NYT Pips Hard

Hard, measured against the other two

The Times publishes one easy, one medium and one hard board every day, so there are exactly as many hard boards as easy ones and the comparison is fair. Averages over all 397 published days, with the badge mix as a share of all regions on that tier:

Two of those rows contradict what most people assume about a hard puzzle. The first is that hard boards are less loose, not more: 29% of their regions offer no exact number to compute from, against 40% on easy and 41% on medium. Difficulty here does not come from vagueness. It comes from size — 26.3 squares against 9.9 — and from the arithmetic being everywhere: 55% of a hard board's regions carry an exact sum. A hard board tells you almost everything and then makes you hold all of it at once.

The second is the doubles. A hard tray carries 3.12 doubles on average, more than twice the easy tray's 1.32. A double is the only tile that can sit in an all-equal region without a partner of the same value, and it is the tile most often forced by a small exact sum — so on a hard board the doubles are not decoration, they are the skeleton. On 3 of the 398 boards measured, all seven doubles in the set are in the tray.

Who builds the hard boards

The puzzle data carries two separate credits, and they are not the same person. The editor — Ian Livengood — is credited on every one of the 397 days. The boards themselves carry a constructor credit each, and on the hard tier that credit is overwhelmingly one name:

Rodolfo Kurchan is credited on 92% of every hard board published. That single fact explains why the hard tier feels so consistent from day to day, why the method below transfers so well between boards, and why the easy tier — which is mostly Ian Livengood's — plays like a different game rather than a smaller version of this one. When you learn to read a NYT Pips hard board, you are mostly learning to read one constructor's habits.

The extremes

Averages hide the shape of a tier. These are the boards at its edges — every one a real published day.

The ceiling: 32 squares

No hard board has exceeded 32 squares — sixteen dominoes — and the tier hits that ceiling constantly: 149 of the 398 boards measured (37%) are 30 squares or more. The tray runs from 6 to 16 dominoes. Among the boards at the ceiling, the ones with the most regions are the densest puzzles the Times has set:

Every square its own region

The most regions ever placed on one hard board is 32, on August 7, 2026 — a board where every single square is a region of its own, so the puzzle is not about regions at all but about which half of which domino can land where. It is the tier stripped to pure tiling.

The 63

The highest exact sum ever demanded of one region is 63, on September 15, 2025. For scale, no other hard board has asked for more than 30 in one region, and the commonest sum target across all hard boards is 4. A region that needs 63 has to be enormous — the largest region on any hard board spans 16 squares — and it inverts the usual logic: instead of a small sum forcing low tiles, a huge one forces the high dominoes to go there and nowhere else.

Nothing loose at all

25 hard boards have no loose region whatsoever — every region is an exact sum or an all-equal — the most recent being September 15, 2026. These are, counter-intuitively, often the slowest solves in the tier: with nothing to eliminate by inequality, the only way in is arithmetic, and on a 32-square board that is a lot of arithmetic to hold.

And the one that is all loose

At the other end, January 27, 2026 has 100% loose regions (15 of 15) — not one exact sum on the board. It plays like a completely different puzzle, closer to a pure placement problem, and it is the clearest evidence that “hard” is a size and a style, not a single mechanism.

The smallest hard board

The smallest hard board ever published had 12 squares, on October 6, 2025 — a hard board physically smaller than the average easy one (9.9 squares). Only 27 hard boards (7%) come in under 20. It is a useful corrective: what the tier is selling is density, and a small dense board still earns its label.

One more number for the file: 86 hard boards (22%) contain a ≠ region. Across the whole game that badge is nearly absent, which is why its rule gets misread so often — but on the hard tier you will meet it roughly one day in five.

Recent hard boards, hints first

The last thirty published hard boards, profiled. Each date opens that day's answers page at the hard board — hints before the solution, so you can take as little as you need.

Every earlier day is in the full NYT Pips answers archive, by month, back to the first board. To play the official puzzle, go to The New York Times; nothing on this site reproduces its boards to play.

If what you want is a hard board you can actually play right now, without a subscription and without waiting for tomorrow, this site sets its own — same rule vocabulary, verified solvable, and ranked by what makes them brutal. The daily board comes in three difficulties too.