elo ratings of two players combined (more specifically, the difference between them), will give an expected result (often abbreviated to ‘We’) for each player. This difference is then subtracted from the actual result (1, 0.5 or 0), multiplied by a ‘K-factor’, and then the result is the rating change for that game.

So, if we take the Elo ratings, we can calculate the We for each player, and then get a prediction for the final standings:

(The score figures have been rounded, so do not add up to exactly 56.)

The massive disadvantage about this is that the highest rated player is predicted to finish 1^st, the second highest rated 2^nd… and so on. It also puts the higher players extremely close, and the whole field is separated by less than a point, so this isn’t an accurate way to predict this at all. It is also difficult to use Elo ratings to predict individual results, even between Caruana and Svidler (the highest and lowest rated), the expected result between them is 55.3% for Caruana, and 44.7% for Svidler. So, as predictions go, this is not a good way to do it at all.

Past Results

Of course, it would be a better idea to take the player’s past ratings against each other, and then see if one has a much better score than the other. For this, if a player had a +2 or better record against their opponent in classical play, then I would predict they would win, if it was -2 or worse then I predicted a loss, and otherwise I predicted a draw. I also did this separately for white and black.

Now there are some possible inaccuracies with this method, one being that some of these result predictions are based on games played 5-10 years ago, maybe even further. For example, Svidler hasn't played a classical game as White against Nakamura since 2010.

Using this prediction, Aronian is predicted to win 6 of his 7 games as White! The most wins by one player across 14 games in the past two Candidates tournaments is 5, by Carlsen and Aronian in 2013, so going 6/7 in games as White is extremely unlikely. His overall score is reasonably realistic though, his 9.0/14 is a +4 result, and the previous two Candidates tournaments were won with 8½ points (+3). Aronian also has been in a good run of form recently in classical play (winning the Sinquefield Cup in September), so he could do well.

Another thing I did not take into account was the number of games played. There are players like Svidler and Nakamura, who haven’t played each other too many times, but others like Anand and Topalov have played each other many times (including a World Championship match in 2010). Another way to do this was to take the percentage score, and if the score was above a particular threshold (e.g. 65%), then I could predict a win. That would take a lot more calculations however.

Most Recent Result

This method is slightly simpler. I have taken the last result between the two players in classical, for white and black individually, and that is the prediction I have made here.

This method has the advantage that it is largely based on most recent form. This could be why Giri finishes top in this prediction instead of bottom; given his recent rise in ranking and his invitations to more top events. Topalov on the other hand, with his recent poor run of form, is also ranked lower. It is a surprise to me that Nakamura is ranked lower using this prediction; however this does include his last classical game against Svidler as Black, a loss from 6 years ago (if that was predicted as a draw, he would move up to 5^th place).

My Intuition

Ok, this isn’t too statistically based, but here are my observations and predictions:

Giri’s solid style means that he is unlikely to lose many games, so if he can win games with white, then he stands with a good chance of winning. However, with this being his first Candidates, and players winning previously with +3 scores, he has to be careful not to take too many draws. He has also not won an elite event since the 2011/12 Reggio Emilia tournament over 4 years ago, so winning the Candidates is likely to be a big challenge for him. Aronian also has shown that he still has what it takes to win big tournaments, after winning the Sinquefield Cup last year. These two also finished 2^nd and 3^rd in the Grand Chess Tour, behind Carlsen, showing that they are two strong contenders in Moscow.

Nakamura is of course as dangerous as ever, winning the Gibraltar Chess Festival and the Zurich Chess Challenge. He certainly will be hungry for success, with the World Championship match being held in the USA; he will be eager to compete in front of a home crowd. His countryman, Caruana, also will be looking to do well, but he didn’t quite show the form last year that sent him towards 7 straight wins and a 3080 performance in the 2014 Sinquefield Cup. But he did show flashes of excellence, in Dortmund ending the tournament with 5 straight wins. He too will be looking for a great performance.

The old guard, Anand and Topalov, will be looking to show that at 46 and 40 years old respectively, they aren’t to be written off. Topalov got 2015 off to a good start, winning Norway Chess, but has slipped a little since then, finishing bottom in London. Anand has had disappointing results in St. Louis and London, as well as a below par performance in the World Rapid & Blitz and the Gibraltar Chess Festival. But he finished 2^nd to Nakamura in the Zurich Chess Challenge, losing out on tiebreaks, so he is not to be written off yet!

And then there is the Russian pair of Svidler and Karjakin. Both qualified after their blunder-filled final in the World Cup, after Karjakin recovered from being 2-0 down with 2 to play, to win the final. Both of these players played in the previous Candidates tournament in 2014 (Svidler also played in 2013), and have done well (Svidler was 3^rd in 2013, Karjakin 2^nd in 2014). These two are certainly the underdogs in the tournament, and although they have had previous experience in the Candidates, and the tournament being on home soil this year, they will have to play a great tournament to get past the favourites, to stand any chance of winning.

So, here are my final standing predictions:

I think that Aronian, despite being the wildcard choice for this event, will do well, given his recent performances and experience at the top level. Nakamura has hit some great form in the past year, and I feel that he has the performance to win this tournament. His exciting play, along with Caruana, might give them an edge on the second tiebreak (number of wins). On the other hand, Topalov seems to have hit a wall, with disappointing results in the last two Grand Chess Tour tournaments, while Anand, despite having great experience in these types of tournaments (he won the 2007 World Championship held in this format, and also played the last Championship match to be played in the USA, in 1995 losing against Garry Kasparov), will have to play some superb chess to come through in this young field.

And finally, a few random facts:

• In 2015, Anish Giri only lost 9 out of his 134 classical games last year (6.7%), with an overall record of +36-9=89; drawing nearly two-thirds of his games and scoring 60% overall.
• Since the start of 2015, Hikaru Nakamura has won all but one of the 6 tournaments that he played in, that Magnus Carlsen didn’t play (the only exception being the 2015 World Cup).
• The average age at this year’s Candidates is 32. Giri is the youngest at 21, and Anand is the oldest at 46. Four of the players are below 30.
• Caruana, Giri and Nakamura will be playing their first Candidates this year. The most experienced players in the field are Aronian, Topalov and Anand; this will be their 5^th Candidates tournament.

So, what are your predictions? Post them in the comments below, and we’ll see come the 30^th March, who predicted the winner (if anyone gets it right!).