Most amateur games are not lost to deep strategy. They are lost to a missed fork, a hanging piece, a back-rank mate that nobody saw coming. That is good news: it means the fastest way to gain rating is also the most trainable one: tactics.
But “do more puzzles” is not a method. I went from roughly 2300 to nearly 2600 not by grinding random puzzles, but by training in a specific way. Here it is.
Tactics are two skills, not one
When people say “I’m bad at tactics,” they usually mean one of two different things:
- Recognition: seeing that a shape (a fork, a pin, a back-rank mate) is even on the board.
- Calculation: working out the exact move order and checking it survives the opponent’s best defence.
You need both, and they improve in different ways. Recognition comes from seeing the same patterns hundreds of times until they jump out at you. Calculation comes from forcing yourself to read lines to the end before you move.
The method
1. Learn the patterns by name. You cannot spot what you cannot name. Spend time with the core tactical motifs and checkmate patterns until each one has a clear picture in your head.
2. Solve at the edge of your ability. Puzzles that are too easy teach nothing; puzzles far above you just frustrate. The sweet spot is where you solve maybe 70–80%, hard enough to make you think, easy enough to keep the feedback loop tight. This is exactly why adaptive difficulty matters: it keeps every rep in that zone.
3. Calculate fully before you move. The single biggest habit change: do not play the move you “feel” is right. Pick the most forcing candidate, follow it to the end, check the opponent’s best reply, and only then commit. Guessing trains guessing.
4. Review every miss. When you get one wrong, do not just click “next.” Ask why you missed it: wrong pattern, lazy calculation, overlooked defence? The lesson is in the mistake, not the solve.
5. Repeat with spacing. You will forget a pattern you saw once. Seeing the same theme again days later is what moves it into long-term memory.
The mistakes that stall people
- Only solving puzzles you can already do. Comfortable, useless.
- Rushing. Speed has its place, but if you never calculate slowly, you never build calculation.
- Never reviewing. Solving 50 puzzles and reviewing none is worse than solving 15 and understanding each.
- Ignoring weaknesses. If you keep missing deflections, drill deflections specifically instead of hoping they come up.
A simple weekly routine
- 15 focused puzzles a day, calculated to the end.
- One theme you are weak at, drilled deliberately a few times a week.
- A quick review of anything you missed.
That is it. Not a marathon. A tight, honest loop, repeated. Consistency beats volume every single time, and the rating follows.
Want the loop done for you: adaptive difficulty, the patterns, and instant mistake review? That is exactly what Chess.PhD is built for.