Most daily games hide one answer. That makes the content cheap to build and cheap to check: type a word, compare it to a string, done. We went the other direction, and the work moved from the game loop into the answer bank.
The format
One prompt, two minutes, about 27 hidden answers to name. You get points for each answer you find, and the points depend on how rare the answer is, not on how many you found. Seven hints are free. The next prompt appears at 00:00 UTC.
That format is easy to describe and annoying to build, because every part of it depends on a list that has to be right before the round opens.
An answer bank is a judgement call, not a word list
"Name things you would find in a kitchen drawer" sounds like a five-minute job. It is not. You have to decide, for a single prompt, what counts as an answer, what counts as a duplicate, and what counts as too obscure to accept.
Take batteries. Is a "dead battery" a separate answer from "battery"? No. Is "AA battery" separate? It is the same object, so also no. Is "battery charger" separate? Yes, it is a different object. Multiply that conversation by 27 entries and 120 prompts in the bank and you have the actual content pipeline.
Every prompt sits in the bank with its answer list, its rarity weights and its hint ladder before it can go live. An answer bank that is 90 percent right is worse than useless: the player who types the missing legitimate answer gets a miss, and a miss feels like the game is broken rather than like the player is wrong.
Matching is where players quit
The matching layer has to forgive the things people cannot control and refuse the things they can.
Forgiven: capitalisation, extra spaces, punctuation, accents, leading articles ("the", "a", "an"), and simple plurals in both directions. "Batteries" and "battery" are the same answer.
Not forgiven: a different word. "Charger" is not "battery charger". If the game accepted loose semantic matches, scores would depend on how generous the matcher happened to be that day, and the whole rarity system would inherit that noise.
Duplicates never score twice, which sounds obvious until you watch someone type the same answer three ways in a two-minute round.
Why the tier is the rarest answer, not the total
Every answer carries a weight from 1 to 5:
- 1 — Common. The first things almost everyone types.
- 2 — Uncommon. A little more thought than the obvious list.
- 3 — Rare. The list is getting interesting.
- 4 — Epic. Most players miss these.
- 5 — Legendary. The deep cuts.
Your score is the sum of the weights. Your tier is the weight of the rarest answer you found, not the total. Finding one Legendary, one Rare and four Commons is 12 points and a Legendary tier.
That split matters in a game like this. If the tier were the total, the winning strategy would be to type as many obvious answers as possible as fast as possible, which is exactly the behaviour the game is not trying to reward. Making the tier follow the single rarest find means the two-minute clock is spent deciding whether to keep listing groceries or to reach for something odd.
Seven hints, no cost
Each round has seven hints: first letters, word lengths, category clues, in that order. They cost nothing.
Free hints are a deliberate choice. Charging for them would turn a two-minute game into a resource-management game, and it would punish the player who is stuck rather than the player who is careless. Hints narrow the search space without handing over the answer, which is the only job a hint has.
After the clock
When the round ends, the full board is visible: every answer, its weight and its tier. That reveal is where most of the learning happens — you see the Legendary you missed and realise it was obvious in hindsight.
Past boards stay available for 90 days, so a missed day is not a lost streak. There is no signup, no download, and no energy system, which means the only thing keeping you in the loop is whether the prompt is good.
What we would do differently
If we rebuilt the answer bank from scratch, we would write the hint ladder first. Writing hints before the answer list forces you to describe the categories out loud, and half of the awkward entries reveal themselves at that stage, before the prompt ever reaches a player.
The daily board is at ZomZam if you want to see the finished version of this pipeline in two minutes.