Research · test #8
Do new crypto listings dump after day one? No edge
The idea: a new coin spikes on the first day, then bleeds - so short it once the hype is over. (The opposite version says new listings keep rising, so buy.) We wrote the rules down first and checked every coin that started trading on Binance futures from January 2024 - 348 of them. Most listings do fall, but shorting them has no edge: the typical trade wins, and a few coins that rose several times over take the profit back.
In short: 348 listings, entry 24 hours after the first candle, exit seven days later. Short: 64% winners, median +8.9%, average +0.61% after costs and funding (t 0.24); on the second half of the data -1.27%. Long: 36% winners, average -0.91%. The worst three shorts lost 219% to 290% of the position.
The rules (fixed before the run)
- Listings: coins whose first hourly candle in our Binance USDT-M perpetual data falls between 5 January 2024 and 1 September 2026 - 348 coins. Older coins (170) are the control.
- Entry at the close of the hour 24 hours after the first candle - by then the first-day high is known. Exit 7 days later (main), 24 hours and 72 hours (secondary). One trade per coin and side.
- Result = price move in the trade's direction + every funding payment during the hold - 0.15% for fees and slippage.
- Control: for each listing, a random old coin at the same entry date and the same hold, 10,000 draws. This removes what the whole market did in those days.
- Edge requires: average above zero, t above 2 and p below 0.05 against the control - on the first part (to 30 June 2025) and the second part (from 1 July 2025) separately; at least 30 listings in each.
Result: no edge on either side
| 7-day hold | Trades | Average | Median | Winners | t | p vs control |
|---|---|---|---|---|---|---|
| Short, 1/24-6/25 | 184 | +2.28% | +9.41% | 65% | 0.70 | 0.0008 |
| Short, 7/25-10/26 | 164 | -1.27% | +8.00% | 62% | -0.33 | 0.99 |
| Short, all | 348 | +0.61% | +8.91% | 64% | 0.24 | 0.50 |
| Long, 1/24-6/25 | 184 | -2.58% | -9.53% | 35% | -0.79 | 1.00 |
| Long, 7/25-10/26 | 164 | +0.97% | -7.58% | 38% | 0.25 | 0.005 |
| Long, all | 348 | -0.91% | -8.77% | 36% | -0.37 | 0.51 |
The short looks good on the first part (p 0.0008 against the control), but the control there was flat and the short's own t is 0.70, far from 2; on the second part it lost money. The long's p of 0.005 on the second part only says it lost less than old coins did - its own average is indistinguishable from zero.
Most listings fall - the average hides it
Seven days after entry the typical listing was 10.2% lower, and only 35% were higher. Yet the average fell just 2.0%: a few coins rose several times over, and a short has no ceiling on its loss. With costs and funding the short's average is +0.61% - positive, but with t of 0.24 it is what chance produces. The worst three shorts lost 290%, 289% and 219% of the position, with no leverage involved.
The sign flips from period to period
A strategy with an edge would not swing between +8.2% and -5.9% depending on the half-year. Only 2025 H1 reaches t of 2.01, and it is one of five periods; the next half-year went the other way. This is descriptive and not part of the verdict (the five listings from late 2026 are left out of the chart).
Other entries and exits
- Entering 1 hour or 6 hours after the first candle, or exiting after 24 or 72 hours, changed nothing: the short's average ranged from -3.3% to +2.3% across the parts of the data and t never exceeded 1.1.
- Funding of +0.10% per 8 hours or more was paid in the first day on 104 of the 348 listings (30%) - new coins often start with a crowded long side. A short collects it, but it did not offset the squeezes.
What this test does not say
- Coins closed before 21 October 2026 are missing. Listings that collapsed and were delisted are not in the data, which makes shorts look worse than they would on the full list. We cannot say by how much.
- Entries are at the hourly close. New coins have wide spreads and thin books, so real slippage can exceed the 0.15% we deducted, and a short may not be possible at all if the exchange limits positions.
- Listing dates are the first candle in our data, not the announcement. Trading ahead of or right at the announcement is a different strategy.
- Listings cluster in time and trades are not independent; the matched control removes the market's move but not everything.
Data and method
Rules were committed before the calculation, and the code before the run; the code and every trade are kept with the project. Data: Binance public futures funding history and hourly candles. The same approach as in our other tests: a control from the same dates, two time periods checked separately, costs and funding included. Related: do pumps always dump (the same pattern of frequent wins and rare large losses), extreme funding, live funding rates.
FAQ
Do new crypto listings go down after the first day?
Most do. Of 348 coins that started trading on Binance futures from January 2024, 65% were lower seven days after entry and the median change was -10.2%. The average was only -2.0%, because a few listings rose several times over.
Is shorting new listings profitable?
Not reliably. The short won 64% of the time with a median of +8.9%, but the average after costs and funding was +0.61% (t 0.24), and -1.27% on the second half of the data. The worst three shorts lost 219% to 290% of the position.
Should you buy new listings?
No edge either: the long averaged -0.91% over seven days after costs and funding, and only 36% of trades were profitable.
Are delisted coins included?
No. Only coins still trading on 21 October 2026 are in the data, so listings that collapsed and were closed are missing. That makes shorts look worse than they would with the full list, by an amount we cannot measure.