How we measure.
Every trader re-read and re-scored on what they actually kept. Here is how the ranking is built, and the proof it works.
Does the ranking work?
Each Monday we take the top of the leaderboard as it stood that day, built only from the 90 days before, and look at the positions its traders open during the next seven days. We compare them with the other ranked traders the same week, with traders drawn at random, and on the same tokens.
| Top | Weeks ahead | Median trade, top | Median trade, others | Winners, top / others | Same tokens, top better |
|---|---|---|---|---|---|
| Top 20 | 19 of 19 | +1% | -11% | 52% / 32% | 68% |
| Top 100 | 19 of 19 | +0% | -11% | 49% / 32% | 69% |
| Top 250 | 19 of 19 | -1% | -12% | 47% / 31% | 71% |
| Top 1,000 | 19 of 19 | -4% | -15% | 39% / 29% | 70% |
May 18 to Sep 28, 19 weeks, recomputed every few hours. Each position at the trader's own entry and exit, $100 each; one still open counts at the market price now.
- The ranking picks traders who do better the next week, every week so far. It is not luck: they sit above 95% of random draws.
- It picks consistency, not moonshots: more winning trades and smaller losses, not more x2s, except the top 20.
Who gets ranked
We follow about 30,000 fomo traders and their full history. A trader is ranked once they have at least 30 closed positions, four active weeks, a complete history on our side, and at least $2,500 at risk at the same time at some point. Only traders who bought something in the last 30 days are ranked. Accounts younger than eight weeks are hidden by default, because a short run looks better than it is.
Peak at risk
Nobody's wallet balance is visible, so we rebuild it: the smallest amount that would have paid for every position a trader held at the same time, with profits reinvested as they came. ROI is total profit divided by that peak.
Realized, unrealized, total
Realized is profit from positions already sold. Unrealized is what the open positions are worth at the market price today, tokens times price, minus what they cost; a token whose pool holds under $1,000 is dead and counts zero. Total is the two together. The score and the ROI use a stricter value, what each pool could absorb if the bag were sold, so a bag nobody can sell never lifts a trader up the ranking. Ranking on realized profit alone flatters traders who sell their winners and keep their losers, which is why we count both.
The score
Each trader gets a percentile on eight measures, averaged with equal weight: growth per trade, win rate, profit factor, ROI, profit without the single best trade, share of positive weeks, downside volatility, and worst drawdown. Small samples are pulled toward the average so that eight lucky trades cannot top the list.
Style labels
Every trader with 10 or more closed positions in the last 90 days gets two labels: how long they usually hold (Scalper under 20 minutes, Swing up to 3 hours, Holder beyond) and how much they usually put in (Small under $150, Mid up to $600, Big beyond). A third tag names their chain when 70% or more of their positions are on it, and the best scores carry Elite (top 1%), Top 5% or Top 10%.
A squad built on a rule can keep only one style or one size, inside its top N: see your rules.
We measured nine habits and kept only those that stay the same from one month to the next. Position size (correlation 0.75 between the two halves of the window) and holding time (0.68) hold. How erratic the sizes are (0.27), which chain (0.28) and how young the tokens are (0.16) do not, so they are not labels. Two thirds of traders keep the same horizon and the same size from one half to the other, where chance would give one third.
Sell waves
When three ranked traders sell the same token within 30 minutes, the price has historically fallen further than usual: 56% over the next 24 hours, against 16% for the same tokens at random moments, across 3,720 tokens.