How to know when you have actually left opening theory

"Out of book" is a fact about a database

When a tool says you left theory on move 9, it is saying one thing only: after that move the position stopped matching any line in the book it consulted. A different book gives a different move number. It is not a judgement, it does not mean the move was bad, and it does not mean your opponent knew more than you did.

This matters because the phrase carries a sting it has not earned. Players read "out of book on move 7" as "you were already lost on move 7", start memorising, and fix nothing. The useful reading is flatter: this is where preparation stopped and thinking started.

The move that leaves theory is rarely the move that goes wrong

In most club games the two are several moves apart. You leave a named line on move 8 with a perfectly reasonable developing move, play four sensible moves after it, and go wrong on move 13 when the position asks a question your preparation never reached. The error is at 13. The interesting fact is at 8, because that is where you stopped being told what to do.

So treat them as two separate marks on the same game. The leaving move tells you the depth of your preparation. The error tells you what you do when preparation runs out. Improving the first is memorisation; improving the second is chess.

Find the exact move, and find who left first

Play the game forward against a named-line reference and note the last position that still matched. The next move is the one that left. Then check whose move it was, because the two cases mean opposite things. If you left first, you chose something outside your own preparation - possibly on purpose, possibly because you forgot it. If your opponent left first, you were the one still holding a book position and facing something new.

The second case is the more common and the more instructive. It means the position after their departure was, at that moment, better understood by you than by them if you could describe it - and describing it is a skill you can practise directly, without learning one extra move of theory.

How Mikhail names it

Mikhail traces your moves through a catalog of named ECO lines and reports the line you were actually in, its ECO code, how many plies stayed in the book, and the exact move that left - including which side played it. The same catalog is the opening database, so you can go and read that line's plan and its per-move notes rather than only its name.

It refuses the claim when the claim would be invented. A game that did not start from the standard position - Chess960, or odds - has no book, so nothing is named. When the trace disagrees with the ECO code the imported game already carried, the trace is treated as a transposition artifact and dropped rather than printed. A named line you were never in is worse than no name at all.

What to do with the position after it

Take the position immediately after theory ended and write three sentences about it before looking at any engine: what the pawn structure asks for, where each side's play is, and what your next two moves are meant to achieve. That is the exercise. If you cannot write the three sentences, more theory will not help you - you do not yet know what the theory was for.

The opening database entries carry the plan behind each line and a note on every move, which is the material for exactly this. Reading why a line plays c4 on move 3 does more for the position on move 14 than memorising three further moves of it.

Three habits that beat memorising more moves

Play the same first move for a while, so your positions repeat and your reviews start to rhyme. After every game, look up the line you were in and read one move past where you left it - one move, not ten. And when your opponent leaves theory first, write down what you played and whether it was still a book move in the resulting position; that habit alone quietly deepens your preparation without a single memorised variation.