Krdytkyu

Any online search about the recent game 12 in Carlsen-Caruana's WCC is filled with the surprise caused by Carlsen offering a draw when in a (possibly?) winning position. It was announced on chess.com that

the top eight chess engines in the CCC will play a 2x-round-robin tournament starting after move 31 of the world chess championship game 12. The time control is rapid chess, 30 minutes plus five-second increment.

At the moment of writing this, I can see the tournament here. But I don't know enough about chess or at least how chess.com works to understand what's happening.

Is the tournament leading to some consensus on how good/bad Carlsen's decision was?

Is the tournament leading to some consensus on how good/bad Carlsen's
decision was?

No. Consensus on Carlsen's play and decision has already been reached, I would suggest.

Psychologically Carlsen made it clear in the post match interview that his goal before this game was a draw to reach the rapid playoff where he thought (correctly) that he was strong favourite to win. Hence his play was aimed at taking no risks and inviting a draw at every opportunity. For instance, he offered to repeat moves around about move 14 or 15. In so far as he practically forced Caruana to accept the draw and then went on to win the playoff 3-0 his decision was a good one.

From a chess point of view the consensus is that a few moves earlier Carlsen had a winning advantage, say after white's 29. Re1, and that Carlsen made quiet moves rather than the aggressive moves he needed to make to win.

While the game was in progress a very powerful computer, sesse, was analysing the positions and showing its analysis and evaluations.

This computer tournament adds nothing to the judgement as to how good or bad Carlsen's decision was.

Is the tournament leading to some consensus on how good/bad Carlsen's decision was?

Not quite. As others have pointed out, Carlsen's decision was based on factors outside that one game. With a stronger position and a large time advantage, Carlsen most likely could have won game 12, but Caruana had just tied Carlsen in 11 consecutive games, several of which had each player blunder away a significant advantage. Caruana showed he could go toe-to-toe with our Norwegian champion, and so attacking his position would not come without risks.

Meanwhile, a draw effectively changed the championship to a 4-game Rapid tournament (then Blitz if that ended in a tie), where Carlsen is a heavy favorite. While Carlsen and Caruana are roughly equal in standard chess (2835 and 2832 respectively), Carlsen is much stronger at Rapid (2880 and 2789) and Blitz (2939 and 2767). Why extend a game in an even format when a draw moves to a format you're stronger in?

Having said that, the Chess community was not happy at all to see that draw offered (source: I was unhappy). I'll concede it was a technically correct decision, but as someone who had been waiting 11 games for a win, seeing one materialize then simply vanish was frustrating. It left a lot of people asking "What would have happened if Carlsen hadn't offered that draw? How would the game have ended?", and that's the question these games are answering.

In other words, the tournament isn't leading to some consensus on how good/bad Carlsen's decision was. It's leading to some consensus on how good/bad Carlsen's position was.

All in all, the tournament showed that Carlsen had a very good position. After 56 games, black finished with 27 wins, white finished with 3 wins, and there were 26 draws. That's a tremendous advantage. Any grandmaster would be ecstatic to hear that, for a chess engine, their position was only losing in 5% of games played.

What's more, this tournament had given both sides equal time. That's not completely accurate, as Carlsen had a significant lead on time (though I've forgotten just how much). That implies these results are even too conservative, and that Carlsen may have had a larger lead.

Having said that, humans and computers play in very different styles. Because it's winning for a computer doesn't guarantee it's winning for Carlsen (though he's practically a computer anyway), but it does show he had a strong opportunity to win it all in game 12.

It will not show how good or bad Carlsen's decision was - humans play
and assess differently from computers; there are positions that are extremely
easy for computers, but terribly difficult for humans and vice versa.

So a computer evaluation of ~+1 might not give any chance for a human win
and in many cases a position with equal computer evaluation is an easy win
for one side in a human game, most often difficult endgame defence, that
computer never gets tired and plays perfectly and gets a draw, but human
can't execute it with tablebase precision.

I think Carlsen did Very OK, last game before tie-breaks, draw with
black pieces - and he won the tie-break - that's the best proof that he
was right :)

The classic game rating of Carlsen and Caruana is very close (3 points difference!). But rapid game rating is another story. The rating point difference is a whooping 100 points. Carlsen's strategy throughout this whole tournament was along these lines - beat me if you can in classic. If you don't, I will sweep you off in the rapid games. And he did exactly that. Sure, Caruana is a quality player; but playing under time pressure, not to mention pressure in playing for World Championshiop, is a totally different thing.

So, Carlsen has no need to risk anything; (yet, it was Caruana who looked more happy after draw was agreed in game 12. This is quite understandable since Carlsen had the advantage in the game.)

As Lord Farquaad points out, the tournament will only show how good Carlsen's position was, not how good his decision was. That said, his position was pretty damned good, if the computers can be believed: so far 49 games have been played, and the score is 23 Black wins to 3 White wins, with 23 draws.

The engine tournament is absolutely pointless. Engines don't make mistakes and they are far better at pushing in advantageous positions. Even then black still managed to lose games.

Carlsen's decision was absolutely the right one for the highest probability of holding the title. Realistically in a GM game black probably had around a 20% chance of winning and a 10% chance of losing at best. Instead he goes for the tiebreakers where he as a 75%+ chance of winning that with his elo advantage.