Rich As A King

Monday, November 20, 2006

Important Official Rules of the Kramnik versus Fritz match


The Kramnik versus Deep Fritz is taking place soon. The rules have been published by the organizer. With a lot of misunderstanding regarding the rules in Elista, it is important that we know the full facts before making judgment. Here are the most important parts of the official rules for the match:

DATES
Six games will be played on the days 1, 3, 5, 7, 9, 11 where day 1 is 25 day of November 2006. The games commence at 3.00 pm and end at 9.00 pm at which time Mr Kramnik will have the right either to continue with the game or to adjourn it to the next day.

TIME CONTROL
In each game the Players shall each have to make 40 moves in two hours followed by 16 moves per hour thereafter provided that in the event that a game has not been completed within six hours it may be adjourned to the following day at Mr Kramniks discretion when play will continue at the rate of 16 moves per hour for a further six hours.

Mr. Kramnik shall have the right to adjourn any game after 56 moves even if six hours of play have not been completed. Should this right be exercised, play shall continue on the following rest day at the rate of 16 moves per hour.

ARBITER AND MATCH DIRECTOR
The match will be ruled and officiated by one arbiter. His decision regarding any interpretation of any of the rules of chess pertaining to this match shall be final and binding.
The arbiter shall be selected by UEP and UEP shall inform Mr. Kramnik and Chessbase (CB) about the name of the arbiter as soon as possible but in no event later than 1 October 2006. UEP informs about its intention to contract Mr. Albert Vasse (Netherlands) being the arbiter of the WCC.

UEP shall have the right to announce a WCC Match Director who will be responsible for all necessary communication and administration during the match (with the Players, the Arbiter, the sponsors, the press, the local organisation at venue and other parties involved). For the avoidance of doubt: The Match Director has no influence or responsibility on the Arbiter’s decisions and shall have no influence regarding any interpretation of any rules of chess and any rules of the WCC which is solely a responsibility of the Arbiter.

OPENING BOOKS
The computer will consult an opening book during the game. During the match, the opening book may not be modified, except that up to 10 ply of additional moves may be added in the opening variation of the game which has most recently been played (not counting adjournment sessions) and the weightings of specific moves may be modified so that the different variations, already present in the opening book, will be preferred by the program.

All opening book modifications will be entered by the Arbiter before the game according to the material confidentially provided to them by the Deep Fritz Team. A member of the Deep Fritz Team will be present and if necessary guide the Arbiter through the necessary steps of operation.

At the conclusion of each game the Arbiter will attempt to replicate the opening of the game on a computer which has the opening book and program as delivered to the Kramnik Team and the Arbiter. If they find any discrepancies, the Deep Fritz Team is required to explain these to the satisfaction of the arbiter.

If a violation of this rule is determined by the Arbiter, the penalty may include loss of the game.
As long as Deep Fritz is “in book”, that is playing moves from memory and not calculating variations, Mr. Kramnik sees the display of the Deep Fritz opening book. For the current board position he sees all moves, including all statistics (number of games, ELO performance, score) from grandmaster games and the move weighting of Deep Fritz. To this purpose, Mr. Kramnik uses his own computer screen showing the screen of the Deep Fritz machine with book display activated.

As soon as Deep Fritz starts calculating variations during the game the operator informs the arbiter. The arbiter confirms this on the screen of the playing machine and then shuts down the second screen.

ENDGAME TABLEBASES
The use of a database of endgame positions (“Tablebase”) is permitted only if the tablebase contains positions with a total five total pieces or less, including kings.

When Deep Fritz identifies the board position in a tablebase, it must inform the Arbiter, who will then stop the clocks.

In the presence of the Arbiter, the Operator will inform Mr. Kramnik that the position has been located in the tablebase.


If the position is evaluated by the tablebase as winning for the side played by Deep Fritz, the Operator will inform Mr. Kramnik of that fact in the presence of the Arbiter. The game will continue, unless Mr. Kramnik chooses to resign.

If the position is evaluated by the tablebase as winning for the side played by Mr. Kramnik, the Operator will inform Mr. Kramnik of that fact in the presence of the Arbiter. The game will continue unless the Deep Fritz Operator chooses to resign.

If the position is evaluated by the tablebase as a draw, the Operator will inform Mr. Kramnik of that fact in the presence of the Arbiter. This will constitute an offer of a draw. The game will continue, unless the offer is accepted prior to the completion of Mr. Kramnik’s next move.

It is recognized that the program will access tablebases in its calculations. The above rules apply only when the position on the board is present in the tablebase.

DRAW OFFERS
Mr. Kramnik may offer a draw at any time, regardless of whose turn it is. The Operator is authorized to accept or decline the draw on behalf of Deep Fritz.

The Operator may offer a draw on behalf of Deep Fritz, however a draw may not be offered unless a previous offer by Mr. Kramnik has been declined.

If Mr. Kramnik feels that the position is clearly drawn, he may notify the Arbiter and the Operator that he is making a claim of “technical draw”. The Arbiter will stop the clock. Mr. Kramnik will then explain his reasoning, and the Operator is obliged to accept the draw unless Deep Fritz can demonstrate that in the previous ten moves progress has been made.

The Arbiter will determine the validity of the claim, and his decision shall be final and binding. Should he uphold the claim, the game will be declared drawn. If he rejects the claim, then the game continues.

During the deliberations regarding a technical draw, the clocks will remain stopped. In the event the Arbiter reject the claim, the penalty will be deduction of the lesser of 5 minutes, or 10% percent of Mr. Kramnik’s remaining time. In the event a second claim of a technical draw is rejected in the same game, a 25% penalty will be assessed.

POST-GAME CONFIRMATIONS
After the conclusion of each game, the Deep Fritz Team shall provide a printout of the computer analysis of the game to the Arbiter and the Kramnik Team. This printout includes evaluation, search depth, expected move and thinking time.

After each game, the Arbiter will be provided with an opportunity to compare the opening book used in the game with the opening book used in previous games.

DISCLOSURE OF MATCH ENGINE
By October 1, 2006, Mr. Kramnik and the arbiter will receive the final match version of Deep Fritz.
After this date only bug fixes are done on the engine, for example obvious crashes or obvious positional errors. No positional knowledge will be added. Should the engine be modified in any way after 1 October 2006 Chessbase will notify Mr. Kramnik’s team and the Arbiter in writing about this specific change and demonstrate its effect on a test position. In any case Chessbase guarantees that any change after October 1, 2006, will not influence general playing style and tactical strength and confirms that the engine code remains practically unchanged after October 1, 2006.

From 1 October 2006 on, the Deep Fritz team will be ready on the request of the Kramnik team to install the final match version on Mr Kramnik’s trainings machine. The Deep Fritz team shall inform Mr. Kramnik and the organizer (UEP) by no later than 1 October 2006 about the specifications of the hardware which will be used during the WCC. In co-operation with a possible hardware producer and Chessbase the organizer (UEP) will make their best endeavours to provide Mr. Kramnik a trainings machine being similar with the machine during the WCC.

At any stage in the match, the Kramnik team may copy the exact playing engine directly from the tournament machine under supervision of the arbiter. The Deep Fritz Team is not required to disclose the exact hash table size for the match. It is understood that hash table size does not influence playing style but rather introduces a small element of non-determinism into the move selection process. The Deep Fritz Team has to notify the Arbiter of the Hash table size so that they can reproduce the programs calculations.

Source: Official match site. Here are the full compete rules.
Posted by Picasa

58 comments:

Frank said...

Susan, what are your feelings about this? I believe all Kramnik has to do is find a few drawing lines and play them to a tablebase draw and memorize the moves to be played over the board.

The ability to adjourn also means the ability to run the engine all night long in analysis mode.

To me this is just a chessbase PR stunt, not man vs machine.

David said...

To: Frank

Memorization of lines can't help, if you read the article carefully you might have found out that the chessbase team can alter the memory given to the engine and so make the probability of playing the exac moves once again a small one - changing the amount of space previded to the engine before each game makes the engine play "random" or in other words it makes it imposible to predict with high degree of accuracy.

+ There are also heat factors involved, it shuld be impossible to duplicate the playing conditions to prefection because of them - for example if the room the computer will be in had a different air temperature then the room which Kramnik used for preparation the heat diagram of the proseeor(s) will differ making the end speed different making the evaluations different ect ...

D.K.

pawnstar3 said...

i think this is going to be a boring match- lots of draws- but i'm rooting for fritz

Anonymous said...

'Sanctuary Chess' ...

The Kramnik-DFritz match is in a variant of chess which I have previously dubbed 'Sanctuary Chess'.

This is because it does not require the human player to defend a possibly hard-to-defend endgame. A pity: computers can/have been programmed to play the opponent as well as the game, identifying and using perceived preferences and fallibilities. You may well feel 'robbed' if this situation arises, as it did (6-man position) in the Kramnik-Topalov game

Kramnik, it would appear, also gets to use the opening book information of the computer.

The computer is not allowed to be pre-loaded with 6-man endgame tables. But what if it computes 6-man EGTs during the game, and uses them. After all, EGT-search is just a specialised form of forward-search.

A three-way compromising of the computer's strengths. Poor set of rules.

Guy Haworth

Anonymous said...

So many special rules in favour of Kramnik. The only thing I am missing is that Viktor is allowed to pull the plug if he feels he is loosing ....

Renzo - IncaKing @ ICC said...

So Kramnik will be able to:

- See Fritz' opening book on the current position (possible continuation plus statistics) while Fritz in "in book".
- Adjourn games and run the engine overnight.
- Know of a tablebase result in advance even if for him the position is unclear.

Besides already having the program version with him.

All this seems to be a group of clauses made to avoid a 6-0 result by Fritz.

Chris said...

Is kramnik allowed to use a computer and team after the game has been adjourned until the next day?

Anonymous said...

Yes, Kramnik is even allowed to use Rybka for overnight analysis.

Dot said...

No one has ever gotten these kind of conditions. It's obvious that the match has been rigged.

Anonymous said...

Mr. Kramnik is a dignified gentleman. He prefers fair play. He would never agree to such biased rules, that would be heavily in his favour and destroy the fairness of the match! Stop with these lies immediately!

Show some respect to the unique player, the strongest player in the world, the magnificent Vladimir Kramnik.

Anonymous said...

Two other points re endgames in the 'EGT zone':

1) the rules do not prevent Kramnik from studying the endgame overnight with 6- or 7-man EGTs.

2) the EGTs are presumably DTM EGTs, and so will necessarily flag up an endgame for which a 50-move rule claim would be available with best play.

Guy

Alex said...

Kramnik should be ashamed of himself. Now we know how Carsten Hensel and Kramnik operate. This is not chess, it's about a fixed match to promote Fritz.

David said...

I think the point is to compare Frizt calculating abilities to that of Kramnik. I don't think there is anything wrong in letting Kramnik see the opening book - there is no point in comparing human memory to that of a computer - that wuld be stupid; the same goes for the tablebases - if the computers plays with help from the tablebases then it doesn't matter if you have Rybka, Hiarcs or Fritz running + it is the same as in the opening database, we all know that computers have "better memory" so there is no point in testing this!?

D.K.

gabor said...

This "show" was over when Deep Blue defeated the world champion at the time Kasparov. That's why IBM took apart Deep Blue and gave no rematch. The entire point of the "man vs. machine" was that men claimed that the machine can never defeat the best chess player. It did. The current Kramnik-Fritz, with weird rules or not, became irrelevant (however interesting). We now know that computers are able to defeat the best chess players. People should forget about that, and re-declare chess what it always was, one person's mind against another person's mind. The computers will be there for learning, practice, analysis, player if no human is available, etc.

Besides, the whole issue is downright silly. IBM is working on an over 100 Ghz CPU. If and when they make it work, eight of those put together will defeat any chess player, no matter what. So, it is only a matter of time. Kramnik won't be able to bring back the "can a computer defeat the best chess player" debate, even if he wins this match.

Gabor

Ken said...

All;

I am very surprised by some of these comments. I think are some well-thought if burdensome rules of play and will likely be the standard in future man-machine matches.

One, the human knows when the machine is reading a book instead of thinking. Since the start of compouter chess there has been an unfairness about the computer's use of opening (and now endgame) books - but if you shut off the book the human has an advantage. I would have preferred allowing Kramnik the use of opening database and EGT, but I am OK with this rule.

Two, it is inherently unfair to allow a computer team to drive a human to exhaustion. The adjournment rule makes a lot of sense, and was considered fair for a century or so by top players.

Three, many of these matches are "haunted" by the possibility that humans are directing computer moves and making strategy choices on openings during the match. The rules meant to prevent such "tampering" are good.

Just my opinions...

Chris said...

I do not think that we should call this cheating my kramnik. Everyone is aware of the rules; the results will be interpreted with respect to these rules.

I would like to see a man vs. machine match of FischerRandom. Has this been done by any top player?

pawntorook4 said...

just think if kasparov had these conditions for deepblue II!?!

elizabeth said...

I note that if the Fritz team objected to the tablebase rules, then they could turn the tablebases off and force the computer to work out the best moves OTB.

I think that using the tablebases at all is an advantage to Fritz, as it is more likely to err in a technical endgame position than Kramnik.

Kramnik would have done better, I think, to drop the request that he have access to the tablebase, as this request appears unfair in his favor. If both players had to compute the endgame positions without consulting a database, and I think there would not be such awkward rules and bad PR. But I am not sure that the Fritz team would agree to this.

Computer chess is beginning to resemble correspondance chess in its rules (though with a clock), given the amount of external data that the program can use simply because the data is stored on the same computer.

Austen said...

I'm looking forward to this match and find the rules to be very interesting. Humans have long been at a disadvantage to computers in chess because of opening books and endgame tablebases - the point is, these matches haven't been fair to the human players in a long time.

The idea that Kramnik will have Fritz's openings book available is entertaining, but ultimately I think it would have made the match more interesting to allow Kramnik to bring his own book to the table and allow Fritz's book to be available only to the computer. Kramnik would not be at a disadvantage to Fritz's book knowledge, but the specific weightings would not be available to Kramnik.

I am surprised to see the adjournment clause, but I am also thrilled by the idea of it. Computers play tirelessly whereas humans do wear out. This helps level the playing field with respect to fatigue. Granted, Kramnik's ability to go and analyse the position with another computer taints the human vs. computer play with the work of another computer, but ultimately I see this as much fairer to the human player.

Overall I don't think that this match will be boring at all. These rules encourage Kramnik to play for a win with a lot less risk than other man vs. machine matches. I hope that this match is less one-sided than the Adams-Hydra match.

Marc Shepherd said...

We already know that in a straight-up match, computers have the edge. These rules seem to be designed to blunt the computer's advantage in the two areas where it is obviously superior: memory and stamina.

Of course, if Kramnik wins big, you'll have to put a big asterisk next to it in the history books, because he had advantage that human opponents haven't been given in the past. If the match is drawn, or if Fritz wins, then I think we can say that human vs. computer chess is over. If Kramnik can't win with all of these extra advantages, then what chance does anyone else have?

Some of the advantages really aren't that significant. Kramnik's opening theory and endgame technique are already pretty solid. The advantage of adjournment is pretty significant, though, because he can go home and analyze all night.

Since the start of computer chess there has been an unfairness about the computer's use of opening (and now endgame) books - but if you shut off the book the human has an advantage.

I have never seen any "unfairness" in a computer's opening book. Any GM has thousands of opening moves memorized. In some lines, GMs play from memory as deep as 15 moves or more. Fritz's "book" is simply an electronic equivalent of the memorized openings that all GMs have.

Endgame tablebases are a little different, without them computers would play quite poorly in the endgame.

Austen said...

I'm also pleasantly surprised that they are going through so much trouble to prove that the moves are generated entirely by the computer (especially showing the logs! IBM could learn a thing or two from this).

Anonymous said...

Most endgames go beyond 56 moves. I would find it totally unacceptable that the human side would have the possibility of analysing these positions with computer aid (Hiarcs, Rybka, Shredder and an army of GMs) it would totally ridicule the fairness of this encounter.

Therefore the Great and fair Mr. Kramnik has not agreed to such rules because it is his nature, and the claims posted here by Mrs. Polgar are untruthful.

Long live Kramnik!

Anonymous said...

Person v. Computer matches are exhibitions, not "PR stunts". The odd rules for this match mean it is not "real" chess, but so what?

These rules reflect the reality that computers have 100% solved the late endgame phase of chess.
So these rules wisely avoid letting the late endgame phase play an unnecessarily large factor in the overall outcome of the match.

In fact, this whole exhibition is mostly just about the middle game phase.

This is a battle of humanity v. itself. We are using our ingenuity from two differing competing angles. Humanity wins either way.

Dan H. made the fair point that this 2006 match is interesting even tho Hydra recently erased GM M.Adams. The Hydra match was not restricted to the common PC computer & software any of us could buy, but this match is.

Gene_M

Anonymous said...

This was posted on Sunday and I thought this guys a jerk ...I now think he is right

hoddy said...
This time he will play what many say was his helping source..I somehow don't thing there will be as many rest room breaks..and no blitz games to decide the winner...the result may very well be pre-arranged and again the real loser being Our Great Game Of chess Sunday, November 19, 2006 6:00:20 AM

Anonymous said...

'Chris' asked (Nov 20 1:13PM):
{
I would like to see a man vs. machine match of FischerRandom. Has this been done by any top player?
}

A.Yusupov played "shuffle chess" against a computer in Frankfurt in 2000. The setup restriction rules were improperly understood, so it was not quite FRChess (bad luck placed the king outside the two rooks). But maybe it was close enuf.
Yusupov lost both games.

P.Leko felt the lack of openings knowledge (in both human and computer) gave the human an advantage.
Yusupov believed the opposite.

The chess world has not yet generated enuf evidence to answer this interesting debate.

Yusupov applied his chess1 opening knowledge is a clever creative way: he adopted a hedge-hog pawn position.

That Yusupov had any opening strategy at all hints at how much rich new chess openings theory humanity will discover when it gives chess960 a serious foot-hold along side chess1.
And these theories will be more elegant and principle based than are the raw concrete move memorization theory of today's chess1.

- - - - -

The Kramnik v. DeepFritz 2006 rule allowing the human to read the openings book stored inside the computer all but proves these human v. computer matches should switch to chess960.

Gene Milener
http://CastleLong.com/

Anonymous said...

Stop with these insults immediately! Vladimir Kramnik is the strongest positional player in the history of the game and a gentleman who stands for fairness and dignity.

He would never agree to such rules!

1. Computers can memorize opening books therefore they are allowed to have them. Humans too would have opening books of such depth if we could memorize the lines.

2. -||- re: tablebases.

3. Human player should rely solely on his own ability. No postponements. Possibly an two hour break during which time Vladimir Kramnik must stay supervized and can be in no contact with outside help.

These would be fair rules. So these are the rules Kramnik will play with.

Now stop with the lies immediately!

Kramnik the gentleman!

Gunther said...

Now we understand why ChessBase was so biases toward Kramnik. He's on their team and they're protecting him. This is bad for chess because ChessBase is helping to promote pre-arranged chess.

Anonymous said...

amen to Gunther. Kramnik is their spokesman or at least they are his sponsor so they are bending every rule for him in every way.

Vinay said...

The rules seem fair to me. What is wrong with them ? Why bother comparing human memory to a computers? Thats silly.

I think it will be an exciting match. Even if Kramnik loses 0-6.

- Vinay

Anonymous said...

These rules are a result of adams being blown off the board from hydra. Kramnik discussed the results as terrifying.

Kramnik just wants to avoid, at all costs, 6-0 from Fritz. And his paying sponsors are obliging him.

Maybe next match he will get pawn and move odds from Chessbase and ability to take back moves.

Ivanb said...

Why nothing is said about the power of the hardware? Is it a weak pocket PC with DeepFritz installed or multiprocessor cluster with many TerraFlops?

In later case the dispute of human vs computer is meaningless, zillions of variations will be calculated in seconds, no chance for any human. And software type becomes less important, would it be Fritz, Crafty, or Rybka. To be honest, in this sence I don't get this match.

Anonymous said...

The Truth:

Kramnik wanted fair rules.

Chessbase insisted on giving these odds (after all they bring the sponsors to the table).

Kramnik, as a gentleman and noble person, said: "OK then. You say the rules and I will be there to entertain the millions of chess fans around the world."

Even if these are the factual rules, Kramnik has acted with impeccable class in this.

Long Live Vladimir Kramnik!

Anonymous said...

Entertain? He's paid $500,000 for 6 games. Entertain my ....

Anonymous said...

How many unsupervised potty breaks is kramnik allowed to consult with rybka?

I'm dying to see if he needs a potty trip per move again.

Anonymous said...

Kramnik is the true World Champion, and he has behaved like a gentleman. Particularly in the Elista match againt Topalov.

It is strange to see all the resentment people express against him.
Eighter they have been fooled by the outrages behaviour of the Topalov team, or they have no understanding whatsoever about the chess he plays. Or perhaps both.

Anyway, I think his positional style should fit pretty well againt a machine. At least better than the style of Kasparov. Though a Karpov at his peak would be the best represntativ for the humans aganst the machines.

Anonymous said...

What about bathroom rules? We need more funny stuff about bathrooms.

hoddy said...

Ok then...the water kramy drinks is full of electrons were via satellite the best move is transmitted to...then he goes to toilet has a quick pinckle and the electrons sperate in the water and form ...say Bxe6 is good...then he plays that move...that's why kramy drinks so much of that electronic water...no joke true fact

hoddy said...

PS: a pincklephone would defiantly trip him up

Anonymous said...

regarding time control:
the rules should point out clearly, in the case the game is adjourned after at least 56 moves and say 5 hours,whether the play shall continue the next day for a further 6 or 7 hours.
hac

Anonymous said...

"If the position is evaluated by the tablebase as a draw, the Operator will inform Mr. Kramnik of that fact in the presence of the Arbiter. This will constitute an offer of a draw. The game will continue, unless the offer is accepted prior to the completion of Mr. Kramnik’s next move."

This is ridiculous, If Fritz considers it a draw does not mean a human can play the end game out to its drawn conclusion. This will mean, Kramnik can expect to escape with a draw in a tough position, he may be unaware of how to defend. The rules are a total baloney.

Rgds M.

Anonymous said...

Of course Kramnik also has unlimited potty breaks. Dufus.

Lee said...

I think the match should be renamed "Kramnik and Fritz vs Fritz".

It certainly is not "Kramnik vs Fritz".

From an earlier anonymous blogger:
"Quite honestly the same program on the same hardware will always pick the same move. All he has to do is play 6 games at home, memorize them and play it out at the board."

I don't dislike Kramnik, but these match rules are just not worthy of a match billed as World Champion vs Fritz.

Anonymous said...

I'm not totally clear on the rules. On October 1st, Kramnik was given the Fritz engine. Then at any time during the match, he can get a machine as much as possible like the one he's playing. So does that include the opening book or not? If Kramnik has had the exact opening book since October 1st, his team could find weak points in it. I think that the opportunity to "win the game before it even starts" would be unfairly in favour of Kramnik. However, unlike many of the other posters here, I'm OK with the other rules.

Jonathan Berry

Anonymous said...

When Kramnik played the computer in Bahrain in 2002, did he get this advantage of seeing Fritz's opening book during the game (rather than using his brain a la normal chess)?
And did Gazza get this advantage either? And are there any sites that prove your answers? Thanks :)

David said...

To: Lee
"I think the match should be renamed "Kramnik and Fritz vs Fritz".

It certainly is not "Kramnik vs Fritz".

From an earlier anonymous blogger:
"Quite honestly the same program on the same hardware will always pick the same move. All he has to do is play 6 games at home, memorize them and play it out at the board."

I don't dislike Kramnik, but these match rules are just not worthy of a match billed as World Champion vs Fritz."

I you read my first post (te second post onthis page) you'll find out that the same cmputer with the same software will NOT pick the came move with high statistical probability.

D.K.

Anonymous said...

If not for the P.R. aspects, they would label this event "Fritz and Kramnik practice sessions".

Lee said...

To: David

"the same computer with the same software will NOT pick the came move with high statistical probability."

Yes, true, but Kramnik and his team get to analyse Fritz for two months before as well as during the match. If they carefully selected conservative lines in certain openings, could they eliminate most decisive winning moves by Fritz in those lines in two months of analysis?

The rules are so disadvantageous to Fritz that it will be no great credit to Kramnik if he wins but certainly an embarrassment if he loses (though this is hardly likely with such rules). In fact the neccesity for such conditions suggests that the result would be a foregone conclusion if Kramnik were to face Fritz as if Fritz were a human opponent, i.e. have access to Fritz’s previous games, but not to its opening book, no obligatory draw offers, no access to Fritz’s engine before or during the match, etc. It was a great credit to Gary Kasparov that he took on Deep Blue on more or less equal terms and conceded defeat. That is true sportsmanship.

I thought that the whole point was to see if a human could beat Fritz, not whether the human (and his aides) can do a better job with Fritz than the programmers of Fritz!

Poor Fritz, no human would accept such terms and conditions!

David said...

To: Lee

I think youre missing the point here, as I stated in my previous posts it is silly co compare caomputer memory to a human one! (A computer has no problems storing the entire Britanica but a human has no way to do that! So such comparison is irrelevat!) The purpose of the match is to compare fritz calculating abilities to that of Kramnik not to compare their memories! As far as preparation goes it is not that difficult to make fritz behave unpredictably during the match - which if you read the rules carefully Chessbase is going to do! So if Kramnik finds flaws which isn't easy agains a Fritz-type engine the chess - base team can correct them or avoid them the next game. (why is it hard to find flaws against a fritztype engine - if you look at the number of positions fritz searches and compre it to that of Rybka you'll find that fritz searches about 10 to 12 times faster, but reaches the same depth as rybka or less -> why is that? It is because Fritz does't search only for "smart" moves but checks bad lines also, which Rybka doesn't - this wasy the probabilyity of finding a tactical trick increases but at the same time it decreases the depth of the search - which means Fritz is more likely to make long term strategical error then Rybka. Why is it then hard to find mistakes in Fritzs play - the thing is that you can forget about any tactical chances and at the same time you must always be on the look out for some + Fritz is predetermined to like open lines and more tactical openings ... which makes the number of variations impossible to analyse and it makes it easy to completly change the way fritz is going to play just by adding or substracting some memory ...

I ran a test match Rybka 2.1 vs Fritz 10 a few days ago (rybka won 4,5 to 1,5) but the interesting par is that after the opening they dissagreed on more then 50% of the moves, Fritz 9 also dissagrees on 40%+ of the moves cmpared to fritz 10 (again if you don't count in the opening moves) Which means that is is easy to cerate a huge diversity even on the same computer just by tweaking minor things.

D.K.

David said...

P.S. I didn't like the match against Garry Kasparov he was the better player in my eyes compared to deep blue but the terms of the match gave an advantage to Deep Bule - team that is why I think that match shuld not be considered as a final verdict (this matc also shuldn't be considered a final verdict because Kramnik isn't going to play the best maschine or the best software available - so the point of the match is simply to compare the calculating abilities of Fritz to kramnik as I said many times).

P.P.S. If they called the match Fritz 10 vs Kramnik and not Man vs Maschine it wuld geat less publicity )

D.K.

robert said...

THIS WILL BE A FREE STYLE MATCH.

Let Kramnik use Fritz 6.0 with the openning book and tablebases 6 pieces available (so he is a centaur). Let us see how powerful Fritz 10 now.

invincible said...

Fritz is like a blunt tableknife.

Just running the dagger Rybka or sword Zappa or any recent world champion would tip the match into kramnik favour.

Yet he has to make 40 moves in 2 hours without blundering.

interesting match!

Ashik said...

I was thinking about the DRAW OFFER RULE. Is it not like, a computer will consider the start of end game position only when there is 5 or less pawns and pieces (including King) present in the board? If this is, then a GM should be able to hold a draw if there is a draw in the position. Irrespective of this assumtion true/false, I would like if the computer operator would offer a draw to Kramnik not the first time in endgame when Fritz evaluates the position to be exactly equal, rather if Fritz evaluates the position equal "3 consequitive times in endgame". This way it will make sure Kramnik is going through the right path to the draw. What do you think about this proposal?

Renzo - IncaKing @ ICC said...

I believe that this type of matches are starting to miss the point very badly.

Why opening books? Why table bases? Let both the human and the machine play by their own means!

In other words, chess with the classic initial piece arrangement (RNBQKBNR) brings the problem of "theory" - which is no more and no less than an attempt to start to solve the game since the beginning. The human has memorized a lot of openings and the machine has access to its opening book, as large as it can be from its creators. As there is no point to compare the human's to the computer's memory capacity then the human is in disadvantage. This lead to allowing the human to see the machine's opening book and even more, to have a copy of the program before the match so it can analyze its opening repertoire - does that make sense? Not at all! The solution? Replace the classical chess match by a Fischer-Random match and have both computer and human "use their own brains" (and circuits) right since the beginning.

Table bases? Another farse. When two humans play, how many times has one of them lost a winning position? Even more, how many times they have not been able to find the right moves to secure the draw? In this case the human is receiving the benefit of a draw offer when the computer finds it in its table bases - why??? If the human wants to draw then let him force the draw and not give it away to him! If he blunders in the intent then better luck next time! Now you can say that would be unfair to the human and I would agree. So the solution is not to allow either human or computer to use tablebases at all. It is OK if at a given time during a game the position is identified as forced win or draw due to table bases but just don't let the playing computer identify that. In other words, you can always have other computers identify the table base result, but not the one playing. It simply should be programmed not to use table bases.

Observing these two simple rules will bring back the point of having man versus machine chess matches. No opening books and no table bases.

Unfortunately, there seems not to be a solution for the issue of the human getting tired so adjournments seem fair. However, it would be ideal to ensure that the human will not consult computers overnight but... that seems just utopic.

Enough of opening theory - it is just killing the interest of chess by a wider number of people worldwide. Many would not simply not compete in tournaments because the initial advantage will always be with the one having memorized more opening lines, thus making chess more elitist and discouraging pure talent. Try to change the rules by adopting Fischer-Random and you will see how a greater number of people will practice the sport, and also more people watching the games as you only need to know to move the pieces and not disgusting theory. This would even be more beneficial from a commercial and sponsorship point of view. Fischer-Random Chess is the future of our sport.

spirit said...

The match clause is interesting but has holes.

1. "All opening book modifications will be entered by the Arbiter before the game according to the material confidentially provided to them by the Deep Fritz Team."

Whether or not Kramnik will be informed of the modification is not stated clearly here...but implied
The extent to which modifications will affect transpositions to another line from the line in question is not clear.

2."At the conclusion of each game the Arbiter will attempt to replicate the opening of the game on a computer which has the opening book and program as delivered to the Kramnik Team and the Arbiter. If they find any discrepancies, the Deep Fritz Team is required to explain these to the satisfaction of the arbiter."

The definition of discrepancy here can only mean that the exact opening move sequence cannot be reproduced,does this need any explaining.

3. "In the presence of the Arbiter, the Operator will inform Mr. Kramnik that the position has been located in the tablebase."

Is he required to prove this?

4."If Mr. Kramnik feels that the position is clearly drawn, he may notify the Arbiter and the Operator that he is making a claim of “technical draw”. The Arbiter will stop the clock. Mr. Kramnik will then explain his reasoning, and the Operator is obliged to accept the draw unless Deep Fritz can demonstrate that in the previous ten moves progress has been made."

The exact definition of "Progress" here is rather subjective and therefore suspect.

The match is clearly a marketing stunt by chessbase in light of the rise of the rybkas.
The conditions of the match is dubious somewhat and misleads us to believe that GM Kramnik has been given overwhelming advantage when in reality there is nothing.
One example stands out:Most people think Kramnik can just set up a closed game, shuffle pieces until the 56th and then adjourned.Go home play out the position to his advantage and then reproduce the exact move sequence OTB the following day.This is exactly what chessbase wants us to believe, but this can never happen because kramnik wont know the exact hash size of HERR FRITZ's brain!

Anonymous said...

In my opinion the computer gets some unfair advantages as well. The programers can after all change Kramnik's opponent during the match by changing, the opening book, the prefences and weights for each variation, and the hashtable. The programers rightly want to avoid losing the smae way game after game. So both sides struck a reasonably fair arrangement. The programers can make limitted changes to the Kramniks opponent but kramnik gets some more limitted information in exchange.

Anonymous said...

Anand, Karpov, Leko and other GMs express their opinions about the match: http://www.chesspro.ru/index_en.html

Xmas said...

I get the impression most of you people have no idea of computer chess and barely speak English.

Anonymous said...

Thanks Susan, for publishing the
rules in a readable format. If you
anything more about the hardware
involved, please tell us.

My impression (as opposed to xmas)
of this is that Kramnik has no idea
of computer chess. He probably
thinks that he can explore some old
opening bug of Fritz'es found by his
hired seconds.

He obviosly also wants to cripple
the tiny Merom laptop computer in
the endgame.

(does anyone know if the Merom is
allowed to run in its 64-bit mode
- wich is just great for the game of chess?)

I'll like to see a match between a real Man versus a real Machine.

After looking at his second game,
this coward could as well be playing
against an ordinary 50-watt lightbulb
instead of this powersaving Intel processor.
With the right machine operator he
would still loose!

My prediction is that he will loose
against any mobile telephone of his
own chooice in the 3 years.

I bet a millon dollars that my old
telephone will beat him right now, if he allows me to connect it to
my home computers!