Market structure

What happens to a launch

Measured on 2,410 pump.fun launches, chosen without regard to how interesting they looked. This is the shape of the market you are betting against.

48% of launches never receive a single trade. Of those that do, half are finished within about four minutes. One hour later, 3.5% are still trading — about 1 in 29.

0%10%20%30%40%50%24.2%8.3%3.5%01m2m5m10m20m30m1hminutes since launch · log scale
A deterministic 1-in-4 sample of every pump.fun launch, each watched from birth so its silence was measured rather than assumed. n = 2,410. The curve ends at one hour because that is where the observation ended — not where the tokens did.

The figures

Share of all launches — including the ones that never traded — whose last observed trade falls at or after each age.

Share of pump.fun launches still trading, by minutes since launch
AgeStill tradingRoughly
traded at all52.45%1 in 2
1 min46.93%1 in 2
2 min41.58%1 in 2
5 min24.19%1 in 4
10 min16.60%1 in 6
15 min12.95%1 in 8
20 min11.00%1 in 9
30 min8.30%1 in 12
45 min5.98%1 in 17
60 min3.49%1 in 29

How it was measured

The sample
A deterministic 1-in-4 slice of every launch observed in the window — selected by an internal sequence number that carries no information about a token's symbol, creator, or deployer buy. Nothing was chosen for looking promising.
The observation
Each token was held under a live trade subscription from birth, so a token going quiet was measured, not inferred from us having stopped watching. That distinction is the whole reason these numbers can be trusted.
The window
13 Aug, 18:35 → 14 Aug, 05:01 UTC · n = 2,410

What this does not tell you

The estimate this replaced said 26.8% of launches are still trading at an hour. It came from a population that only included tokens already showing activity — selected for being alive. Finding that out is what this measurement was for. How we work