Queen vs. Bishop, Bishop and Knight: What Tablebases Really Say

0. Short Summary

In classical endgame literature, the endgame Queen vs. Bishop, Bishop and Knight (Q vs. BBN, with the two belonging to the same side and placed on opposite-colored squares) is usually described as “generally won” for the three minor pieces.

Such evaluations were necessarily based on human intuition, practical experience, and limited analysis.

Using complete Nalimov Tablebases (6 men, no pawns), I conducted a systematic and exhaustive investigation of this endgame, covering all legal positions under clearly defined conditions.
The results strongly contradict the traditional assessment.

1. Methodology

Material: K+Q vs. K+BBN, no pawns
Tablebases: Nalimov (6-piece)
All legal positions included

Positions classified by:

• side to move,
• whether the queen-side king is in check,
• theoretical outcome (win / draw / loss),
• distance-to-mate where applicable.

No engine evaluations, heuristics, or pruning were used — Tablebases only.

2. Main result

With the queen side to move and the queen-side king not in check, over 90% of all legal positions are winning for the queen.

Statistics:

Queen side to move, king not in check:
Win 90.75% · Draw 9.09% · Loss 0.15%.

BBN side to move, king not in check:
Win 39.46% · Draw 39.08% · Loss 21.46%.

Thus, the widespread claim that Q vs. BBN is “generally won” for the three minor pieces is statistically misleading.

3. Why this result is non-intuitive

From a human perspective, BBN appears to have:

• control of both color complexes,
• strong mating potential,
• good apparent coordination.

Tablebases reveal a different picture:

• the first accurate queen check often disrupts coordination;
• the knight frequently becomes the weakest link;
• maintaining harmony between three minor pieces over many moves is extremely fragile.

The queen typically wins not by immediate tactics, but by gradual structural degradation.

4. Examples

Representative positions — including:

• typical queen wins,
• fast and correct BBN wins,
• extremely long forced mates by the queen (exceeding 100 moves),
• and rare drawing fortresses —

are available in the accompanying Lichess Study:

5. Verification

In addition to direct Tablebases enumeration, the results were independently verified using Wilhelm (v. 1.50, by Rafael Andrist), which provides direct access to Nalimov Tablebases and allows systematic exploration of long forced lines, zugzwangs, and unstable configurations.
All conclusions presented here are fully consistent between Nalimov Tablebases and Wilhelm-based analysis.

6. Conclusion

Q vs. BBN is not “generally won” for the three minor pieces.
Under common and clearly defined conditions, the endgame is strongly biased in favor of the queen.

This case highlights the limitations of heuristic endgame theory and the necessity of exhaustive Tablebases analysis in low-piece endgames.

Author: Musin Bulat.
Date: February 28, 2026.

Quite interesting.
However, a position does not appear out of nowhere.
The side with KBBN will try to put pieces on squares so they mutually protect each other.

Interesting post.
I imagine that it might be different for human players though. As the player with the queen, it's extremely hard to find the correct sequence of moves. Because on a nearly empty board, the Queen has so many "potentially good" squares at every turn, it becomes impossible to calculate all of them and you will go wrong easily.

Fot the side with BBN, promising drawing or even winning plans (e.g. to build a fortress) may be easier to find for a human brain.
That said, I'm crap at endgames, so would be interesting to hear what endgame experts think.

It would be interesting to look at a few examples where this occurred in real games.

@tpr said ^

It would be interesting to look at a few examples where this occurred in real games.

See, for example, the game Rozentalis (2623) - Sutovsky (2657), 05.02.2010.
After White's 95th move, they got drawn position KQ vs. KBBN.
And finally, draw after White's 109th move.

@tpr said ^

Quite interesting.
However, a position does not appear out of nowhere.
The side with KBBN will try to put pieces on squares so they mutually protect each other.

The easiest way to get an assured draw is to to sacrifice an extra bishop and construct a Karstaedt fortress, e.g., Bishop on g7, Knight on e5, King on g8, h7, h8.

@Panagrellus said ^

That said, I'm crap at endgames, so would be interesting to hear what endgame experts think.

Actually, endgame "Queen vs. three minor pieces" is not studied scrutinizingly yet.

Grandmaster John Nunn, the well-known endgame expert, even didn't mention that type of endgame in his book "Secrets of
Pawnless Endings" (2nd ed., 2002.)