White 70 vs Black 639

The two setups

White: setup #70 N N B R K B Q R

King on the e-file, rooks on d, h, queen on the g-file; no bishop starts in a corner; the knights start away from the centre. Castling follows the Chess960 rule: king and rook end on the classical squares. In ordinary symmetric play setup 70 gives White +0.40 (Morozov Balance Index, 400M nodes).

Black: setup #639 R N K Q R N B B

King on the c-file, rooks on a, e, queen on the d-file; one bishop starts in a corner; the knights start on central files. The king already stands on a castled square, so one of the castlings moves only the rook. In ordinary symmetric play setup 639 gives White +0.30 (Morozov Balance Index, 400M nodes).

Why the pair is balanced

At the deepest rungs (400M, 800M, 1.6B, 3.2B nodes) the evaluation is exactly 0.00 each time, reaching search depth 58.

Admission rule: |cp| ≤ 5 at four budgets up to 4 billion nodes, no one-sided drift along the ladder, and confirmation by two independent engines — Berserk and Lc0 — each below its own median on symmetric positions. The same corridor is required at every deeper budget (8 G, 16 G) at which the pair has been evaluated.

The shallow rungs wandered between -24 and +23 cp before settling — a reminder that quick engine checks overstate imbalances, and why the ladder goes deep.

For comparison, in symmetric play setup 70 gives White +0.40 and setup 639 gives White +0.30: with both sides on the same rank the first move is worth something. In this pair Black's own setup absorbs that edge completely.

The engine's first choice for White is d4; the alternatives are slightly weaker but keep the game level.

Along the engine's main line (23 moves) the largest swing was 45 cp (around move 1), ending at 0 cp — the position stays playable for both sides.

Engine self-play agrees: 10 games, White won 0, drawn 10, Black won 0 (White scored 50.0%).

Verification ladder

Nine independent searches of the starting position, each with a larger node budget. Evaluation in centipawns from White's point of view.

Deep verification

The search ladder reaches 1.6 billion nodes; admission needs a 4-billion rung and two further engines on top of that. Beyond even that, every pair was re-checked at budgets up to ten times larger and in thousands of self-play games. This section shows what each check returned for this pair.

This pair stays within the ±5 cp corridor up to 16 000 000 000 nodes.

Swap the two setups, giving White the black back rank and Black the white one, and the evaluation jumps to +115. The same two armies dealt the other way round are no longer balanced at all: that is how precisely this pair pays for the first move.

Independent engines, run on the same position:

• Berserk 14: -5
• Lc0 0.32.1: -11

Independent confirmations: 3 of 3. An engine counts as confirming when its reading of the pair is closer to equality than its own reading of a typical symmetric Chess960 position (+10 for Berserk, +18 for Lc0). Where only one engine confirms, the pair is balanced according to that engine alone, and this page says so.

Own weight of each setup, measured against the classical array: White +43, Black +131, difference -88. A balanced pair needs White's setup to be weaker by about the price of the first move.

Extended self-play: 160 games, 0 White wins, 146 draws, 14 Black wins, White scored 45.6%.

Best first moves

The 3 best first moves for White, searched together at 400 000 000 nodes.

Main-line play-out

23 moves, 5 000 000 nodes per half-move. Largest |evaluation| 45 cp, mean 7.1 cp, final 0 cp.

Engine self-play

10 games at 500 000 nodes per move: White won 0, drawn 10, Black won 0 — White scored 50.0%. Average length 30.6 moves.

1. d4 f5 2. Nc3 g6 3. g4 Nc6 4. f4 Nb4 5. gxf5 Nxa2 6. Nxa2 Bxa2 7. fxg6 Nxg6 8. f5 Nh4 9. b3 Bf6 10. Qg4 a5 11. e4 a4 12. e5 axb3 13. Nxb3 Bxb3 14. cxb3 Ra2 15. Rd2 Ra1 16. Rd1 Ra2 17. Rd2 Ra1 18. Rd1 Ra2

1. d4 g6 2. g4 f5 3. gxf5 Nc6 4. fxg6 Nxg6 5. Qg4 Bxa2 6. Nc3 Bc4 7. h4 e6 8. Bg5 Nge7 9. Nb3 Rg8 10. Qe4 Qf8 11. Qxh7 O-O-O 12. e3 Bxb3 13. cxb3 Bf6 14. Bh6 Bg7 15. Bg5 Bf6 16. Bh6 Bg7 17. Bxg7 Rxg7 18. Qe4 Rf7 19. Rh2 Kb8 20. Bd3 d5 21. Qxe6 Nf5 22. Qg6 Nfxd4 23. exd4 Rg7 24. Qh5 Nxd4 25. Rh1 Rg2 26. Rd2 Qf4 27. Nxd5 Nf3+ 28. Kf1 Nxd2+ 29. Kxg2 Rg8+ 30. Bg6 Qf7 31. Kh3 Rxg6 32. Nf4 Qd7+ 33. Kh2 Rg8 34. Qd5 Qxd5 35. Nxd5 Nf3+ 36. Kh3 Ng1+ 37. Kh2 Nf3+ 38. Kh3 Ng1+ 39. Kh2

Play it in the CH960 app

Open this page on a phone with the CH960 app installed, or find pair 70 / 639 in the app's encyclopedia, then choose how to play.

• Local game — two players on one device, with a clock.
• With a mentor — a clearly labelled training bot with hints, at your level.
• Challenge a friend — an online game by link; both players get the same pair, colours swap on rematch.

Play this pair against the computer, right here Play a friend

FEN (Shredder, castling rights by rook files): rnkqrnbb/pppppppp/8/8/8/8/PPPPPPPP/NNBRKBQR w DHae - 0 1

Mirror twin: White 886 vs Black 224 — a different position (castling squares are fixed), not evaluated by the search.

Current place in the live balance rating: 19.