Is it legal to have a chess engine in a fide rated tournament?

[Uri Blass]

I am not sure but I found the following by fide that suggest that it is legal.

https://handbook.fide.com/chapter/E03

The main problem with playing with a knight handicap for fide rating is that I see no logical opening book that lead to normal knight handicap position.

An engine may have an opening book like
1.Nf3 Nc6 2.Ne5 Nxe5 3.Nc3 Nc6 4.Nb1 Nb8 but black does not have to follow the book at moves 3,4 and move 4 is illogical and people can claim that forcing the human to play some opening book line is against the rules.

[lkaufman]

I am not sure but I found the following by fide that suggest that it is legal.

https://handbook.fide.com/chapter/E03

The main problem with playing with a knight handicap for fide rating is that I see no logical opening book that lead to normal knight handicap position.

An engine may have an opening book like
1.Nf3 Nc6 2.Ne5 Nxe5 3.Nc3 Nc6 4.Nb1 Nb8 but black does not have to follow the book at moves 3,4 and move 4 is illogical and people can claim that forcing the human to play some opening book line is against the rules.

I recall at one point that FIDE declared a $20,000 cost to have a chess engine get a FIDE rating, I don't know if that is a current rule or not. As for rating tournaments with specified openings, "thematic tournaments", I don't recall any precedent. Logically it would not make sense to rate games between humans at knight odds for example, but for an engine where it would be called a knight odds rating it's less clear. FIDE has rated events with the "football" 3-1-0 scoring and even with stalemate getting more than 0.5 score, so they seem to be somewhat flexible.

[Uri Blass]

I am not sure but I found the following by fide that suggest that it is legal.

https://handbook.fide.com/chapter/E03

The main problem with playing with a knight handicap for fide rating is that I see no logical opening book that lead to normal knight handicap position.

An engine may have an opening book like
1.Nf3 Nc6 2.Ne5 Nxe5 3.Nc3 Nc6 4.Nb1 Nb8 but black does not have to follow the book at moves 3,4 and move 4 is illogical and people can claim that forcing the human to play some opening book line is against the rules.

I recall at one point that FIDE declared a $20,000 cost to have a chess engine get a FIDE rating, I don't know if that is a current rule or not. As for rating tournaments with specified openings, "thematic tournaments", I don't recall any precedent. Logically it would not make sense to rate games between humans at knight odds for example, but for an engine where it would be called a knight odds rating it's less clear. FIDE has rated events with the "football" 3-1-0 scoring and even with stalemate getting more than 0.5 score, so they seem to be somewhat flexible.

If the only problem is that the computer does not get a fide rating and games against the computer are not for rating then it is not so important(it is possible to calculate an estimate for the rating of the computer based on the results).

A bigger problem may be if fide refuse to consider the games of humans against humans in the same tournament for fide rating.

Of course it is possible to make a tournament not for fide rating but in this case I am afraid
it is going to be harder to find chess clubs who are going to support organizing the tournament.

[lkaufman]

I am not sure but I found the following by fide that suggest that it is legal.

https://handbook.fide.com/chapter/E03

The main problem with playing with a knight handicap for fide rating is that I see no logical opening book that lead to normal knight handicap position.

An engine may have an opening book like
1.Nf3 Nc6 2.Ne5 Nxe5 3.Nc3 Nc6 4.Nb1 Nb8 but black does not have to follow the book at moves 3,4 and move 4 is illogical and people can claim that forcing the human to play some opening book line is against the rules.

I recall at one point that FIDE declared a $20,000 cost to have a chess engine get a FIDE rating, I don't know if that is a current rule or not. As for rating tournaments with specified openings, "thematic tournaments", I don't recall any precedent. Logically it would not make sense to rate games between humans at knight odds for example, but for an engine where it would be called a knight odds rating it's less clear. FIDE has rated events with the "football" 3-1-0 scoring and even with stalemate getting more than 0.5 score, so they seem to be somewhat flexible.

If the only problem is that the computer does not get a fide rating and games against the computer are not for rating then it is not so important(it is possible to calculate an estimate for the rating of the computer based on the results).

A bigger problem may be if fide refuse to consider the games of humans against humans in the same tournament for fide rating.

Of course it is possible to make a tournament not for fide rating but in this case I am afraid
it is going to be harder to find chess clubs who are going to support organizing the tournament.

So I think the question you need an answer to is "Can a tournament be FIDE rated with some games specified in advance as not for FIDE rating?" I don't know the answer, but there is a sort of precedent, in that many American tournaments had some rounds of some sections played at fast time limits that were not FIDE rated, while the slower games were FIDE rated. So I would say that it is allowed, but I'm not the one making the call.

[Milos]

I am not sure but I found the following by fide that suggest that it is legal.

https://handbook.fide.com/chapter/E03

The main problem with playing with a knight handicap for fide rating is that I see no logical opening book that lead to normal knight handicap position.

An engine may have an opening book like
1.Nf3 Nc6 2.Ne5 Nxe5 3.Nc3 Nc6 4.Nb1 Nb8 but black does not have to follow the book at moves 3,4 and move 4 is illogical and people can claim that forcing the human to play some opening book line is against the rules.

You can't do it with one line, but it is fairly trivial (10 lines max) to have a book where white sacs a knight for a pawn at move 3 no matter what black plays.

[lkaufman]

I am not sure but I found the following by fide that suggest that it is legal.

https://handbook.fide.com/chapter/E03

The main problem with playing with a knight handicap for fide rating is that I see no logical opening book that lead to normal knight handicap position.

An engine may have an opening book like
1.Nf3 Nc6 2.Ne5 Nxe5 3.Nc3 Nc6 4.Nb1 Nb8 but black does not have to follow the book at moves 3,4 and move 4 is illogical and people can claim that forcing the human to play some opening book line is against the rules.

You can't do it with one line, but it is fairly trivial (10 lines max) to have a book where white sacs a knight for a pawn at move 3 no matter what black plays.

But then Black will play whichever line leads to the least compensation for the knight, resulting in some weird position that doesn't resemble real knight odds.

[Milos]

But then Black will play whichever line leads to the least compensation for the knight, resulting in some weird position that doesn't resemble real knight odds.

One can tell black i.e. human what is white's strategy. For example white's strategy is to sac knight for a pawn on 7th row in move 3 (and white can force that no matter what black plays). So then black can decide what opening provides him best chances and play responses in moves 1 and 2 accordingly. The only ambiguity really would be what pawn on row 7 does white take on move 3 (if it can be chosen), but the fairest there is that white decides randomly. So if black values one pawn more than the other than black can force it.

[lkaufman]

But then Black will play whichever line leads to the least compensation for the knight, resulting in some weird position that doesn't resemble real knight odds.

One can tell black i.e. human what is white's strategy. For example white's strategy is to sac knight for a pawn on 7th row in move 3 (and white can force that no matter what black plays). So then black can decide what opening provides him best chances and play responses in moves 1 and 2 accordingly. The only ambiguity really would be what pawn on row 7 does white take on move 3 (if it can be chosen), but the fairest there is that white decides randomly. So if black values one pawn more than the other than black can force it.

Assuming QN odds, play might go 1Nc3 Nf6 2Nd5 (not Nb5 as White could be forced to take the low-value rook pawn) e6 3.Nxc7+ QxN. But Black already has huge development lead for the pawn, probably it's already worse for White than knight odds, and in any case not similar to it.

[Milos]

Assuming QN odds, play might go 1Nc3 Nf6 2Nd5 (not Nb5 as White could be forced to take the low-value rook pawn) e6 3.Nxc7+ QxN. But Black already has huge development lead for the pawn, probably it's already worse for White than knight odds, and in any case not similar to it.

Well if you want repeatable and balanced starting position, than it's like original proposal. White has 2 move book with a single line:
1.Nc3 ... 2.Ne5
After that SF continues.
And black knows this. So essentially every game would start from:

[d]rnbqkb1r/pppppppp/8/8/4n3/8/PPPPPPPP/R1BQKBNR w KQkq - 0 3
Ofc, black might refuse to play 1...Nf6 2...Nxe5, but then black might also resign on the first move (which is totally equivalent to not capturing the white knight).

[lkaufman]

Assuming QN odds, play might go 1Nc3 Nf6 2Nd5 (not Nb5 as White could be forced to take the low-value rook pawn) e6 3.Nxc7+ QxN. But Black already has huge development lead for the pawn, probably it's already worse for White than knight odds, and in any case not similar to it.

Well if you want repeatable and balanced starting position, than it's like original proposal. White has 2 move book with a single line:
1.Nc3 ... 2.Ne5
After that SF continues.
And black knows this. So essentially every game would start from:

[d]rnbqkb1r/pppppppp/8/8/4n3/8/PPPPPPPP/R1BQKBNR w KQkq - 0 3
Ofc, black might refuse to play 1...Nf6 2...Nxe5, but then black might also resign on the first move (which is totally equivalent to not capturing the white knight).

OK, at least that's fairly close to real N odds, after 3.d3 Nf6 4.e4 it's like Alekhine's defense with 2.d3, all reasonable moves. Giving the knight away on d5 might be a bit better for White, after 3.e4 Nf6 (not forced) it would again be Alekhine's defense, but White could choose e5 instead of d3.