Rustam Kasimdzhanov proposal

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 2:27 pm

Re: Rustam Kasimdzhanov proposal

Post by Don » Wed Jun 05, 2013 11:09 am

Albert Silver wrote:
Don wrote:
Albert Silver wrote:
I disagree that the matter is whether the relative strengths are the same. Let us suppose that you decided to take two marathon runners and test them at 100 meters. In theory, you can assume that as they are both similarly specialized, their handicap will also be similar. Let us even suppose it is true. So what? The only thing in common is that both are foot race events, but they are otherwise very different competitions. Even if two players perform similarly in blitz and long games relative to each other, the two events are simply completely different.
But that is not my point, my point is not that they will be both seriously handicapped in an equal way but that they will both be playing top notch chess for the time control itself.
What is the difference? I say they are similarly handicapped at 5 minutes, and you say they are both playing as well as possible. If they are similarly handicapped, they are still playing their best possible within those limitations, no?
My point is that they are still playing world class chess - at least world class for 5 minutes.

If your argument is that 5 minutes chess is lower quality than 4 hour chess then you get no argument from me. But if your argument is about quality then why not play 8 hours chess or 16 hours chess per side with breaks every 20 moves or so? I guarantee you that over the board chess does not compare to correspondence chess in quality.

It's for logistical reasons partly, but more to the point these are primarily a competition and a test of skill, not a format designed to create the highest quality games possible (which they don't even come close to doing.)

You are stuck in the stone age of think it has to be done a certain way because that is "tradition" and anything less than a game per day is against the laws of nature or something.

By the way, I don't necessarily agree that this should come down to 5 minute chess either. The FIRST game should be sped up to allow time for escalating speedups and that we should not suddenly go from 4 hours to 5 minute chess.

However, I think you seriously underestimate the difference between blitz and slow games. It might be because you are so accustomed to seeing machines only gain X plies and start playing 14 plies depth in blitz instead of 20 (or whatever).
Are you kidding? The difference in 5 minute chess and 4 hour chess is enormous, I am well aware of the difference. Even a couple of ply puts you in another class. The "scalability" if you want to call it that of humans is superior to machines also so that humans improve more with more time than computers do. The shape of the curve is very similar though, just steeper. In fact most humans don't realize that they play something like 100 ELO stronger for every doubling in time control. It depends on the level but it's basically whatever a computer gets plus somewhat more. Humans don't realize it generally because their opponents play stronger too so they don't notice that the quality of the game itself just jumped significantly, they just think they are playing "about the same." They DO notice how crappy they play in 5 minute chess though and think that it means THEY are particularly weak at it.


More mistakes accrue there is no question, but for humans the difference goes much deeper. Deep planning and counter-planning (fighting your opponent's ideas) go right out the window. In fact planning becomes general ideas at best.

I think you will find very very few grandmasters who think the two time controls are comparable. The vast majority of players are outraged at the idea of the world championship being decided by blitz tiebreak games.
Well then they need to get over it. It's nothing more than prejudice and old-school conservatism and fear of change. I don't know if you realize the prejudice over algebraic in the U.S. but it was just plain "wrong" too in the eyes of many players - but they got over it.

If this were incorporated into top level play it would be very quickly accepted as normal.

In tennis, you have the situation that a game, a set or a match could be decided by a single point - perhaps even a bad ball bounce. In chess a game can be decided by a single bad move. That's life and everything works that way in life and in games and having a single grueling long game does not change that as it will still come down to the skill of the players.

Albert
Capital punishment would be more effective as a preventive measure if it were administered prior to the crime.

tomgdrums
Posts: 736
Joined: Wed Dec 02, 2009 8:48 am

Re: Rustam Kasimdzhanov proposal

Post by tomgdrums » Wed Jun 05, 2013 11:41 am

Don,

I think the tennis analogy is not the best one to make as it is a continuation of the same game.

If a chess draw was then decided by a new blitz or rapid game then that is a DIFFERENT game being played to decide the outcome of the first game.

This is why (although exciting) sudden death goal shootouts in Hockey and Soccer are tough to watch. And seemingly less satisfying to me the observer. (unless I have an extreme rooting interest in the team that won)

I am not against draws! They are part of what makes chess, chess!

Someone who is losing who can fight their way to a draw is exciting.

However there are too many quick draws. I say they should somehow change the point system. (in fact isn't there a point system like this--)

User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 2:27 pm

Re: Rustam Kasimdzhanov proposal

Post by Don » Wed Jun 05, 2013 11:46 am

tomgdrums wrote:Don,

I think the tennis analogy is not the best one to make as it is a continuation of the same game.
But the proposal is to change the definition of a "game" so that it really becomes a mini-match.

If a chess draw was then decided by a new blitz or rapid game then that is a DIFFERENT game being played to decide the outcome of the first game.
But use your imagination, this is only semantics. You could say that a tie-breaker in tennis is a separate game too, or what about the second and third sets? Those are not related at all to the first set.

So the proposal is to expand the definition of what a "game" is.

This is why (although exciting) sudden death goal shootouts in Hockey and Soccer are tough to watch. And seemingly less satisfying to me the observer. (unless I have an extreme rooting interest in the team that won)

I am not against draws! They are part of what makes chess, chess!

Someone who is losing who can fight their way to a draw is exciting.

However there are too many quick draws. I say they should somehow change the point system. (in fact isn't there a point system like this--)
Capital punishment would be more effective as a preventive measure if it were administered prior to the crime.

tomgdrums
Posts: 736
Joined: Wed Dec 02, 2009 8:48 am

Re: Rustam Kasimdzhanov proposal

Post by tomgdrums » Wed Jun 05, 2013 1:15 pm

Don wrote:
tomgdrums wrote:Don,

I think the tennis analogy is not the best one to make as it is a continuation of the same game.
But the proposal is to change the definition of a "game" so that it really becomes a mini-match.

If a chess draw was then decided by a new blitz or rapid game then that is a DIFFERENT game being played to decide the outcome of the first game.
But use your imagination, this is only semantics. You could say that a tie-breaker in tennis is a separate game too, or what about the second and third sets? Those are not related at all to the first set.

So the proposal is to expand the definition of what a "game" is.

This is why (although exciting) sudden death goal shootouts in Hockey and Soccer are tough to watch. And seemingly less satisfying to me the observer. (unless I have an extreme rooting interest in the team that won)

I am not against draws! They are part of what makes chess, chess!

Someone who is losing who can fight their way to a draw is exciting.

However there are too many quick draws. I say they should somehow change the point system. (in fact isn't there a point system like this--)


I don't think a chess game should be a mini-match. And blitz and long control are different. You said it earlier in the thread. People (of all skill levels) play better when there are longer time controls.

At least when a tennis match goes into sudden death it is the same set of skills being tested.

When a tennis game goes into sudden death they don't all of a sudden use smaller balls. And they don't let the person with serve get closer to the net etc. etc.

Or in a tennis doubles match they don't make sudden death one person vs. one person. All the above changes the skill requirements of the tennis match.

Changing tie breaks to blitz or rapid in chess change the skill set. That is why it is inherently unsatisfying. Sudden death in tennis is satisfying. It is the same game just with higher stakes.

As I said above, the changing of the skill set being tested is why I find sudden death goal shootouts in Hockey and Soccer to be unsatisfying. Different game altogether to decide who wins.

Basketball, Baseball,Tennis. They all have overtime (or sudden death) figured out. They don't change the skill set being tested. Just the stakes!

So I am still against blitz or rapid as a decider for one game.

Uri Blass
Posts: 9111
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Rustam Kasimdzhanov proposal

Post by Uri Blass » Wed Jun 05, 2013 1:25 pm

Don wrote:
After my speed chess experience I talked to the local master about this and it was his opinion and his experience that if you are good at one you are good at the other and vice versa and that these imagined disparities do not exist or at least they are really rare.
I agree that disparities are not common but
I believe that there are going to be big disparity for some players.

I expect bigger disparity in the lower level.

I believe that people who do not know much about chess but think better may perform 400 elo better at 90+30 time control relative to blitz(meaning something like rating 1800 at long time control and rating 1400 at blitz).

I also believe that 100 or 200 elo difference may be possible also at the high level.

I have more respect for people who can do well at longer time control.
I dislike the fact that fide decided to change the time control to faster time control(I know that in the past people used slower time control than 90+30 and I remember tournaments with 120/40+60/20 when I was a child) and I feel that the idea to use faster games to prevent draws is another step to the wrong direction.

User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 2:27 pm

Re: Rustam Kasimdzhanov proposal

Post by Don » Wed Jun 05, 2013 1:51 pm

Uri Blass wrote:
Don wrote:
After my speed chess experience I talked to the local master about this and it was his opinion and his experience that if you are good at one you are good at the other and vice versa and that these imagined disparities do not exist or at least they are really rare.
I agree that disparities are not common but
I believe that there are going to be big disparity for some players.
Even if this is true, so what? All we would be doing is bringing a bit more depth to the game instead of being so anal-retentive and conservative.

I see no problem with considering a players skill set to include enough versatility to play at all time controls.

I expect bigger disparity in the lower level.
There will be disparity but as I said, so what?

However I do think that among some players the disparity is psychological, not real. If you are convinced that you suck (relative to others) at 5 minute chess then you probably will. And every time you lose a game at 5 minute chess you will believe that 1 games is convincing proof of it.

I believe that people who do not know much about chess but think better may perform 400 elo better at 90+30 time control relative to blitz(meaning something like rating 1800 at long time control and rating 1400 at blitz).

I also believe that 100 or 200 elo difference may be possible also at the high level.
I don't think that is a likely difference but it's certainly possible. However I say, "so what?"

Every game that has ever had rule changes to improve the game or sport is like this, the change will benefit some players more than others. That is NOT a bad thing. Since when are you or I the one who thinks that what we do now is THE ONLY TRUE WAY? I think it's arrogant. (I'm not saying you are, just the general attitude of resistance in general.)

If this happened it would quickly become accepted after a brief period of fierce resistance and then everyone would start to believe that this new way was the ONLY TRUE WAY.

I have more respect for people who can do well at longer time control.
I dislike the fact that fide decided to change the time control to faster time control(I know that in the past people used slower time control than 90+30 and I remember tournaments with 120/40+60/20 when I was a child) and I feel that the idea to use faster games to prevent draws is another step to the wrong direction.
The time control is a purely a matter of personal preference, it's not a universal principle of right and wrong.

What is pretty clear is that we could use some serious modernizing of the tournament format and that probably involves speeding things up even more.

Here is my basic rant on time control in general, ignoring the draw issue and this is learned from computer chess testing:

To decide a world championship we general play a small handful of games. In fact a ridiculously small number of games. And we play them at a rate of speed where you cannot always even finish a game in one sitting. We PRETEND that we are after the quality of the result to see "who is really better."

But a FAR MORE effective way to do this is to increase the number of games played. We don't test Komodo at 2 hours + 5 minutes per move even though we could generate super high quality games that way. If that was the right way to do this and if that had far greater meaning we would. But it is far better to play thousands of much faster games. OF COURSE we would love to play much longer games but we know for a fact the results are far less valid.

So if you REALLY desire the highest quality (most relevant) results at high levels you should consider increasing the number of games played. It wont' be taxing or logistically difficult as long as you speed up the time control accordingly.

I'm not advocating speed-chess for world championship matches, but I am advocating faster time controls which will provide more relevant statistical evidence on who is best and also provide more GAMES for us to relish and more drama.
Capital punishment would be more effective as a preventive measure if it were administered prior to the crime.

tomgdrums
Posts: 736
Joined: Wed Dec 02, 2009 8:48 am

Re: Rustam Kasimdzhanov proposal

Post by tomgdrums » Wed Jun 05, 2013 4:42 pm

Don,

If longer time controls are less valid why does Larry make such a big deal that Komodo can always catch Houdini on longer time controls?

And Chess Matches (or any competition) are not about finding out who is statistically better. They are about a dramatic competition. The drama of a long tense chess game is much more interesting than a bunch of clock slapping chess for the masses who think Poker is interesting on TV.

And as an aside:
Still haven't figured out the excitement of poker on tv? How is it good television to watch a bunch of people wearing sun glasses pick up cards and then place them back on the table again? I think Poker has done a great job of selling it's sport. And poker IS fun to play. But to watch poker? I don't get it.

User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 2:27 pm

Re: Rustam Kasimdzhanov proposal

Post by Don » Wed Jun 05, 2013 4:56 pm

tomgdrums wrote:Don,

If longer time controls are less valid why does Larry make such a big deal that Komodo can always catch Houdini on longer time controls?
I never said they were less valid. My point is that long time control testing is just not very valid when you cannot run thousands of games. If I had hundreds of CPU's to work with I would prefer testing at something around 10 minute time controls - or the longest time I could muster and still get thousands of games in a reasonable amount of time.


And Chess Matches (or any competition) are not about finding out who is statistically better. They are about a dramatic competition.
Yes, and long time controls games is spectator unfriendly, reducing the drama.

Championships are in fact about determining who is the better player - even if statistically it's pretty flawed.
The drama of a long tense chess game is much more interesting than a bunch of clock slapping chess for the masses who think Poker is interesting on TV.
Maybe for YOU, but this is about what is best for the game, not for you or I. But what is best for the game is best for you and I too.

Do you imagine that I would not benefit from a format change that actually would have a fighting chance of making chess more popular? Do you imagine that this would not be good for chess?

I think shortening the time control is pretty much a necessity to make chess popular again. You might well imagine that a 5 minute chess tournament advertised on television would be a big success, especially if the personalities and players were developed like they are in tennis. Ever watch Wimbledon coverage? It's not just the tennis.

And as an aside:
Still haven't figured out the excitement of poker on tv? How is it good television to watch a bunch of people wearing sun glasses pick up cards and then place them back on the table again? I think Poker has done a great job of selling it's sport. And poker IS fun to play. But to watch poker? I don't get it.
I'm with you on this one. But you hit the nail on the head, this is about salesmanship and I would love to see a little salesmanship in chess of this calibre.
Capital punishment would be more effective as a preventive measure if it were administered prior to the crime.

Ron Langeveld
Posts: 138
Joined: Tue Jan 05, 2010 7:02 pm

Re: Rustam Kasimdzhanov proposal

Post by Ron Langeveld » Wed Jun 05, 2013 8:32 pm

Don wrote:This is an old proposal dating back to 2011, but I think a very good one.

http://en.chessbase.com/home/TabId/211/PostId/4007387

Rustam advocates eliminating draws completely from chess by using a tie-break system similar to what is done in tennis.

I would love to see some computer chess tournaments set up using that system.

Don
I think this is a very bad idea and it is one that can be expected from someone not being a passionate chessplayer. The idea coming from Rustum therefor is bizar.

The beauty of chess is in the moves and positions, not in the results. This trend of finding a 'solution' to draws is making me sad.

Ron Langeveld

carldaman
Posts: 2115
Joined: Sat Jun 02, 2012 12:13 am

Re: Rustam Kasimdzhanov proposal

Post by carldaman » Wed Jun 05, 2013 11:16 pm

tomgdrums wrote:Don,

I think the tennis analogy is not the best one to make as it is a continuation of the same game.

If a chess draw was then decided by a new blitz or rapid game then that is a DIFFERENT game being played to decide the outcome of the first game.

This is why (although exciting) sudden death goal shootouts in Hockey and Soccer are tough to watch. And seemingly less satisfying to me the observer. (unless I have an extreme rooting interest in the team that won)

I am not against draws! They are part of what makes chess, chess!

Someone who is losing who can fight their way to a draw is exciting.

However there are too many quick draws. I say they should somehow change the point system. (in fact isn't there a point system like this--)
One possible innovation I haven't seen yet would be to score short draws, under 30-40 moves, as 1/3 of a point, while continuing to score longer fighting draws as 1/2 a point.
Not completely foolproof, but it would be more awkward to pre-arrange a 40+ move draw than a 15-20 move "grandmaster draw".

CL

Post Reply