Competition that pays what it said it would.
A leaderboard is the simplest way to reward playing well rather than playing constantly. Any game on the network can run one, in any token, paying as many places as it likes — and what each place is worth is fixed before a single score is submitted.
Who is actually ahead.
Two tables, kept apart on purpose. A game board ranks scores a studio signed and pays a prize. The Union table is community standing — playing, finishing quest chains, taking part — and carries no prize and no token.
The Union
One table across every game. Playing counts towards it the day you play — link your game account with /linkgame in Discord.
Loading…
A player is named only if they have made their profile public. Everyone else keeps their place and their score — the ranking is complete either way.
Game boards
Loading…
Every entry on a game board was signed by the studio that ran it, and the ranking is published so anyone can recompute it. How that works is below.
Pay for outcomes, not for hours.
A pool handed out per hour played settles in one place: automated accounts arrive until an extra one earns roughly what it costs to run, and a real player’s share collapses. A ranking does not work that way. Finishing first is not something you can do more of by leaving a machine on overnight.
Any token
A board’s prize pot is held in whichever ERC-20 its organiser chooses. A studio can fund a season in its own game token, in ARCU, or in a stablecoin, and a sponsor can top up a board they do not run.
Any number of places
Winner-take-all, a podium, top ten, top fifty. The payout curve is a list of shares that must add up to the whole pot, and it is recorded when the board is created.
Fixed before play
The curve cannot be edited afterwards. Whoever settles the board decides who came where, and nothing else — what a place is worth was already settled.
What each place is worth.
Shares are recorded in basis points and must sum to exactly 10,000, so a board cannot be created that pays out more or less than it holds. These are the ready-made shapes; any other set of weights works too.
| Shape | Places | Split | Suits |
|---|---|---|---|
| Winner takes all | 1 | 100% | A single headline prize |
| Podium | 3 | 50 / 30 / 20% | Short events and one-off cups |
| Equal split | 10 | 10% each | Rewarding a bracket rather than a winner |
| Decaying | Any | 43 / 26 / 16 / 9 / 6% | A long season with a real tail |
A prize pot is frozen at the moment results are published, so money arriving later cannot change what an earlier winner was owed. Rounding remainders stay in the pot and return to whoever funded it.
The game signs, never the player.
Why it has to work this way
Nobody outside a game can tell a real score from an invented one. So every entry is signed by the game’s own server, and an entry signed by anyone else is refused. The network checks that the studio put its name to a score; it does not pretend to judge the score itself.
The tools refuse to sign from inside a game client rather than doing it quietly, because a client that can sign a score can sign any score at all.
The honest limit
A game with no authoritative server cannot secure a prize board. There is no way around that, and we would rather say so than sell a guarantee that does not exist.
Such a game can still run a board with no prize attached. A leaderboard people play for pride is worth more than one that quietly pays whoever automated it first.
A ranking nobody can recompute is not a result.
When a board settles, every entry behind it is published — scores, timestamps and signatures — alongside the final order. Anyone can rebuild the result from that file and compare.
Checkable by anyone
- That the published ranking follows from the published entries
- That every entry really was signed by the game that claims to have sent it
- What each place is worth, from the curve recorded when the board opened
- That the prize pot matches what was paid out
Stated, not proved
- Whether a signed score reflects what actually happened in the game — only the studio can know that
- Whether a high-placing player was a person or a very good script
- Whichever party operates the board, which is a disclosed centralisation rather than a solved problem
A published result stays inert for a dispute window before anything can be withdrawn, so a wrong one can be found and cancelled while the money is still in place. The ability to cancel sits with a different party from the one that publishes, because a guardian who is also the publisher guards nothing.
Two lines in your game.
Leaderboards are part of GameKit, the same package that pulls a build through the node network. Reading a board needs nothing at all; recording a score needs the key that lives on your server.
What you write
One call to show standings wherever you want them, and one call from your server when a run ends. No wallet handling, no contract calls, no Merkle anything.
Players who never touch a wallet can still appear on a board. A prize only needs an address at the point somebody claims one.
What you do not write
The ranking, the tie-breaking rules, the settlement, the proofs, the claim flow and the audit trail. Those are the parts that are easy to get subtly wrong and tedious to get right, which is the whole reason for offering them.
The same machinery records anything else your game vouches for, which is what a player’s passport is built from.
Seasons are the shape player rewards take.
The reasoning behind rewarding placement rather than playtime, and what the games allocation is proposed to fund, is set out in full in the whitepaper.