Capablanca's Theorem Q+N > Q+B

[Lyudmil Tsvetkov]

Jon, it is not clear to me: which side won, what were the precise conditions?

[Lyudmil Tsvetkov]

In the meantime a test is launched here :

http://tests.stockfishchess.org/tests/v ... 16ff64aff4

the way it goes, it will never succeed.

while watching countless numbers of SF games, I never had the impression SF has problems with this.

one thing I don't understand is why the bonus should be only eg; of course, this is a general mg/eg bonus, and concerns all configurations featuring Q+N, not just QN vs QB.

but engines are crazy things, after all, more than half of the code depends on redundancies and peculiarities, rather than necessarily sound knowledge.

[Nordlandia]

Capablanca's Theorem test positions

[d]8/1p3pkp/2p3p1/p1n1p3/PqP3PP/1P2P3/2Q2PK1/3B4 b - - 0 27

[d]5b2/7k/p3q1p1/2p2p1p/2Np4/3P1PP1/1P1Q1PKP/8 w - - 0 41

[d]6k1/pp1q1ppp/5n2/2p1p3/P1P1P2P/1P1B1PP1/8/5QK1 b - - 0 32

[d]8/p3q1k1/1p1p3p/2pP2pn/P1P2p2/2PQ1P1P/5BPK/8 w - - 0 34

[hgm]

Positions with many Pawns in general favor the Knight, whether accompanied by a Queen or not.

[Lyudmil Tsvetkov]

Capablanca's Theorem test positions

[d]8/1p3pkp/2p3p1/p1n1p3/PqP3PP/1P2P3/2Q2PK1/3B4 b - - 0 27

[d]5b2/7k/p3q1p1/2p2p1p/2Np4/3P1PP1/1P1Q1PKP/8 w - - 0 41

[d]6k1/pp1q1ppp/5n2/2p1p3/P1P1P2P/1P1B1PP1/8/5QK1 b - - 0 32

[d]8/p3q1k1/1p1p3p/2pP2pn/P1P2p2/2PQ1P1P/5BPK/8 w - - 0 34

perfect example of biassed testing.

positions 1-3 heavily favour the knight side due to some blocked/backward pawns on squares the colour of the bishop, others might heavily favours the bishop side, so one really does not quite know how the positions will be split in a random test set.

in order to test perfectly, one needs perfect conditions.

[Nordlandia]

Lyudmil Tsvetkov: you're right, my example positions isn't reliable for this kind of experiment.

May i ask if you can find eligible positions

[Nordlandia]

Maybe this position qualifies as eligible.

[d]6k1/1p2q1p1/p3b2p/3pP3/3Pp3/P1N3P1/1P5P/5QK1 w - - 0 24

https://lichess.org/nNlotAY8#46

Source:

bishop vs knight the verdict - [1997]

Practical Endgame Play - Beyond the Basics - [2007]

[MikeGL]

Maybe this position qualifies as eligible.

[d]6k1/1p2q1p1/p3b2p/3pP3/3Pp3/P1N3P1/1P5P/5QK1 w - - 0 24

https://lichess.org/nNlotAY8#46

Source:

bishop vs knight the verdict - [1997]

Practical Endgame Play - Beyond the Basics - [2007]

[d]3qk1n1/2pppppp/8/8/8/8/2PPPPPP/3QKB2 w - - 1 1

Maybe something like this position.

.

[Lyudmil Tsvetkov]

Lyudmil Tsvetkov: you're right, my example positions isn't reliable for this kind of experiment.

May i ask if you can find eligible positions

I can't.

you are much better at that stuff.

I guess none of existing chess software has the necessary filtering options to successfully mask out an acceptable set, so maybe we should just test with engine games from a neutral position.

[Lyudmil Tsvetkov]

Maybe this position qualifies as eligible.

[d]6k1/1p2q1p1/p3b2p/3pP3/3Pp3/P1N3P1/1P5P/5QK1 w - - 0 24

https://lichess.org/nNlotAY8#46

Source:

bishop vs knight the verdict - [1997]

Practical Endgame Play - Beyond the Basics - [2007]

who was the author of those 'insights'?

whatever conclusions he made, he counted in irrelevant positions, never checked what the position looked like, but just if QN, QB were present.

it is easy like that, but your conclusions should be mostly wrong.

how could a bishops be weaker than a knight in the endgame?

this simply makes no sense, it is obvious that the small sample of less than 50 positions contained mainly noise/irrelevant/random games, where the knight has been favoured in some way.

concerning your position, this is simply obvious draw, the 2 passers make so that perpetual queen check for one side is unavoidable.