Frequently asked questions

Because it measures something different, not because either figure is wrong.

PredCheck counts only short-term crypto markets — the 5m, 15m, 1h and 4h Up or Down markets on supported assets. Anything a trader did in sports, politics, equities or longer-dated crypto markets is excluded entirely, so a trader active across the whole venue will show a smaller figure here.

The bigger difference is that Trade PnL is net of costs. A headline profit figure is usually gross — it counts what a trader won and ignores what the trading cost them. PredCheck rebuilds the result from raw fills and subtracts the fee each market’s published schedule implies for every one of them. On a 5-minute market a round can pay up to $1.75 per 100 shares, which at even odds is 3.5% of its own volume. For someone trading a hundred rounds a day that is not a rounding difference; it is frequently the whole result.

What Trade PnL does not do is add the venue’s rebates and rewards back in, because those are paid to an account rather than to a market. That sum is published as a second figure, Overall PnL, which is what the account took home and the one to compare against the venue’s own profile page.

Because they missed at least one of the bars a score requires for that window: enough completed rounds, at least ten rounds on each day they trade, an average stake of $20 or more, a minimum net profit, an average ROI of 5% per round, and a profit factor of 1.2. The profile names the bar that was missed.

A number computed from thirty cheap rounds over two days would look authoritative while being mostly luck, so PredCheck declines to publish one. The score is computed separately for the rolling day, week and month, so a trader can carry a monthly score and still show “Not enough data” on the day.

In Overall PnL, yes. In Trade PnL, deliberately not — that is what separates the two figures.

The venue pays these to the account as a whole, across every market the trader touches — politics, sports, long-dated, everything — and no payout says which market earned it. PredCheck only measures short-term crypto. One wallet had 719 of its last 1,000 trades outside that scope, so its account-wide rebates say very little about how it did here.

Apportioning the payouts by volume would replace a wrong number with a guess, so neither figure tries. Trade PnL stays what it claims to be: profit from the short-term crypto markets, gross less fees. Overall PnL states the sum and says plainly that its halves cover different markets. Reading them together is the point — a trader can be under water on the trading and comfortably ahead on the account, and that gap is the most useful thing on the page.

The payouts also arrive in batches rather than round by round, so on a short window Overall PnL can look worse than the account really is: the trading loss is already counted and the rebate that covers it has not been paid yet.

A trader whose payout feed has not been read yet shows “Unavailable” rather than zero, and no Overall PnL is stated at all. Unread and genuinely nothing are different claims, and PredCheck will not make the second one until it has checked. Referral income is paid by the venue but PredCheck does not read that feed, so it is always Unavailable.

The fee is computed from the fee schedule each market publishes, which is exact arithmetic: rate × size × price × (1 − price), peaking at $1.75 per 100 shares at even odds and falling away towards the extremes. What cannot be determined from public data is whether a given fill was the aggressing side, and on these markets only the aggressor pays — the resting side pays nothing.

PredCheck starts by charging every fill as if it were the aggressor, then uses the maker rebates the venue paid the wallet to work out how much of that flow was maker flow, and hands that share of the fee back. For a wallet whose rebates come mostly from markets outside the ones measured here the split cannot be read off, and the full upper bound is kept. It is an estimate either way, and it is labelled that way wherever it appears.

Everything one trader did inside one short-term market. A trader who buys, sells, takes the opposite side and re-enters inside a single BTC 5m market has traded one round, not five trades.

That is why win rate and average ROI here are comparable between someone placing two orders a day and someone placing two hundred.

ROI is measured against the most capital the trader ever had committed at one time during a round, not against total turnover.

Someone who recycles the same $50 three times has $150 of turnover but only ever risked $50. Dividing by turnover would understate an active trader’s efficiency threefold, so PredCheck does not.

Short-term crypto Up or Down markets on Polymarket, in four timeframes: 5 minutes, 15 minutes, 1 hour and 4 hours. Supported assets currently include BTC, ETH, SOL, XRP, DOGE, BNB, HYPE and ZEC.

Sports, politics, elections, equities, FX, commodities and longer-dated crypto markets are out of scope and never enter the dataset.

PredCheck builds its own history from raw fills rather than reading a stored one, so “All time” means all of the history in this dataset — not a trader’s lifetime on the venue. The window reaches back about a month and is being extended.

That is why a wallet active for a year shows a smaller all-time figure here than it would elsewhere, and why the longer periods can agree with the shorter ones on a trader whose record does not yet reach that far back.

Continuously. New markets are picked up within a couple of minutes, settlements within about ninety seconds of resolving, and fills are imported once a market has settled — which is the point at which a round can be finalised.

Metrics, Copy Scores and classifications are then recomputed on a rolling basis, so a trader’s profile reflects their latest completed rounds rather than a nightly snapshot.

Most often because they have not traded a supported short-term crypto market. A wallet that only trades sports or long-dated markets will not appear, by design.

PredCheck also tracks a curated active universe rather than every wallet on the venue. Traders are discovered from the public crypto leaderboards and from the participant list of every settled short-term market, so anyone genuinely active in these markets is picked up on their next round.

A trader is treated as inactive once they have gone seven days without taking part in a supported short-term crypto market, and inactive traders drop out of the leaderboards and discovery surfaces.

Nothing is deleted. Their profile page stays reachable by direct link with its full history, and they return to the leaderboard automatically on their next qualifying round.

By results alone. Searches, clicks and profile views count for nothing — the ranking is computed entirely from settled rounds in the last 24 hours, so no amount of traffic can put a trader on it.

A trader must first clear every minimum: Trade PnL, ROI, win rate, Copy Score and completed rounds. All of them, not one — a board that admitted a trader for a single strong number would mostly promote lucky days. Those who qualify are then scored on profit, Copy Score, win rate, median ROI per round and consistency.

Profit is scored on a compressed scale, so each further doubling counts for less. A wallet trading with a hundred times the capital cannot take the top slot on size alone; it has to trade well.

The board is often empty, and that is the intended behaviour rather than a fault. On a day when nobody clears the bar, the section simply does not appear.

The homepage shows the strongest few. The full board, everyone who cleared the bar that day, is at predcheck.xyz/hot-traders.

Email [email protected] with the wallet address and the period you were looking at.

PredCheck retains the raw fills behind every metric, so any figure can be recomputed from source and checked line by line.

No. PredCheck is a read-only analytics site.

It does not execute or route orders, does not connect to a wallet, and never asks for a private key, seed phrase or signature. It holds no funds and cannot act on a trading account.

No. They are analytics classifications describing patterns in past results.

“Copy-Friendly” does not mean you should copy a trader, and “Fade Candidate” does not mean you should trade against one. Neither describes a guaranteed strategy, and past performance does not indicate future results.

No. There is no sign-up and no wallet connection.

Your theme choice and your watchlist live in your own browser’s local storage, on your device only, and are never sent to us — which is also why a watchlist does not follow you to another browser. PredCheck sets no cookies of its own; Google Analytics sets its own so that a repeat visit is not counted as a new one. The privacy policy has the detail.

No. PredCheck is an independent project and is not affiliated with, endorsed by, or sponsored by Polymarket or any other venue.

It reads public data and calculates its own figures. It makes no claim about numbers published anywhere else.

The trading data shown here is public on-chain and public venue data, which PredCheck reads but does not own and cannot amend at source.

If you have a concern about how a profile is presented, email [email protected].