This game is quite remarkable, because not only does MC manage to swindle Jones, it is also important for theory. It discusses what I regard as the critical line of the whole system and it seems that Jones had prepared an important improvement over the line given in his book. I added some quick analysis by Stockfish, in which direction this may be heading.

The cat is out of the bag. Now we know why the players from China are going for the early Bf5. Drummroll please for the variation that curtains the mainline of the Petroff:

The biggest advantage of the Petroff is that it not only dodges the discussion in the Berlin, it can also dodge the Italian, the Scotch and the Four-Knights, although black has to go slightly offbeat to do so.

The following game may or may not lead to another renaissance of the French Winawer, but this variation obviously got the stamp of approval from the World Champion, at least for one game against a known theoretician. For convenience I also added the known draws.

On a sidenote: Kramnik scored another full point against his former second Svidler, this time in 24 moves. For some very weird reason former seconds don’t seem to be able to cope with the pressure of meeting their former boss and fail to recall their home analysis, their playing strength, their resilience and pretty much everything else.

In this position Inarkiev gave check on e3 to which Carlsen replied with Kd3. Inarkiev immediately called the arbiter and claimed an illegal move by Carlsen and got awarded the win. This is the oldest trick in blitz and it shows how low it can get in professional sports these days. My deepest respect goes to Carlsen for staying calm, because I have already seen fights over unsportsmanlike behavior like that.

Here is the incident:

Source: Youtube

And here is the lucky winner talking about it with a grin on his face:

Source: Youtube

Finally the good news: The arbiters looked at the incident again and offered Inarkiev to continue the game from the final position. He declined to play on and got his well deserved zero. That’s an amazing turn-around and two thumbs up for the arbiters!

You can say a lot of things about Anand, but he never has a bad hair-day and he certainly knows how to play tournaments. It turned out that playing Stock Exchange Chess can be enough to win. Obviously it is easier said than done, but he certainly executed it to perfection.

Btw, has anyone seen Gelfand? I wonder why he didn’t participate.

This is why you shouldn’t bet on chess or maybe you should. Anyways, Volkov ends up in a lost position in his pet-line after 14 moves against Vitiugov and Svidler beats Malakhov in a totally one-sided affair. That led to a tiebreak between Svidler and his friend and second from the candidates Vitiugov. Svidler won the first game with black and then Vitiugov basically threw the second game. Surprisingly or not Svidler became the new Russian Champion, for the 8th time. Meanwhile Fedoseev and Dubov shared third place.

Here are the most memorable games:

 

Source: Youtube

Going into this round Fed was the sole leader with seven points. Svidler was one point behind and his buddy Vitiugov even had a chance to take the lead if he beats Dubov and Svidler wins. So basically Svidler had every reason to play a complicated game. The only problem is that this is incompatible with his free-roll style. Playing unclear positions where he doesn’t have a forced draw in his back-pocket is simply not part of his repertoire. Obviously that’s only half of the story, because the Fed self-destructs in this game for absolutely no reason. Svidler didn’t win the game, Fedoseev lost it. Maybe losing in rounds 6 and 7 had something to do with this.

Apparently AlphaZero has a predecessor named Giraffe. The developer Matthew Lai now works for DeepMind. Surprise?

The paper on Giraffe explains everything in much greater detail. For instance it took 72 hours to calibrate the net on a workstation with 2×10-core Intel Xeon E5-2660 v2 CPU. If we use a rough shortcut and divide 72 by 4, then Google’s high-end cluster with 64 TPUs did the job 18 times faster.

In 2016 Giraffe peaked at Elo 2410 in engine competitions, which is remarkably weak as even the current version of the good old GNU Chess is rated Elo 2800+. Given the initial results, the decision to keep investing in the idea is quite remarkable too.

As I wrote in the previous article, my Stockfish, casually running on just one core on an Intel i7 4760 3.60 GHz, took roughly 75 minutes to find the star move at depth 41. Hardware is the bottleneck. Just for comparison: Massive hardware upgrades almost doubled the playing-strength of AlphaGo. It simply expands the search-horizon.

Looking at the difference in pure hardware power this reminds me of David vs. Goliath. Running on identical machines, Stockfish should beat AlphaZero easily. Drawing 72 games with such a handicap is actually amazing. Let’s not kid ourselves, Stockfish on 4 TPUs would beat Stockfish running on much weaker hardware convincingly too.

One thing is clear: It will take a few CPU-generations until mere mortals like you and me will be able to run AlphaZero at home.

 

Update (21.12.2018): The results by Leela are so encouraging, there may not be a need to run AlphaZero at home anymore.

Fourty years ago the New Yorker published an interview with the math professor Paul Magriel aka X-22 who became Backgammon World Champion shortly after. In order to develop his tournament strategy he did the following:

“I used to play backgammon against myself,” he said, “and once I had a private tournament with sixty-four imaginary entrants, whom I designated X-l, X-2, and so forth, through X-64. In the final, X-22 was pitted against X-34, and X-22 won.”

Source: New Yorker, “Playing X-22”, 5th of December 1977.

According to this paper, AlphaZero pretty much did the same:

In AlphaGo Zero, self-play games were generated by the best player from all previous iterations. After each iteration of training, the performance of the new player was measured against the best player; if it won by a margin of 55% then it replaced the best player and self-play games were subsequently generated by this new player.

This sounds rather easy in theory, but it’s not that easy to code. While Magriel could make the deliberate decision to play for certain points or use the cube in a certain way, AlphaZero modifies each player based on what? There is certain difference in style between Tal and Petrosian, but how do you formulate this in numbers? In other words, it’s not easy to describe a style in a formal language or as an object. Stockfish is much easier to configure, because you can just give weights to certain positional features and you can modify the value of pieces. I guess the solution to this problem is worth the 400 million dollars that Google paid for DeepMind in 2014.

Initially I thought that there is a pretty good chance that the whole story is just a scam, like the match Slyusarchuk vs. Rybka. There is even an incentive for manipulation. Just check out how the stock market reacted to the annoucement.

After I saw the following game, I could pretty much exclude all of that. The move 21. Bg5 is simply an amazing bolt out of the blue that basically wins on the spot. It takes Stockfish over an hour to evaluate the move correctly at depth 41. The idea is hidden so well, that it could easily qualify as preparation for a World Championship match.

Well, it seems like the new age of chess has arrived. Obviously even AlphaZero cannot improve upon forced draws, so this project is still quite safe. Nevertheless I wouldn’t be surprised if the american players get access to AlphaZero opening analysis for the candidates and it could even lead to an american world champion in this cycle.