Opening vs End game databases

So I was wondering what's the reasoning behind allowing opening databases on CC while prohibiting end game tablebases. Aren't both of those pre-calculated moves?

Although I do acknowledge that end game tablebases are solved analysis while opening databases are less precise, at the very least what we are doing is letting grandmasters play for us up to move 15, 20 (since we don't have to memorize the lines, just consult them). Why would that be desirable in any game at all?

You mean endgame tablebases, which in no manner resemble printed books. Databases of endgames are allowed; tablebases are tanatamount to engine use.

If both players access opening databases, they will still get to a position that is within about a half pawn of equality and the use of the database won't significantly favor either player. However, an endgame database can allow you to win or draw an endgame that is so difficult you wouldn't have been able to do it on your own. This means that you can get a significant advantage from an endgame database - a half of the game instead of a half of a pawn. Also, the losing side in an endgame would have no use for the database (they will lose no matter what if the other player uses it), so only one player really gets an advantage in a non-drawn endgame, which is certainly unfair.

SirDavid, your reasoning is flawed on several points, but I'll confine my refutation to your misunderstanding of the site rules.

Database use is not limited to openings, as long as the database does not include computer analysis. Tablebases (what I believe the OP means by endgame databases) ARE computer analysis, and are specifically prohibited:

You may NOT get any help from any person or any chess engine throughout the course of a game, including tablebases. You MAY use books, magazines, or other articles. You may also use Game databases which do not include computer analysis (including Chess.com's Game Explorer) for the Online (correspondence) Chess only.

http://support.chess.com/index.php?_m=knowledgebase&_a=viewarticle&kbarticleid=17

"Game databases which do not include computer analysis" is a very stupid rule, assuming the database is from grandmaster games it will include computer analysis.

I don't think I understood the terminology. I thought of an endgame database and tablebase as the same thing. I was referring to a database (well, tablebase if you prefer) as a list of positions and the evaluation (win, loss, or draw) of each one.

Surley opening databases invlove computer analysis as GMs use computer analysis for opening prep.

I think the distinction is rather that an opening database will not tell you the best move. It will give you a list of moves played historically, how many times each one was used, and what the results were for games with that move. You still have to think about that information to determine what your move will be.

A tablebase will tell you the best move in a given position. It will tell you exactly how to win if you can win, and exactly how to draw if that's the best possible result.

Now, the "opening" databases I mentioned in the first paragraph are really game databases. If you can find your endgame position in one of those games, you could see what someone else did in that position, but that wouldn't necessarily be the best thing to do in that position.

Exactly, thanks. I've corrected it.

THe rule is not stupid if its intent is clear. I agree that the phrasing renders all databases suspect. Indeed, all chess books published in the last few years contain computer analysis because the annotators check their analysis. I believe the intent is to exclude computer analysis of specific opening and endgame positions, not all annotations that might be checked or even informed by analysis. But, that's not what it says.

SirDavid,

The way you are explaining it now, you would seem to exclude the "Table of Computer Database Results for Pawnless Endings" that is published in Muller and Lamprecht, Fundamental Chess Endings (2001). This table does not tell you how to play the ending, but lets you know, for example, that a queen vs. a bishop and knight is a win in 33 moves. It does refer you to a section of the text that explains the ending in some detail. Tablebases, on the other hand, contain every move in every conceivable position with these pieces + kings (five piece tablebases).

It is my understanding that Muller and Lamprecht's book is permitted. Also a database containing games that might have reached this endgame is permissible. But, a tablebase (see Shredder endgame database) is not. For instance, The standard ChessBase database contains search keys that allows quick access to this game following 68...Qxa1.

If I were in a game with a queen against these two pieces, I would be allowed to access my database for the purpose of looking at this game and other similar games. If I found that this game (or the others) had been annotated, I could look at those too.

In short, the distinction in the rules is not between opening databases and others, but between engines analysis, which are forbidden, and instructional materials, which are elements in the appeal of correspondence chess.

Concerning my example of Q vs NB, I checked my database. I've had this endgame once (in over 45,000 games). I was on the losing end in a three or five minute game (almost the only blitz time controls I have ever used at FICS where the game was played), and was saved by the bell.

The position occurs after my opponent's 56...b1Q

My explanation is rooted in the knowledgebase article cited here. But the TOS explains matters differently:

No Cheating or Computer Help

You can NEVER use chess programs (Chessmaster, Fritz, etc) to analyze current ongoing games unless specifically permitted (such as a computer tournament, etc). The only type of computer assistance allowed is games databases for opening lines in Turn-based Chess and Vote Chess. You cannot receive ANY outside assistance on Live Chess games.

http://www.chess.com/legal.html#termsofservice

The TOS does limit database use to opening lines, but does not exclude any books. Hence, according to the TOS, I could consult Muller and Lamprecht, as noted above, but could not employ my ChessBase program to locate similar endgames.

If this distinction was the OP's concern, I, too, question why? SirDavid's explanation, thus, might be correct.

SirDavid has acknowledged the confusion of terminology (tablebase/endgame database) that we see is also present in the usage by the Shredder site linked above. So, it is time to address other flaws in his statement.

"...they will still get to a position that is within about a half pawn of equality and the use of the database won't significantly favor either player."

I think this is a hasty generalization that will not be supported if more data is examined. I also believe that equality in the position is not the criteria for determining whether assistance is acceptable or not.

Large databases contain many opening line busts. Moreover, some lines have a strong scoring percentage favoring one side, but are nevertheless unsound, and not within one-half pawn of equality.

I'm thinking of 9.a3 in a certain line of the Fried Liver Attack as one example. The moves scores 75% for White in my database (16 games), 100% for White in the chess.com Master Games database (2 games), but is -3.00 according to most engines, which favor 9.Kd1.

No. Nor have general clarifications been published such as those we discussed in the Cheating Forum group forum (concerning a specific online database that is disallowed).

This thread brings out another point of potential confusion that warrants clarification from the site's staff.

That's true, I had not considered that. I was trying to make the point that use of a tablebase may favor one side and not the other and that openining database use will favor both players, but I didn't think about bad moves that may have good stats. Thanks for pointing that out.