Magnus Carlsen Invitational: Carlsen vs Nakamura Live!

"Chess960 is young, and some of its lack of draws can be attributed to its theory being young, but don't expect that to last forever."

Of course it will last forever, a human cannot memorise the theory for more than 2-3 starting positions/960 in a lifetime. If you try to spread out that same amount memorisation over 960 positions, it is so minimal for each position as to be useless. So the draw rate would be expected to be lower thaLen chess indefinitely. What that rate is, I do not know, but I am not against draws in chess in and of itself. There are 2 main problems, one is the kind of draw: players who play 25 moves of theory and then agree to a draw. They did not actually play each other, they played against their computers at home. And the other problem is when the rate gets too high and you cannot determine the outcome of a match (like Carlsen-Caruana 2018 world championship) without resorting to rapid, blitz, or armageddon. And a third problem are the kind of sterile games that Leavenfish attributes to computer chess.

@lajosarpad, I do not see how you can be in favor of a coin toss unless you know that each side has a 50% chance of winning the game, which is not known because no one has analysed the data to know how much time black should have for the draw odds. So you are advocating one side getting a huge advantage at the most critical portion of the match!

"Of course it's knowledge, and in Chess960 it will be useful 1/960 of the time. Your argument that it is "arbitrary" could be equally applied to any rule of the game, be it the way pieces move and capture, the stalemate rule, the castling rules, en passant, etc."

I have no problem with changing the way pieces move or capture (GM Seirawan and others have done this), nor the stalemate rule, castling, en passant, etc. The point is that any of these changes can be made and a player can translate their ability from chess to the new game quite easily. The knowledge is useful. But change the starting setup and all the knowledge memorised is NOT useful. No less an authority than GM Kasparov made this point. You mentioned shogi and there are other forms of chess (chinese, korean, etc). I have played around 10 games and won most of them, against ppl from those countries who have played many more games than me. How is it possible? Because there was a transference from chess knowledge. But I guarantee that memorising chess opening moves did not help me a bit.

"In Chess960 you can play zero moves and find yourself in a bad position, because many of the setups are imbalanced."

Of the 960 starting positions, more than 2/3 of them are LESS imbalanced than the starting position in chess. So if that is something to be concerned about, it is an argument against chess, and in favor of one of these other positions. But it does not bother me, because it evens out in the long run. And one does not have to memorise reams of information to avoid getting a bad position, like in chess. That was the point. How many hours/day do top players spend memorising openings?

I am not sure anyone is arguing for 'more interesting' for more interestings sake. The idea is to get a decision after a given match is tied and hopefully via a method most similar to the chess that is actually being played by the contestants. A coin toss (or roulette wheel spin - aka Smyslov - Huebner Candidates) does not do that. This after the attempt at doing so via more 'boring, precise' games have yielded no decision.

If adding more errors into the game makes it more interesting, then let's focus on novice play only. Plenty of errors there, so it must be pretty interesting. As about armageddon, that's forcing a decision where none was achieved via conventional means. If an armageddon is needed, then my preference is a coin toss. It's quick, honest and there are no mistakes. I'm of those masochistic types who is more interested in precise chess games than the more interesting faulty games some of us regards to be more interesting here. As about chess960 I'm not very interested in that. I'm interested in the boring, precise game.

Fgkdilkag:
"By continuing with chess, there is an entire universe of pawn structures and middlegame positions of which we deprive ourselves."

Jeh:
That's a fair argument. But like I said earlier, there are some pleasures that come from having one starting position that I would feel deprived of, too.

I do think you're attributing too many of the draws in chess to opening knowledge. The reality is that chess is systemically drawish.


Let me just weigh in with two points.
1. Non-standard starting positions also leave most spectators without a rudder in the ensuing journey…don’t like that. It essentially becomes a different game despite with different starting positions. Deciding with an essentially different game does not set well.
2. Yes, chess is systematically drawish…a ‘correctly played game’ will usually end in a draw. The most reasonable way to decide a drawn match seems to be to induce more errors in play as naturally as possible in the deciders…and speeding up the game does this. No 'weirdness' required. I personally like dealing with the ‘white to serve’ edge by giving black a chance to win the tiebreaks early from his disadvantaged point…induces psychological pressure on White, but YMMV…

Jeh – I know some disagree with me about the ‘sterility’ of Computer chess. Sure, engines can lead to innovations players can use – but most of those are just as a result of humans pushing engines through their paces in certain opening/middlegame positions to find an edge here they can surprise a player with OTB to try to get an edge. That’s their main use to a chess player. Actual engine vs engine ‘play’ is like a mathematical exercise…sterile in that an engine simply counts out the moves based on an algorithm a programmer has supplied. Without a human playing…it’s like two bots following pre-programmed instructions to generate a move…no real ‘planning’ (watch them play without starting positions) few calculation errors. New Neural Network engines are more interesting though. But we can disagree on engines…I just can’t find engine vs engine game play interesting. Human vs Human is a totally different animal.

One tiebreak system I always thought would be worth an experiment is CUTTHROAT. We used to play this all the time [on analog clocks, no less] and it is a lot of fun. Today you could add a second or two per move, or even a delay ... Anyway, what we did was put 5 minutes on each clock and play. Whoever wins gets one minute less for the next game [ie, 4 to 5 in game 2), and so on until you win the game in which you start with 1 minute. If a game is drawn, the times remain the same, but you switch colors every game regardless. Fairly well negates who gets white first.
Massive excitement at the club!

"I would not call opening theory "knowledge". The fact that all this information becomes useless by switching a couple pieces in the starting position shows it is not particularly valuable information, and actually very arbitrary."

Of course it's knowledge, and in Chess960 it will be useful 1/960 of the time. Your argument that it is "arbitrary" could be equally applied to any rule of the game, be it the way pieces move and capture, the stalemate rule, the castling rules, en passant, etc.

"Today there are plenty of top games where a player has forgotten some preparation and ends up in a very bad position. "

In Chess960 you can play zero moves and find yourself in a bad position, because many of the setups are imbalanced.

"By continuing with chess, there is an entire universe of pawn structures and middlegame positions of which we deprive ourselves."

That's a fair argument. But like I said earlier, there are some pleasures that come from having one starting position that I would feel deprived of, too.

I do think you're attributing too many of the draws in chess to opening knowledge. The reality is that chess is systemically drawish. Compare it to Shogi, where like chess, there's only one starting position and a huge body of opening theory, but unlike chess the draw rate is no more than 2%. (The reason is because Shogi's rule that captured pieces can return to play maintains the complexity of the position, whereas in chess the trading of pieces reduces the complexity.) Chess960 is young, and some of its lack of draws can be attributed to its theory being young, but don't expect that to last forever.

@Keith Homeyard, it's rather ironic that you bring up chess and Star Trek because Kirk and Spock are seeing playing a three-dimensional chess. They could not be convinced to keep the original game and starting configuration... :)

"given that one eventually always has to play a hitherto unseen position"
For the amateur player, yes, but at the top level we have seen a plethora of games that are agreed as draws a few moves outside of established theory.

"It throws away over a century of knowledge and tradition"
It does not throw away anything, people will still study the great games of the past and the legendary battles. I would not call opening theory "knowledge". The fact that all this information becomes useless by switching a couple pieces in the starting position shows it is not particularly valuable information, and actually very arbitrary. Chess has always evolved, and I am afraid that by holding on to what has become the standard starting position will lead to the stagnation and decline of chess. It gets worse. every decade. I have heard more chess trainers recently suggest their players go elsewhere for opening theory because they are not interested in it. Draws are more of a problem today than 50 years ago and in 50 years it will be a bigger problem. Top players today spend a higher percentage of time on openings than at any point in the past.

Today there are plenty of top games where a player has forgotten some preparation and ends up in a very bad position. Chess has become a memory contest. No one disputes this. But 100 years ago, the top players had to use creativity and problem-solving much more.

I almost exclusively play 960 these days and frequently get into middlegame positions that I have not seen anything like and never imagined I would see in a lifetime of playing chess. The ratio was the opposite in chess, it is very unusual to get into a novel middlegame position. By continuing with chess, there is an entire universe of pawn structures and middlegame positions of which we deprive ourselves. Plus there is a totally different feeling of the pieces. No longer are the queen and rooks relegated to doing nothing for the opening moves.

Another thing Chess960 does is presume, without justification, that repetition makes the game less interesting. But how many of us have our own "pet" openings that we are happy to let define us and our friends? How many of us like playing the same systems against our friends multiple times, each time taking delight not in a game that was different start-to-finish, but in the very fact that it started out similarly but evolved differently?

Leavenfish, I agree with you about Chess960. It throws away over a century of knowledge and tradition and makes it more difficult to compare the play of grandmasters to one's own play. I personally don't see a problem with memorization, given that one eventually always has to play a hitherto unseen position and that memorization is a necessary feature of a lot of other human skills.

I also agree with you about the tiebreaks. While I think the Armageddon system could easily be improved by letting the players decide how much time black should get, I would still prefer continued pairs of games until someone wins. Just continue lowering the base time, like you said. It seems like tournament organizers are afraid of the pairs of games going on forever, which I find somewhat inexplicable.

But I do disagree with you about computer chess. It is anything but sterile and pointless. Like any chess game, computer vs. computer games are rich with ideas for humans to investigate by analyzing them. And computers still manage to defeat each other despite playing at superhuman levels. It's just that the errors made are far more subtle. Computers may have ripped the veil off the play of human grandmasters, but they haven't ripped the veil off each other, so I still find their games very interesting.

Stupido - That is what you get with shorter time controls (and arguably because they are unrated) BUT...I would argue that errors at this level make the games even more fascinating....human. 'Gamesmanship' comes more into play than exactitude - and we saw that in the final Ding - Magnus game among others. Errors great and small make chess fascinating. 'Computer chess' by contrast is sterile and pointless.

The only real problem I have it with the technology...a couple of late game disconnects have marred things.

fgkdilkag I agree with you (except for the 960 comment!). Not to argue for an impossible ‘perfect’ to replace the ‘good’, but I just think you hold the tampering to get a result to a minimum – difficult to do in a ONE game decider! As we are talking about blitz levels, why shy away from a multi-game solution? You do not need so quickly to introduce probabilities as the errors will likely decide.
Perhaps a mini-match of this sort where AND EARLY BLACK WIN DECIDES: G1: 0-1, match over (automatic pressure on White). 1-0 and you switch colors for G2 with G1 winner now having draw odds, White needing to win to equalize. If you get to G3, you shorten the TC further and continue that way with the first to win being the decider – or lower White’ time by (say 15%). But let the chess itself decide as much as possible. Manipulated probabilities plays out over a large sample sizes, not single games where a huge variance must exist.

The level of games is appalling. Not only today, the whole tourney is subpar.

Ding 2.5 - 1.5 Carlsen Just wait......

fgkdjlkag I understand your comment but then to misquote Spock in StarTrek " It's chess Jim, but not as we know it" :)

I believe you are right that FIDE has nothing to do with this tournament.
I'm not sure where you are getting these ideas regarding armageddon. The 3 solutions proposed below equalise the chances by giving each side (black and white) a 50% chance of winning. It is not gambling, nor arbitrary. Contrariwise, the method used in the tournament gives one side a decidedly large advantage. There is an absolute difference in the probabilities of winning between the unfair current method being used and the 3 proposals below. Are all armageddon systems imbalanced? Yes, since white has an advantage in a standard game, the only way to get a 50% chance of winning in a single game is to introduce draw odds for black and use the clock to adjust for the discrepancy.
IMO the way to substantiallly lower the probability of tied matches/armageddon is just to replace chess with chess960.

They are 'better', yes one could make that argument. It is a bit relative, but I simply think there are no truly 'good' solutions given that you are attempting to decide a match in ONE game by setting up some arbitrary 'imbalance' outside the board... forcing each side from the idea of 'correct play' to gambling...both on and off the board. Again though...FIDE has nothing to do with this tourney, right? I heard that somewhere so I might have that wrong.

@Leavenfish, there are plenty of good armageddon systems. A few have been mentioned below by RichardEaston, Jeh, and genem.

FIDE haters...(and there is always good reason to be!), I don't think FIDE has anything to do with this, do they? It's not even rated from what I understand.
Whatever the case, there is really no good armageddon system. But you have to decide somehow and TIME is the only real randomizer one can use and it remain traditional chess. Perhaps you simply rotate White/Black, offer White a 1 sec increment and Black a 2 sec increment...and the first player to win the blitz game, wins. Keeps it simple.

The two sides secretly bid how much time black gets (versus white) with each side getting an additional 2 seconds per move. The person bidding less time gets black.

This is not a FIDE rule in this case, it's the fault of the tournament organizers. Agree with others regarding the unfairness of the armageddon system used.

And this is why chess is not a sport. Silly rules made up by the corrupt FIDE organization. Keep it simple stupid!

Why have Amageddon here? Totally unnecessary. come on, the players are home, comfortable, there's no pressure for the organizers, etc. What's the purpose of ruining the tourney with such a dumb game.

@genem
Or, use the Pie Rule. One player chooses how much time Black gets, after which the other player chooses what color to play with.

@Green22 -- Victory of the match, not individual games.

How does Firouzja not have 1 point? I keep checking the standings and each day he has 0 points..

"If a players gets the victory without needing the sudden-death tiebreaker, he will get 3 points, while a win in the playoff will give the winner 2 points and the loser 1 point."

I agree with genem re bidding on the Armageddon game.

FIDE should ban Armageddon games where the Tournament Organizer pre-assigns the time amounts to White and Black. Instead, FIDE should require that players be allowed to BID for the amount of time they would want in exchange for the advantage of playing as Black (lowest time bid wins the right to play Black). There should be enough Armageddon games played by now to have enough data to see whether 5 minutes to 4 minutes is proving to be fair to both colors.

@PurpDriv2
Well, Grischuk is one of the official commentators for the tournament, so he must feel pretty good about it.

Does anyone know who is paying for the prize fund ?
Is it coming out of Magnus' pocket, is Chess24 paying or is there some other benefactor ?

So the arbiter sent Naka a message to abort his final game with Magnus...problems with the clock or some such. And clearly it broke Naka's concentration in deciding on Kf1 or Kh1 (he chose Kf1?!)...is this no different than an arbiter coming up to a player when it is his move in sudden death and talking to him? What would the decision be there? I suppose Naka should have protested then and there (or aborted) but....well, wonder what others think of this.

I think he should have aborted. If the arbiter is wrong in saying to do so then it's on him and you replay the game.

How would Grischuk Alexander and Wang Hao feel about this tournament.