The daily puzzle game is one of the strangest successful formats on the internet. One puzzle. Once a day. No levels, no progression tree, no reason to stay. By every conventional engagement metric it is designed wrong, and it works anyway.
Having built one, here is what I think is actually load-bearing.
The obvious move for any game that people enjoy is to let them play more of it. Almost every daily game refuses, and the refusal is the point.
One puzzle a day does three things at once. It makes the puzzle an event rather than a feature, so there is a reason to show up today specifically. It makes the result comparable - everyone played the same puzzle, so a score means something socially in a way that a personal-best never can. And it caps the experience before fatigue arrives, so the session ends while you still want more.
Unlimited play destroys all three. The moment there is always another puzzle, no individual puzzle matters, nobody shares a result nobody else can relate to, and players churn through the content until the fun wears out. Scarcity is not a monetisation tactic here. It is the thing that makes the format work.
The implicit deal is: give me a minute and a half, and I will give you a complete experience with a real outcome. Both halves matter.
Short enough that it never competes with anything - it fits in a queue, a lift, the first coffee. And complete: a beginning, a genuine risk of losing, and a definite end. Not a fragment of a longer grind. A daily game that starts taking six minutes has broken the contract, and players do not renegotiate. They just stop.
A streak is not a reward. Nothing is unlocked at seven days. It is a number that only goes down, and its entire power is that losing it feels disproportionately bad relative to how good earning it felt.
That is loss aversion doing what it always does, and it is the strongest retention mechanic available to a game with no progression. It is also the one to handle carefully. A streak that punishes a missed day feels like an obligation rather than a habit, and obligations get abandoned. The distinction is small and matters: the streak should be something you are protecting, not something the game is holding hostage.
The single most consequential design decision in Wordle was the spoiler-free result grid, and it is worth being precise about why it worked.
It shares a result, not the content. Coloured squares communicate how the game went - triumph, disaster, a near miss - while revealing nothing that would ruin it for the reader. So it can be posted anywhere, by anyone, without a spoiler warning, and the reader's only way to resolve the tension is to go and play.
Most games get this backwards and share a screenshot of the actual content, which spoils the thing it is advertising. A result grid is an advertisement disguised as a brag, and it costs nothing to produce because the player makes it for you.
It is worth adding a wrinkle. In ALLEGEDLY*, optional clues are free in scoring terms but print on your share receipt. The cost is not points, it is the record - everyone can see you needed help. That turns an easy decision into an actual one, and it is the only real strategic choice in the game.
Global leaderboards do not suit daily games. They reward the obsessive, they are meaningless to a casual player, and by week two the top is unreachable, which converts aspiration into resignation.
What works is horizontal comparison: how did I do against the people I actually know, and against everyone who played today. "Only 34% of readers cracked today's case" is a better motivator than a ranked list, because it is a fresh fact every day and everyone starts level each morning.
Past puzzles serve a completely different person: someone who found you today and wants more, right now, before the habit exists. A daily game with no archive converts that person into nothing.
The trap is letting the archive undermine the scarcity. If yesterday's puzzle is identical to today's in every respect, the "one a day" contract quietly dissolves. Our Morgue publishes closed cases with the answer and full story attached, so they read as archive material rather than more turns. The distinction sounds pedantic. It is the difference between a bonus and a leak.
Every mechanic above is a retention mechanic. Streaks, receipts, archives and comparison all assume a player already found you. None of them produce that first visit.
That is the honest gap in daily-game design writing, this piece included: the format is unusually good at keeping people and no better than anything else at reaching them. Build all of it correctly and you have a machine that converts arrivals into regulars, sitting quietly, waiting for arrivals.
Today's edition is here. Ninety seconds.