|
Grischuk also suggested doing the tie breaks first.
I'm baffled Magnus offered the draw, but I'm not going to criticise the world champion for doing what he feels is best to protect his title. The way this championship has gone, FIDE need to seriously consider a change of format to prevent this being a thing in the future. They need something to structurally disincentivise draws. The easiest change up is to do the tie breaks first... but then the guy who wins is going to play SUPER defensively.
Bobby Fischer maybe had the right idea to just keep playing until one side wins, or set an arbitrary limit and after that point the champion retains on a draw. If Magnus knew that all he had to look forward to was another game with Caruana as white, maybe he'd have played that possibly winning position out to its conclusion, and maybe won it for realsies.
|
Yeha the 'show' of the Chess World Championship is not something for the players, its for the fans. The players are playing as a test of who is best, nothing more, and should be able to do as they see fit within the rules. Whether its part of some greater strategy or simply because it was down to nerves (Magnus has said on many occasions that he's nervous and different this year) people should give Magnus a break.
The players have absolutely no responsibility to play up to the show aspect of this, and never have in chess. Its not something you get into to win fans.
Shit look at the insanity of Fischer's World Championships, and how he was psychologically effected by it.
|
On November 27 2018 18:25 xM(Z wrote:
no one owes the viewers anything, be it in football or chess. competitors duke it out based on rules and you, the viewer, are the incidental at their show. if a competitor wants to give you the finger, you can do nothing about it. if it's not infringing on some rules, your whining and bitching amounts to nothing.
the only thing you might think you have on them is the money you directly or indirectly support them with, and if they tell you 'fuck your money', you're done; you indignation has no power(moral or legal) to do anything.
you are not the chess, so when you're sad, the chess doesn't care. stop being offended because the world doesn't spin based on your assumptions.
(once you get that, you can start talking about a middle ground(as Kramnik?) else just get out because in any negotiations here, you'll kill the game for your entertainment)
I don't completely agree. Sports are made by the people that are invested in it. That is not just the people playing the sport, but also the administration body and the viewers. Soccer finally had their VAR implemented because it became so necessary, I'm pretty sure the fans were most vocal about this.
Drama/story tends to be what most fans thrive on, though, and I agree it's often times overblown and you don't always get what you want, but knowing the highs your sport can deliver and then having snoozefest after snoozefest is definitely a blow to the face. You're being insulted for following the sport because the excitement (or depth, or creativity) is what once drew you to it.
I heavily encourage the world championships to have no outsider influence at all. No books, no supercomputers, no contact with outsiders. Only the player and the coach get to prepare, based on their innate knowledge of the game. It will result in more human like games, will rely more on intuition and will give more unique (identity based) approaches to the game. The team that has the best knowledge/intuition/creativity will then probably be able to prevail in the 12 games of classical time format.
|
It's been a weird match. I said they were both playing inconsistently a few pages back, but I don't think that was the best phrasing. They're not playing inconsistently or badly overall, it's just that no player has managed to string together an entire inspired game with no low points (except arguably Caruana game 6, but that just barely remained a draw/inhuman win). All in all, they've played quality chess with very few blunders and very tenacious defense by both (I excepted that from Fabi, but Magnus hasn't always been the most resilient defender in the past). So you can't really fault either of them for their overall play, but every time someone got real winning chances, they inevitably went astray. It's like the weakest part of both their play this match has come from when they were pressing. Maybe just nerves, maybe a too risk-averse approach, who knows.
This is also why I find the 'how do we fix chess' discussions a little misplaced. Yes, the absurd amount of theory, the influence of computers. all those things are very serious issues for modern chess, but I really don't think this match showcased that especially. There were 3 games where one player was just flat-out winning (1, 6, 8), 1 gigantic mess that magically turned into a draw (10), an extremely promising game spoiled by a draw offer (12) and several more that weren't just dull draws. This match could've easily been extremely wild if they'd just played a little better in promising positions or a little worse when defending. Just imagine how different everything would've been if Carlsen had won game 1 (and it's kinda absurd he didn't).
The tiebreak first idea is interesting but probably a little too counter-intuitive to ever actually be implemented. You could always go with chess tradition and go back to tie = reigning champions wins, but I'm not a big fan of that. I generally just hope we can go back to at least 16 game matches - 12 feels incredibly short and several players including Carlsen and Caruana have said they'd prefer longer matches. Whatever is done though, I don't think this match showcased any new problems in chess we weren't already aware of. If anything it was a lot more lively and interesting chess-wise (though not in drama) than Carlsen-Karjakin.
|
seems the problem is computers, they are getting closer and closer to solving chess. there are 7 piece tablebases now i believe, and openings are completely solved for the first 15 moves or so? so its just lame, the game at the highest level at that point is just a boring memorization. there's no chess involved, just boring preparation crap
i was reading this book modern chess move by move, and i recall how even well into a dozen moves, the author would say - oh yeah this position has been played before in these games - these other people tried x but its since been refuted and proven equal etc etc.
time goes on, there will be fewer "novelties", and fewer novelties that will give any real advantage to these chess gods. i'm sure chess960 will be solved eventually too, but fischer random chess seems a good way to counter this lame memorization crap. its the closest way to introduce a new balance patch to chess to interrupt the stale centuries old meta, without changing the core game.
that being said, the 12 draws weren't bad. some of the games were actually exciting. but perfect chess results in a draw. and these are the #1 and #2 players in the world rating wise, so even when one person had an edge the other person got the draw. still game 12 didn't make sense since magnus had better position and fabi had huge time pressure. huge upside no downside, he could always offer the draw after fabi reached time control.
|
what is "openings are solved for the first 15 moves" even supposed to mean?
|
it means the best moves are known for every opening for the first dozen plus moves or so and its known which ones are known to be dead ends for white (completely equal). it also means that all these games between GM are not diverging from games that have already been played before, for the first dozen plus moves
|
United Kingdom13775 Posts
Wouldn’t say that chess is anything close to solved - computers are quite good at it but they’re still inching upward over the years. The existence of opening tablebases doesn’t change that. Don’t think that that really has too much bearing on human games - the current top engines are already significantly better than any human ever will be.
Wonder if anything will come of the deep learning machines though - so far all we really have is one case of fraudulent drive-by PR, and a decent but rather flawed open source engine. It’s certainly an interesting idea.
|
Claiming that openings are "completely solved for the first 15 moves" is putting it a bit strongly. Some openings (like mainline Grunfelds) are theory for much longer than that in fact, but top players are still routinely surprised by novelties or near novelties much earlier than that (several times in this world championship in fact). And of course theory and our understanding of which lines are best does still evolve reasonably quickly.
As for endgame tablebases (which are perfect unlike opening books) they are a dead-end when it comes to "solving chess" given that large ones are completely computationally infeasible.
I doubt we'll even see an 8-man tablebase anytime soon. 7-man tablebases took five months to calculate on machines with 1 TB of RAM, and take 16.7 TB of disk (heavily compressed). Based on that and previous tablebases, an 8-man tablebase would take over a PB of disk, and at least 50 TB of RAM (since the retrograde analysis algorithm used requires all the states to be in memory at once). This is possible to be fair if someone was willing to dedicate months of supercomputer time to it.
|
I think that chess for human play will never be stale. Google's AI beat Stockfish in 50% of the games played as white. The best players can't even dream about beating Stockfish, and it losing so convincingly shows how high the ceiling is. There's just way way too many for even all the people in the world working together to memorize, that chess will always be a game of tactics.
I think that even if full 32 piece tables were solved, it would hardly change the game for that reason... If on turn 12 your opponent plays a non-optimal move you wouldnt know what to do, because it would branch out into one of 10 trillion possible lines you couldn't have remembered. The best that humans can do is learn some openings, some possible lines for the first few moves, learn end games, and thus know which simplified positions will be winning for them... The rest is experience of what positions lead to threats, calculating lines, and tactics from situations experienced in the past.
|
On November 28 2018 15:50 FiWiFaKi wrote:
I think that chess for human play will never be stale. Google's AI beat Stockfish in 50% of the games played as white. The best players can't even dream about beating Stockfish, and it losing so convincingly shows how high the ceiling is. There's just way way too many for even all the people in the world working together to memorize, that chess will always be a game of tactics.
Google played an inferior version on short time control, with no opening or end game bases. There's a fair bit of controversy as to whether stockfish would crush Google without its handicaps.
I do agree with the rest of your post though!
|
On November 28 2018 15:50 FiWiFaKi wrote:
I think that chess for human play will never be stale. Google's AI beat Stockfish in 50% of the games played as white. The best players can't even dream about beating Stockfish, and it losing so convincingly shows how high the ceiling is. There's just way way too many for even all the people in the world working together to memorize, that chess will always be a game of tactics.
I think that even if full 32 piece tables were solved, it would hardly change the game for that reason... If on turn 12 your opponent plays a non-optimal move you wouldnt know what to do, because it would branch out into one of 10 trillion possible lines you couldn't have remembered. The best that humans can do is learn some openings, some possible lines for the first few moves, learn end games, and thus know which simplified positions will be winning for them... The rest is experience of what positions lead to threats, calculating lines, and tactics from situations experienced in the past.
Big misconception about google AI vs Stockfish , Google had a mega super computer (which only they have by the way ) playing vs an I7 desktop stockfish , it had nothing to do with ceiling of Stockfish or chess , Alpha Zero didn't teach us anything.
The main problem with computers in today's chess (as I see it) is the mental aspect of the players , they are not part of the game in the first 15 moves or so , they are just playing the computer lines , and once their prep is over they are lost , their confidence drops as they KNOW they are going to play inaccurate moves and they get nervous , Magnus is only human , he sees how the rest of the world caught up to him and computers prep helped a lot , because the super GM`s rarely outright blunder the opening to give the opponent a real chance.
When you know you are not the best and there are entities that can pick a better move then you then its really a blow to your confidence , no longer you are the best chess player in the world , you might be the best human but that's feels less impactful when you know that the best players (computers) are helping your opponent.
The feeling that you are out of your prep and your opponent is playing the computer line is devastating , in the old days the worst case for you was that you are out of prep and some other human player suggested a line you are unfamiliar with , which might be good or not , now days , they KNOW that the computer has the best move suggested.
|
I think that's the most common critique I've read/heard. That players are just playing computer lines for 20-25 moves, then don't know what to do and just seek a draw to get out of the game alive. It's much about having an advantage position enough after 20-ish moves to either easily convert it to a win or bail out with a draw. I don't know how accurate that is but seeing how many games in Sinqfield(or something similar) Cup etc. were played, I can't help thinking the same in these 12 games. Perhaps it's confirmation bias.
|
I don't follow chess that much outside these World Championship matches, so this may be totally silly, but...
Would it be possible to play some experimental chess tournaments with football (soccer) kind of scoring system where you get 3 points for winning, 1 for draw and 0 for losing? This would reward players for seeking sharp positions and mastery of aggressive play.
Such system obviously doesn't help directly in 1 on 1 matches, especially with these long time controls, but if for example canditates was played favouring aggresive players, maybe some of that practise and study could shape the matches in the actual World Championship too. It's more of an attempt to shape the overall style of play towards more crowd pleasing mindset rather than addressing the current issue right here and right now.
Just to be clear, I do appreaciate the overwhelming finesse and complexity of chess and don't want to force in any kind of crowd pleaser just for the sake of it. However, right now even the top players seem to be rather disappointed by the outcomes and it's definitely a very tough and demanding competition to follow even for a dedicated casualish viewer.
|
On November 28 2018 19:47 Bacillus wrote:
I don't follow chess that much outside these World Championship matches, so this may be totally silly, but...
Would it be possible to play some experimental chess tournaments with football (soccer) kind of scoring system where you get 3 points for winning, 1 for draw and 0 for losing? This would reward players for seeking sharp positions and mastery of aggressive play.
Such system obviously doesn't help directly in 1 on 1 matches, especially with these long time controls, but if for example canditates was played favouring aggresive players, maybe some of that practise and study could shape the matches in the actual World Championship too. It's more of an attempt to shape the overall style of play towards more crowd pleasing mindset rather than addressing the current issue right here and right now.
Just to be clear, I do appreaciate the overwhelming finesse and complexity of chess and don't want to force in any kind of crowd pleaser just for the sake of it. However, right now even the top players seem to be rather disappointed by the outcomes and it's definitely a very tough and demanding competition to follow even for a dedicated casualish viewer.
It's definitely a reasonable suggestion. It's even been tried before. The issue from what I remember was that the results were kind of disappointing, i.e. there wasn't any noticeable change in players' willingness to go for speculative positions and no change to the draw %. Partly I think that's due to the nature of chess - games will frequently fizzle out even if both players are actively looking to make things interesting and spicing a game up at all costs is often going to be -EV even if you get 3 points for a win. The extra point does not make up for playing dubious moves. On the other hand, you could also argue that the players simply hadn't fully adjusted to the new system yet, i.e. that they were still stuck in their classical mindset and not milking the 3 points for a win as much as they could.
|
On November 28 2018 19:47 Bacillus wrote:
I don't follow chess that much outside these World Championship matches, so this may be totally silly, but...
Would it be possible to play some experimental chess tournaments with football (soccer) kind of scoring system where you get 3 points for winning, 1 for draw and 0 for losing? This would reward players for seeking sharp positions and mastery of aggressive play.
Such system obviously doesn't help directly in 1 on 1 matches, especially with these long time controls, but if for example canditates was played favouring aggresive players, maybe some of that practise and study could shape the matches in the actual World Championship too. It's more of an attempt to shape the overall style of play towards more crowd pleasing mindset rather than addressing the current issue right here and right now.
Just to be clear, I do appreaciate the overwhelming finesse and complexity of chess and don't want to force in any kind of crowd pleaser just for the sake of it. However, right now even the top players seem to be rather disappointed by the outcomes and it's definitely a very tough and demanding competition to follow even for a dedicated casualish viewer.
Such a scoring system has been used in the recent past. If I remember correctly it didn't change much. I wonder how tournaments would run if the point difference between a win and a draw were much larger (5 to 1 or even 10 to 1). Might still not change much, as a draw is also about preventing your opponent from getting a win.
1vs1 matches are easier to fix: make draws not count. Play a best of 5 or 7 where only decisive results count. The match might take longer, but the match strategy will at least change from "don't lose" to "must win".
|
United Kingdom13775 Posts
On November 28 2018 15:50 FiWiFaKi wrote:
I think that chess for human play will never be stale. Google's AI beat Stockfish in 50% of the games played as white.
Google used a supercomputer to beat a deliberately crippled version of Stockfish and has spurned just about every request for providing useful insight into their program to allow something resembling a scientific peer review to happen. That's not winning, that's drive-by PR fraud. I bet what happened is that they tried Stockfish under "fair" conditions and got their ass handed to them. So they did what AI researchers tend to do under situations like that: if you can't prove your method is better, then you just change the baseline so that you can pretend it's better.
It's easy to feed on the fact that most people know little to nothing about the inner workings of chess engines to be able to pretend to beat Stockfish. As far as I'm concerned, until they play a real showmatch under mutually agreed upon "fair" conditions, they haven't beaten anything.
|
Doesnt look good for Fabiano...
|
Magnus what happened to you?
|
|
|
|
|
|
|
|
EPT
