Training Mode: Practicing One Motif at a Time

Most puzzle training is mixed: you get a position, you find the tactic, you move on to an unrelated position. That is good for testing whether you can spot something, and much less good for learning to spot it in the first place.

Training mode does the opposite. You pick a motif — forks, pins, skewers, mating nets, endgames — and solve positions that all contain it, with no timer running.

The Problem With Mixed Solving

When puzzles arrive in random order, you are being tested on recall. You either see the pattern or you do not, and when you fail you get a solution, a small amount of disappointment, and a new unrelated position.

That works if the pattern is already in your head. If it is not, mixed solving is a slow way to put it there. You meet a particular kind of deflection once every forty puzzles, which is not enough repetition for it to become automatic.

This is why players sometimes grind thousands of puzzles and still miss the same motifs in their games. The volume is there; the concentration is not.

What Isolating a Theme Does

Solving twenty pins in a row is a different exercise from solving twenty mixed puzzles. By the fifth one you stop hunting randomly and start looking for the specific geometry a pin requires — a valuable piece on a line behind a less valuable one.

That is the actual skill. Not "can I find the answer to this position", but "do I automatically check whether this geometry exists". Concentrated repetition builds the second one much faster than scattered exposure does.

It also makes your failures informative. If you miss six skewers in a row, you have learned something specific and actionable about your vision, which a mixed session would have hidden by spreading those failures across a hundred puzzles.

Why There Is No Timer

Training mode deliberately has no clock. Timed modes train speed of recognition for patterns you already know, which is a real and separate skill — but under time pressure you fall back on what is already automatic, which is exactly the wrong condition for learning something new.

Without a timer you can sit with a position, work out why the tactic exists, and check your reasoning before you move. That is slower per puzzle and considerably more useful per puzzle.

Choosing What to Train

The most efficient approach is to let your own games pick the theme. When you review a game and find you missed something, note what kind of tactic it was, then drill that theme rather than whatever you feel like solving.

If you have no particular weakness in mind, the highest-value themes for most players below roughly 1800 are the basic ones: forks, pins, discovered attacks, and back-rank mates. These decide an enormous number of amateur games, and being fast at them is worth more than knowing exotic motifs.

Endgame themes are the other common gap. Most players train middlegame tactics almost exclusively and then lose winning endings, which is a strange allocation of effort given how often endgames decide games.

How This Fits With Everything Else

Training mode is for building patterns. Rated puzzles are for testing them under something closer to game conditions, and timed modes are for sharpening speed once the pattern is solid.

A reasonable rhythm is to drill a theme until it feels automatic, then go back to rated or timed solving and see whether it survives contact with mixed positions. If it does not, the pattern is not learned yet — go back and drill it again. That loop is the point.