White 304 vs Black 386
White starts from setup 304 (BBNQRKRN), Black from setup 386 (BQRNNBKR). In the encyclopedia since 2026-10-15: neither side starts better.
|cp| ≤ 5 at 400M · 800M · 1.6B · 4B nodes, two independent engines agree
How to play with different setups
White sets up the pieces by position No. 304, Black by position No. 386, mirrored on their own side; the pawns stand as usual.
Moves as usual: White starts; everything else follows the Chess960 rules.
The two setups
White: setup #304 B B N Q R K R N
King on the f-file, rooks on e, g, queen on the d-file; one 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 304 gives White +0.29 (Morozov Balance Index, 400M nodes).
Black: setup #386 B Q R N N B K R
King on the g-file, rooks on c, h, queen on the b-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 386 gives White +0.14 (Morozov Balance Index, 400M nodes).
Why the pair is balanced
At the deepest rungs (400M, 800M, 1.6B 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 -3 and +26 cp before settling — a reminder that quick engine checks overstate imbalances, and why the ladder goes deep.
For comparison, in symmetric play setup 304 gives White +0.29 and setup 386 gives White +0.14: 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 c4; the alternatives are slightly weaker but keep the game level.
Along the engine's main line (21 moves) the evaluation stayed within ±27 cp the whole way: the balance holds as the pieces come out, not only on move one.
Engine self-play agrees: 10 games, White won 0, drawn 7, Black won 3 (White scored 35.0%).
Verification ladder
Nine independent searches of the starting position, each with a larger node budget. Evaluation in centipawns from White's point of view.
| Stage | Nodes | Depth | Eval, cp | Main line |
|---|---|---|---|---|
| S0 | 100 000 | 16 | -3 | 1. b3 b5 2. Nd3 c5 3. c4 bxc4 4. bxc4 Qxh2 |
| S1 | 500 000 | 20 | 0 | 1. c4 c5 2. b3 b5 3. cxb5 Qxh2 4. e3 Bxg2+ |
| S2 | 2 000 000 | 24 | +26 | 1. c4 c5 2. b3 b5 3. cxb5 h5 4. O-O Qxb5 |
| S3 | 8 000 000 | 30 | +23 | 1. c4 c5 2. b3 b5 3. cxb5 Qxb5 4. Nd3 d5 |
| S4 | 32 000 000 | 35 | +14 | 1. c4 c5 2. b3 b5 3. cxb5 h5 4. Nd3 Qxb5 |
| S5 | 128 000 000 | 41 | +3 | 1. c4 c5 2. b3 b5 3. cxb5 Qxb5 4. Nd3 h5 |
| S6 | 400 000 000 | 52 | 0 | 1. c4 c5 2. b3 b5 3. cxb5 Qxb5 4. Nd3 d5 |
| S7 | 800 000 000 | 57 | 0 | 1. c4 c5 2. b3 b5 3. cxb5 Qxb5 4. Nd3 d5 |
| S8 | 1 600 000 000 | 58 | 0 | 1. c4 c5 2. b3 b5 3. cxb5 Qxb5 4. Nd3 d5 |
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.
| Nodes | Eval, cp |
|---|---|
| 4 000 000 000 | 0 |
| 8 000 000 000 | not checked |
| 16 000 000 000 | not checked |
This pair stays within the ±5 cp corridor up to 4 000 000 000 nodes.
Independent engines, run on the same position:
- Berserk 14: +1
- Lc0 0.32.1: -7
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 +138, Black +122, difference +16. A balanced pair needs White's setup to be weaker by about the price of the first move.
Extended self-play: 10 games, 0 White wins, 7 draws, 3 Black wins, White scored 35.0%.
Best first moves
The 3 best first moves for White, searched together at 400 000 000 nodes.
| # | Eval, cp | Depth | Main line |
|---|---|---|---|
| 1 | -4 | 41 | 1. c4 c5 2. b3 b5 3. cxb5 Qxb5 4. O-O h5 5. Nd3 h4 |
| 2 | -25 | 40 | 1. e4 c5 2. c4 b5 3. cxb5 Qxb5+ 4. Ne2 h5 5. O-O Qb6 |
| 3 | -29 | 41 | 1. e4 c5 |
Main-line play-out
21 moves, 5 000 000 nodes per half-move. Largest |evaluation| 27 cp, mean 4.1 cp, final 0 cp.
| # | White | cp | Black | cp |
|---|---|---|---|---|
| 1 | c4 | +27 | c5 | +27 |
| 2 | b3 | +18 | b5 | +14 |
| 3 | cxb5 | +15 | Qxb5 | +5 |
| 4 | O-O | 0 | h5 | 0 |
| 5 | Nd3 | +2 | h4 | -10 |
| 6 | f3 | -6 | h3 | -4 |
| 7 | g3 | 0 | d6 | -16 |
| 8 | e4 | -20 | Nf6 | -12 |
| 9 | Nhf2 | 0 | Ne6 | 0 |
| 10 | f4 | 0 | Qb6 | 0 |
| 11 | Nb2 | 0 | Qb7 | 0 |
| 12 | Nc4 | 0 | Nd4 | 0 |
| 13 | Ne3 | 0 | e6 | 0 |
| 14 | f5 | 0 | exf5 | 0 |
| 15 | Nxf5 | 0 | Nxf5 | 0 |
| 16 | Bxf6 | 0 | gxf6 | 0 |
| 17 | Ng4 | 0 | Nh6 | 0 |
| 18 | Nxf6+ | 0 | Kg7 | 0 |
| 19 | Nh5+ | 0 | Kg8 | 0 |
| 20 | Nf6+ | 0 | Kg7 | 0 |
| 21 | Nh5+ | 0 | Kg8 | 0 |
| 22 | Nf6+ | 0 |
Engine self-play
10 games at 500 000 nodes per move: White won 0, drawn 7, Black won 3 — White scored 35.0%. Average length 43.5 moves.
Sample game 1 (1/2-1/2)
1. c4 c5 2. O-O e6 3. e4 b5 4. cxb5 h5 5. b3 Qxb5 6. Nd3 h4 7. f3 h3 8. g3 Nc6 9. Nhf2 Qb6 10. f4 Bb7 11. Nb2 Qd8 12. e5 Nd4 13. Be4 d5 14. exd6 Nxd6 15. Bxb7 Nxb7 16. Ng4 g6 17. Nc4 Bg7 18. Nce5 Na5 19. Bxd4 Qxd4+ 20. Rf2 c4 21. Qb1 Qd5 22. Ne3 Qd6 23. b4 Nc6 24. N3xc4 Qc7 25. Qe4 Nxe5 26. Nxe5 O-O 27. g4 Bf6 28. Nf3 Qb6 29. a3 a5 30. bxa5 Qxa5 31. Rfe2 Qc5+ 32. Kf1 Rcd8 33. Qb4 Qc7 34. Qe4 Qc5 35. Qb4 Qa7 36. Qe4
Sample game 2 (0-1)
1. Nd3 c5 2. b3 b5 3. c4 bxc4 4. bxc4 Qxh2 5. Ng3 Qh6 6. e4 Qa6 7. f4 h5 8. Rh1 Rb8 9. Kg1 h4 10. Nf5 Qxc4 11. Ne5 Qe6 12. Rh3 Nf6 13. Bd3 Rb4 14. Bc3 Rxe4 15. Bxe4 Bxe4 16. Nxh4 c4 17. d3 Qb6+ 18. Kh2 Bxd3 19. Nxd3 cxd3 20. g3 e5 21. fxe5 Nd5 22. Bd2 Qf2+ 23. Ng2 Rxh3+ 24. Kxh3 Nc6 25. Rf1 Qd4 26. Bf4 Nxf4+ 27. Rxf4 Qd5 28. Qb3 Qe6+ 29. Qxe6 fxe6 30. Ne3 Nxe5 31. Re4 Bd6 32. Ng4 Bc5 33. Re1 Nc4 34. Rd1 d2 35. Kg2 d5 36. Kf3 e5 37. Ke2 Kf7 38. Rb1 Ke6 39. Rb5 Bb6 40. a4 e4 41. Rb3 e3 42. Nh2 d4 43. Nf3 Kd5 44. g4 Nd6 45. Nxd2 exd2 46. Kxd2 Kc4 47. Rh3 Ne4+ 48. Kc2 g5 49. Ra3 Bc5 50. Rb3 a5 51. Rh3 Nf2 52. Rf3 d3+ 53. Kd2 Ne4+ 54. Kc1
Play it in the CH960 app
Open this page on a phone with the CH960 app installed, or find pair 304 / 386 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): bqrnnbkr/pppppppp/8/8/8/8/PPPPPPPP/BBNQRKRN w EGch - 0 1
Mirror twin: White 367 vs Black 759 — a different position (castling squares are fixed), not evaluated by the search.