When do memecoins rug? The hours of the day, from 507,266 launches

The rug clock

Anceta reads every pump.fun coin from the chain as it launches and keeps its first minutes and its outcome. That gives every coin two stamps: the hour it launched, and whether it rugged. Put 507,266 of them on a clock and the day has a shape. Coins launched in the small hours of Europe, before New York is awake, rug the most: 33.1% of the ones launched at 08:00 UTC. Coins launched in the American afternoon rug the least: 25.4% at 19:00 UTC. A coin launched at the worst hour was about a third more likely to rug than one launched at the best.

The clock above can be read in Tokyo, Paris, New York or your own time; the teal bars are each city's stock exchange hours. The rate climbs through the Asian night, peaks as Europe's morning begins, falls as New York opens, is lowest through the American afternoon, and climbs again after the New York close.

Before the clock, the floor. Rugs happen around the clock, every day of the week: not one hour of the twelve days went without them, the quietest single hour of all still had 92, and the “safest” hour of the day, 19:00 UTC, still saw one coin in four rug: 277 rugs an hour, more than the 170 an hour of the riskiest, because so many more coins launch then. What the clock moves is a coin's odds, between one in four and one in three. Nothing on it is a quiet hour.

Two clocks, running opposite ways. The hours with the most launches have the fewest rugs per coin; the hours with the fewest launches have the most. The busiest hour, 20:00 UTC with 2,437 launches an hour, sits in the safest stretch; the quietest, 07:00 UTC with 1,139, sits at the top of the risk.

Twelve days, hour by hour

A pattern over an average can hide a few bad days. This one does not. Each row below is one of the twelve days, each cell one hour of it, coloured by the rug rate of the coins launched in that hour. The dark band through the European morning is there on every row, weekends included; the pale band through the American afternoon too. On every single day, coins launched between 06:00 and 10:00 UTC rugged more often than coins launched between 16:00 and 20:00: the gap ran from 0.8% to 16.6% points.

Whose market hours

Crypto never closes, but the people trading it live in time zones, and stock exchanges are a fair clock for when each region is at its desk. Anceta sorted every launch by which exchanges were open at that moment, in each city's own local time (daylight saving ended in Europe on October 25 and in the United States on November 1, inside the window; each coin was placed by the real local clock of its day). The sessions: the Tokyo Stock Exchange, 09:00 to 15:00; Euronext in Paris, 09:00 to 17:30; the New York Stock Exchange, 09:30 to 16:00.¹²³

The two hours a day when Paris and New York are both open, 14:30 to 16:30 UTC, were the safest in the window: 24.9%. New York alone, the rest of its afternoon, came next at 25.0%. The weekday evenings and nights with no exchange open sat in the middle (28.3%), Tokyo alone a little above (29.2%). Europe alone, the morning hours before New York opens, was the riskiest session of the week at 30.5%, and the weekend, with nobody's exchange open, ran close at 30.4% against 27.7% on weekdays. Launches an hour tell the same story backwards: 2,102 an hour with New York open, 1,327 with Europe alone.

It is not the volume

The easy reading is that quiet hours have less money, so coins die. The data says something more precise, and the easy reading gets the direction wrong twice. First, the quiet hours do not have fewer rugs; they have fewer coins, and more rugs per coin. Second, the money each coin meets is not that different: a morning coin saw a median of 11.6 SOL traded in its first ten minutes, an afternoon coin 16.2 SOL, and its crowd was about the same size (32.4% of morning coins had twenty or more separate buyers by minute ten, 36.4% of afternoon coins).

The test that settles it is to compare coins of the same crowd size at different hours. If the hour only mattered through the size of the crowd, the gap would vanish once the crowd is held fixed. It does not vanish; it moves.

The whole morning penalty lives in the small coins. With three to nine buyers, a morning coin rugged 26.3% of the time against 18.7%; with ten to 29, 41.9% against 31.9%. From 30 buyers up the gap is gone, and with a hundred or more the afternoon is the worse hour. A coin that gathers a real crowd is as safe or as unsafe at any hour; a coin with a dozen buyers is far more fragile at 08:00 than at 19:00.

The creator's exit

What does change with the hour is what the creator does. Anceta reads whether the creator wallet has sold by minute ten. In the morning it had sold in 69.1% of coins that found buyers; in the afternoon, 62.4%. Hour by hour, that share follows the rug clock closely: highest from 06:00 to 10:00 UTC, lowest from 18:00 to 20:00.

And the same exit ends differently. When the creator sold, the coin rugged 39.7% of the time in the morning and 34.3% in the afternoon. When the creator kept its coins, 14.5% against 11.7%. Anceta's rug rule needs a fall of 80% from the peak after the creator or the launch block sold; in a thin hour that sale meets fewer buyers and takes the price down further, so the same act is a rug more often. The creators sell more in the morning, and their selling costs the coin more.

It is not a different kind of launcher. Wallets that launched twenty or more coins the same day made 45% of all launches, about the same share at every hour, and their coins rugged more in the morning too (52.1% at 08:00 UTC, 44.5% at 19:00), as did everyone else's (19.6% against 15.5%).

Crowd size: a U, not a line

Holding the crowd fixed turned up a finding of its own. More buyers in the first ten minutes does not mean fewer rugs. The share runs up, down and up again.

Three to nine buyers: 22.1% rugged, almost none graduated. Ten to 29: 36.0%, the classic small rug, a launch block and a few strangers. Thirty to 99: 14.4%, the calmest band of all, with 7.5% graduating. A hundred or more: 50.3% rugged, and 23.4% graduated. The coins with the biggest crowds are both the likeliest to make it to a pool and the likeliest to be dumped: a big first-minutes crowd is what a pump needs, and a pump is what a dump needs.

When the rugs are recorded

The clock above is by the hour a coin launched. By the hour the rug itself was recorded, the picture is the opposite one, and the simpler: rugs happen when launches happen. Half of all rugs are recorded within 19 minutes of launch, so the rug clock by recorded time follows the launch clock, with its peak in the American afternoon and its trough in the European morning.

The most rugs were recorded in the 16:00 hour, 2,313 over twelve days, and the fewest at 09:00, 1,181. So the two questions have two answers: the hour with the most rugs is the hour with the most launches, and the hour where a launch is most likely to rug is on the other side of the clock.

What this means when you look at a coin

The hour is not a verdict. A coin launched at 08:00 UTC is not a rug, and a coin launched at 19:00 is not safe: the rates are 33.1% and 25.4%, both far from zero and far from certain. What the hour tells you is how much a small crowd can carry. In the European morning and on weekends, a coin with a dozen buyers lives or dies on what its creator does next, and creators sell early more often then. Once a coin has thirty separate buyers, the hour stops mattering. So the first ten minutes matter more at the hours when fewer people are watching: who bought, how many, and whether the creator is still holding, which Anceta reads for every launch on its coin pages and in RugUAV. Pattern analysis, not financial advice.

How we counted

Sources

1. 1.Japan Exchange Group, trading hours of the Tokyo Stock Exchange. 09:00 to 15:30 JST since November 5, 2024; this report counted 09:00 to 15:00, half an hour short of the close
2. 2.Euronext, trading hours and holidays. 09:00 to 17:30 CET in Paris
3. 3.New York Stock Exchange, hours and calendars. 09:30 to 16:00 Eastern

Related research

1. 1.Solana rug pull report: October 2026. the same definitions over eight days
2. 2.How long do memecoins last?. 57,168 coins followed hour by hour
3. 3.How to check if a memecoin is a rug pull. which first-minutes checks held up