Queens vs Knights

[hgm]

[d]nnnnknnn/pppppppp/8/8/8/8/PPPPPPPP/1Q1QK1Q1 w

Score around +8.5 in a lost position.

[Moderation] This was originally posted in another thread, but because the following discussion was too much off-topic there, it was split off.

[Lyudmil Tsvetkov]

[d]nnnnknnn/pppppppp/8/8/8/8/PPPPPPPP/1Q1QK1Q1 w

Score around +8.5 in a lost position.

shall we quarrel on that one again?

minors vs queen really do gain some nice bonus, but 6 full pawns is way too much.

no one demonstrated until now how this position plays out.

besides, your position is illegal for normal chess: 3 queens + 8 pawns.

[Nordlandia]

[d]nnnnknnn/pppppppp/8/8/8/8/PPPPPPPP/1Q1QK1Q1 w

Score around +8.5 in a lost position.

Fritz don't allow illegal positions to be analysed. Can you recreate that position legally. About same position with pawns removed in exchange for promoted pieces.

Position is lost due to extensive queen redundancy.

[Lyudmil Tsvetkov]

I think main message should be chess is pretty much meanigless game, please do not waste your time on it, as there are no sensible rules.

for example, the knight, the weakest non-pawn piece, is worth some 3 full pawns, while the queen, the strongest piece, is worth precisely 0 full pawns.

[d]8/4n3/4k1n1/8/4K3/8/8/8 b - - 0 1

draw

[d]8/3n4/4kn2/8/8/4KQ2/8/8 w - - 0 1

draw again

so, it basically does not matter if one has spare queen or not.

[Nordlandia]

Queen redundancy

SF Dev -16

Komodo: -1.0

[d]3qqqkq/5ppq/7q/8/3N4/3NNN2/3NNPPP/2NBBKRR w - - 0 1

SF Dev: -2.6

Komodo: +0.70

[d]2qqkq2/1pppppp1/8/8/8/8/1PPPPPP1/1NBQKBNN w - - 0 0

[Lyudmil Tsvetkov]

[d]nnnnknnn/pppppppp/8/8/8/8/PPPPPPPP/1Q1QK1Q1 w

Score around +8.5 in a lost position.

Fritz don't allow illegal positions to be analysed. Can you recreate that position legally. About same position with pawns removed in exchange for promoted pieces.

Position is lost due to extensive queen redundancy.

no one ever proved that.

6 full pawns are quite a lot.

besides, such positions never happen is normal chess.

in normal chess, you will have a single queen vs couple of minor pieces, or couple of minors and rooks, and the bonus for the non-queen side is usually quite smaller.

with the so-called queen redundancy, one side would have just a single queen more, 2 vs one, or 3 vs 2, so the effect is much smaller.

I also do believe that pair of queen are worth a bonus instead of penalty, for unusual coordination and swift king attacks.

but once you must have already penalised the queen against the minor pieces, otherwise nothing will work.

[hgm]

Fritz don't allow illegal positions to be analysed.

Fortunately the world is flooded with GUIs that do not have this defect. More serious is when the engine would refuse it.

Can you recreate that position legally. About same position with pawns removed in exchange for promoted pieces.

You could try this one:
[d]1nnnknn1/3ppp2/2n2n2/8/8/8/3PPP2/Q2QK2Q w

Position is lost due to extensive queen redundancy.

One would think Knight redundancy would be even larger...

In fact redundancy is a red herring. The position is lost by the 'Elephantiasis effect', identified by Reinhard Scharnagl. (Perhaps we should rename it the 'Scharnagl effect' to honor his memory?): Strong pieces lose value in the presence of opponent weaker pieces. And there is no way to relax that here, for lack of suitable trading partners.

[Nordlandia]

Fritz don't allow illegal positions to be analysed.

Fortunately the world is flooded with GUIs that do not have this defect. More serious is when the engine would refuse it.

Can you recreate that position legally. About same position with pawns removed in exchange for promoted pieces.

You could try this one:
[d]1nnnknn1/3ppp2/2n2n2/8/8/8/3PPP2/Q2QK2Q w

Position is lost due to extensive queen redundancy.

One would think Knight redundancy would be even larger...

In fact redundancy is a red herring. The position is lost by the 'Elephantiasis effect', identified by Reinhard Scharnagl. (Perhaps we should rename it the 'Scharnagl effect' to honor his memory?): Strong pieces lose value in the presence of opponent weaker pieces. And there is no way to relax that here, for lack of suitable trading partners.

That's correct. Also strong pieces like the queen, is be more easily affected by Redundancy/Scharnagl effect. Multiple knights never occurs in practise beside setup position like this.

Analysis by Komodo 11.01 64-bit:

1.Qdb1 d5 2.Qa4 Nbd7 3.Qa6 Ncb6
+- (3.61) Depth: 33 00:00:25 371MN, tb=561
(16.07.2017)

Analysis by Stockfish 150717 64 BMI2:

1.e4 e6 2.d3 d5 3.f4 Nbd7
+- (7.10) Depth: 21/27 00:00:04 82615kN, tb=10
(16.07.2017)

[hgm]

OK, so Komodo is not completely naive. (Or is that because you analyzed so deep that the loss of the first Queen is already within the horizon? How much does it give at low depth?) Still, +3.6 is pretty bad for a hopelessly lost position.

In other positions Komodo's evaluation seems far off, though. The 3Q versus 4N+2B positions looks like a total win for the Queens to me.

[Nordlandia]

Mark said that he couldn't find any extra queen penalty in SF8 code.

Komodo is better adjusted for imbalances, such as this.