Which first move is best — and why

Setup no. 489 — your move

Next setup

How a move is judged

By what it LOSES against the best move, not by its place in a top three. The reason is measured, not assumed: the top of the list sticks together — in a typical setup about five first moves out of twenty are indistinguishable — and the order inside that cluster changes between search budgets, while the loss does not.

So "as good as the best" is a band, not a single square. Inside that band the trainer says so plainly instead of correcting you: being told off for a move the engine itself cannot tell apart from the best one teaches nothing.

What the numbers are

• Engine Stockfish 17.1, one thread, a fixed budget of 100 million nodes per setup, evaluation taken from the last COMPLETED iteration — the run reproduces bit for bit.
• ALL legal first moves are evaluated, not just the strongest few: whatever you play, the answer comes from the table — instantly, offline and identically on every device.
• For each move the engine's main line is kept for eight half-moves: that is the "where it would lead" line under the verdict.

Table generated 2026-10-05.

Open data · Methodology · The whole table as JSON

Why the first move

In classical chess the first move comes from books. In Chess960 there are no books: a setup you have never seen appears on the board and the clock starts. Deciding there, from scratch, is the skill this trainer builds — and it is the one skill classical opening study cannot give you.

The explanation is assembled from facts that can be checked on the board — occupying the centre, developing a piece, freeing a bishop locked in by its own pawns (a Chess960 disease), making castling possible — plus the engine's own main line. It is not a grandmaster's commentary, and we do not pretend otherwise.