Databases - what is allowed?

Here is a game I played a while ago in which I was totally trashed by a database user. It was a very sharp line in Marshall attack and my opponent came with a counter-intuitive move. To me it looked as a novelty, after checking two well-known online databases. I did not have the courage to go all in for a kingside attack and I allowed my opponent to swap some pieces, losing miserably in the end.

However, it turned out that the "novelty" of my opponent was a known move among the users with cyborg databases. Also, the novelty is an engine's top choice in the given position, and that's why probably it appeared in "modern" correspondence games. As far as I know, that move has not yet appeared in OTB games. 

The whole game and the very good advices I received from other players can be found here:

http://www.chess.com/forum/view/game-analysis/hot-novelty-in-marshall-attack

Right, I don't know where you were directing this statement or to whom but I for one can agree with you. For the sake of talking to people who don't understand terminology for what it really is I speak out of context a lot.

(I have actually created custom database engines and can name about 10 of the most popular off the top off my head (innodb, falcon,federated etc.) and be able to tell you about the properties they support as far as being transactional and what not....)

The point of Richie was that using a database with millions of games and searching through it is very far from searching through books/notes/paper databases which were the tradition of correspondence chess in the past. I mean, looking for a particular position in millions of games is way beyond human possibilities. 

costelus wrote: ...there are not too many options to deviate in Marshall, and, if I analized each of the best possible 10 moves for all the moves from 18 to 28...

I think you've lost track of the numbers involved, and the difference between engine search and database...Doing this move by move, using an engine to analyze the 10 best possible moves is easy (assuming we trust our engine's analysis capabilities).

But, if this is a databased search, with all our searched positions already stored and awaiting our search -- then those 10 best possible moves branching out from moves 18-28 will branch 10 to the 10th times... that's 10 billion -- your 20 million game database isn't gonna to be the dragon slayer you think it is.

There are _many_ more possible games of chess than there are atoms in the universe. The comprehensive chess game database isn't a realistic fear until we can among other things, a) use sub-atomic particles to store data, or b)store and retrieve data outside the current universe!

This is beginning to sound like a classic Luddite argument against technological progress.  I can't help but feel that for some, chess and technology are simply incompatible and any new tool that becomes available will immediately be viewed with suspicion and portrayed in the most negative light possible.

I doubt if you'll find many Super GM's who would be willing to give up their mega databases.  Data mining has become an integral part of chess and placing restrictions on what kind of databases people can and can't access is a bad idea for the simple fact that it would be impossible to enforce.  Unenforceable rules are useless rules.

Seriously, do you want to ban everyone who doesn't play chess just like you?

Jgpyth: You are right, but you exagerrate a little bit. OK, if you want to take the best 10 moves from move 18 to 28, that is unfeasible (yet). But, if you want to take the best 10 moves from move 18 to 23, that is only 10.000 possibilities. Not too much. Agree, not easy to do this, but still, if you have access to an engine you can customize with scripts (like Crafty), then one can do it. But also, you have thousands of cyborg games played each day on ICC, playchess, ICCF.

I am convinced that I lost the game posted above against such a database. My opponent's novelty is not present in any classical database.

Of course such a database will not win the whole game for you. But, it is enough to give you an edge. And that's the only thing a reasonable player needs.

The comparations with databases the superGm's have is not OK: the GM's cannot use those databases during a match. 

Eastendboy: such a rule can be enforced! Easily. Each move that is out of game explorer and is "computer correct" counts as cheating. Too many such "coincidences" and you go to play somewhere else. 

costelus, you should be happy that you discovered such a novelty here on chess.com (for free!) instead of OTB when a win would give you a title but you lose to a prepared novelty.

So if I understand correctly, and assuming it was generated by a database of engine versus engine games, use of that novelty is forever tainted by engine use and will henceforth be considered cheating?

well... I don't think that this is prohibited by the rules of the site. It might be prohibited by the rules of the costelus however.

Costelus, there's a huge gaping hole in your argument and the example you cited is the proof.  In the Marshall game you played, you were surprised by a "novelty" supposedly taken from a "cyborg database" but you seem to be forgetting that the a4 novelty was a terrible move!  As pointed out in the replies to your message, black had a move that completely refuted the "novelty" but you just missed it during the game (understandably, as it wasn't an easy move to find).

  What you seem to complaining about is the fact that someone else had a better database than you.  The "cyborg novelty" was a novelty best forgotten -- regardless of whether or not it's an engine move -- if it WAS an engine move, it was a very poor one since there's a complete reversal of fortune a few moves later!

Complaining about "cyborg databases" and then citing a very weak move as the reason why those database should be banned is illogical.

While I don't have an opinion on this specific game, I think that the notion that other people may have a different database than you seems to be part of the lay of the land.

I don't have a problem whatsoever with how people create any opening DB they like. Granted, I think only professional players really have the time, energy, or whatnot for the undertaking. But if a player wants to put the King's Gambit into a round robin tournament and analyze it however much they want, and go beat Gonnosuke because he thinks it's ok to play it with white, then more power to that guy, right?

1.e4 c6

I'm afraid you've totally lost me. Are we still talking about chess?

I think there is no hole in my argument: that novelty was a bad move only for a player with a database of cyborg games and with an engine. As you can see from that topic, the refutation of that move can only be found in cyborg games. For me it was definitely a strong move I could not cope with it.

For a computer that move is optimal. Probably it would take hours of analysis for it to go to a very high depth and see that the move is indeed bad. The engine also does not think that my answer in the game was so bad. But it never managed to find the refutation of that "terrible" move (or I didn't have enough time to run it).

Well for what it's worth, I think that he is free to have a custom transformer, or he can run it in Europe, and perhaps have both more power and more current.

Common, let's not exagerrate. A single move will never be considered cheating. Moreover, there are lines in today's fashionable openings which were discovered by engines (and later on tested OTB).

The whole point of this thread is that it is extremely hard to play - in some situations- against a huge database, when your opponent might have all the answers prepared - for the key moves.

Your argument for banning DB's is, to paraphrase "I should be able to play bad moves because if no one had a better DB than me they couldn't have found  any unknown to me refutation of my move"? Coupled with the rule appendices you suggested you could only ever have mediocre players here. What about people Like Kasparov? Go through some of his games with an engine...the man is deadly accurate, like less than 3% errors in most long games that I have looked over. By your rules he could not ever play here at chess.com....because he might have spent more time studying analysis' than you? Does this really make sense to you?

I will quote you from post #36:

"Eastendboy: such a rule can be enforced! Easily. Each move that is out of game explorer and is "computer correct" counts as cheating. Too many such "coincidences" and you go to play somewhere else." 

I don't say that databases should not be allowed. I just say that it should be a limit after which consulting a database with cyborg games is cheating. I also think that a database like Game Explorer is good enough.

We have here many people who play much better than Kasparov, even if they do not even have an ELO. And yes, I *love* to play against mediocre players, I *love* to make mistakes, then my opponents to make another mistake, and the winner to be the one who makes the last mistake. I *hate* to play against opponents with perfect play, whether they are cheaters or heavy database users.

About Kasparov: depends on what do you call "error". So, what is an error? 

The Grobe: I don't see the contradiction. Yes, I do think that let's say 5-6 computer correct moves after Game Explorer is extremely close to cheating. In any case, it is likely to give you an edge over your opponent.

I really don't know how someone could make the difference between consulting a giant database and using an engine to make 5 moves for me after I got to a position which is not encountered in Game Explorer.

An error in this case is anything that is not considered the strongest move by the engine. and what if two moves score equally well? can that player play niether of those moves now because they are to good?

You may enjoy playing against mediocre players and not being able to get better but to a lot (most?) players the idea is to reach their maximum potential. Spending time studying massive DB's, reading chess articles, playing countless games with the same openings to test there knowledge, playing blind matches, all these things give you an edge over someone who would rather everyone else be mediocre than to take the loss as a lesson and try to improve.