Hard chess puzzles

Calculating deeper: what changes after 1400

Hard chess puzzles for 1400+ players are not just "harder mates", the test shifts from seeing the tactic to finishing it. Below 1400, most positions resolve in two or three moves and the side to mate is usually obvious. From 1400 up, the median solution is four to six half-moves, the first move often looks quiet, and the win frequently runs through a forcing line where every reply has to be checked. My own jump from 1500 to 1900 on Lichess came almost entirely from forcing myself to count one move further before I clicked. If you have a tactical idea that works after the opponent's only good reply, you do not have a hard puzzle yet, you have half of one. Treat the slider band 1400-1800 as a calculation gym, not a pattern-recognition warm-up. (For pure motif drills, the easy puzzles set is the right room; this page is the harder one next door.)

Concrete cues that a position is in "hard" territory: more than one capture on the first move that looks plausible, a king that is loose but not yet checkable, a piece that is pinned and overworked, or a back-rank theme covered by a single defender. When you see two of those at once, slow down and count attackers vs. defenders on the critical square before you reach for the mouse.

Mate-in-N and only-moves: how to read the line, not guess it

The biggest jump in difficulty on this page comes from longer mating nets. A mate-in-3 usually has one forcing move and a clean follow-up; a mate-in-4 often hides a quiet move or a king-step inside the sequence; a mate-in-5 almost always has at least one "only move" reply you must spot for the other side before yours is justified. My rule of thumb: name the final mating pattern out loud (back-rank, smothered, Anastasia, h-file) before you start hunting the first move. If you cannot picture the last square the king sits on, you are guessing.

Only-moves on this page work the same way. The position is winning only if you find the exact reply, and a strong-looking second-best move often loses a tempo or the whole game. Read the candidate moves in the order forcing → semi-forcing → quiet (check, capture, threat, then everything else). On hard puzzles the answer is the quiet move about a third of the time, so do not skip past it. If you keep clicking the loudest check and losing, browse the quiet-move puzzles as a side drill for a week and the pattern will start showing up here. For the kingside-attack flavor specifically, the kingside-attack set overlaps heavily with what hits 1700+ rated boards.

Calculating hard puzzles honestly: the "two-minute" rule

Hard puzzles tempt you to guess the moment a line gets fuzzy. Do not. The fastest improvement I saw between 1600 and 1900 came from giving every position a full two minutes of human thought before I committed to a move. Two minutes is long enough to spot a defender you missed, a check that does not work, or a stalemate trick that ruins your mate-in-4. Less than that and you are guessing, not calculating.

Use the hints as a ladder, not a crutch: the first names the motif, the next points to the piece that moves. Read the candidate moves in order, forcing then semi-forcing then quiet, and only reach for a hint once you have genuinely sat with the position. When you miss, the trainer tells you "Not the move." and advances to the next puzzle, so there is no engine line to lean on, that is the point. Track your wrong-first-move rate instead. If it sits above 40% at your current slider band, drop the rating 100 points and rebuild from there; that is honest practice for stronger players, not punishment.

If you want a softer ramp before climbing back up, the middlegame puzzles page mixes king-side attacks, isolated-pawn positions and time-pressure recognition, same difficulty range, different cognitive load (positional pattern-recall instead of long forcing nets).

FAQs