How to Run a Successful Trading Competition

A perp DEX puts up a prize pool to bring volume onto its book. By day three the top three places belong to wallets nobody recognizes, trading the same pair against each other around the clock. The volume is real. It counts. It is worth nothing.

What a competition buys is decided before it opens, by the number the leaderboard ranks.

‍

Which metric to reward

A competition does not reward trading. It rewards one number, and traders optimize for that number.

Volume is the simplest number to rank on and the cheapest one to fake. Two wallets run by the same person trade against each other all day and post any figure you want.

PnL is what most competitions pay on, and wash trading cannot inflate it. Its weak spot is one trader with two accounts, long on one and short on the other: one of them always finishes in profit.

That is not a reason to drop PnL. It is the reason to run filters that detect wallets controlled by the same person.

ROI ranks on the percentage gained, not the dollars, so a small account can compete with a large one. Its weak spot is leverage: a small account on high leverage can post an unbeatable number in a single trade.

A minimum volume or number of trades closes that gap. To rank, a trader has to keep trading, not place one bet.

Rank on PnL or ROI, set a minimum activity threshold, and run those filters before anyone is paid.

‍

How to structure the leaderboard

Once the metric is set, the shape of the board decides how many people compete.

1. The window. One or two weeks creates urgency and caps your exposure to any single farming strategy. Long windows favor whoever automates first.

2. The cutoff. Publish the exact timestamp that closes the window, in UTC, before the competition opens. A cutoff that moves is the fastest way to lose the argument with the trader who finished fourth.

3. The tiers. Group the ranks instead of paying each position its own number: 1 to 3, 4 to 10, 11 to 50, 51 to 200. The prize should fall more slowly than the rank does, so a trader who cannot reach the podium still has a next tier worth trading for.

4. The spread. Winner takes most looks generous and empties the board: once one wallet is far ahead, everyone below stops trading and the volume leaves with them. Pay on share of the metric instead, with a proportional pool.

‍

Settle after the window closes

Competitions that pay in real time give a farmer a fast loop and no window in which anyone can look at what they did. Settling on a published schedule, after the window closes, takes the loop away: the farmer has to commit capital and wait, knowing the reward may never become claimable.

Waiting also buys the room for everything in the next section. A flagged wallet that has already been paid is a clawback conversation. A flagged wallet that has not been paid yet is a decision.

‍

Which filters run before anyone gets paid

The filters are what separate the leaderboard from the payout list. Four are worth running on any competition.

Cluster detection groups wallets that look like one person, reading more than 30 onchain signals from funding sources to transaction timing. It is what catches the trader with a long on one account and a short on another.

Self referral detection catches a referrer and an end user who share the same browser session.

Bot detection reads frontend and onchain behavior together to identify automated software.

Caps limit what one account earns in a time window, and what an affiliate earns from referrals relative to their own trading.

Whatever they catch goes to a review queue where your team approves or rejects it, and the reason stays attached to the payout.

‍

What a competition cannot do, and what to run underneath

A competition is time-boxed. It ends, and the volume it bought tends to leave with it.

Treat it as the peak of a program, not the program. What runs underneath is what keeps the traders it attracted:

• A referral program that pays on fees generated.
• A rate that differs by segment, so the wallet that stayed earns more per dollar traded than the one that arrived for the prize.
• A standing reward on open positions held over time, at a lower budget, all year.

‍

What Fuul does

Fuul tracks the metric the competition ranks on, builds the leaderboard from it, and pays by position or by share of a pool.

None of the filters in this post is something your team builds. They run before a payout exists, and every decision, including the rejected ones, keeps its record next to the reward. Rewards are paid from your own contract, in tokens or points.

The competition gets you the attention. The program underneath decides how much of it you keep.

If you are planning a competition, book a demo at fuul.xyz.