Should you buy a pump.fun CTO at the alert, wait for confirmation, or bid below it?
Buying at the moment of the alert is the baseline: across 44 takeovers in the last 45 days, 41% reached +50%, the median peak was 1.42x and 11% were down 80% or more at the end of the window. Waiting 60 minutes and buying only if the coin had held reached +50% 43% (n 35); the record does not separate that from buying at the alert. Resting a bid under the alert price filled on 14% to 34% of signals; the bid table gives each depth with its n. Updated 2026-10-27.
Cite as: Belphor Research (2026). Should you buy a pump.fun CTO at the alert, wait for confirmation, or bid below it? Belphor Data, https://www.belphor.com/data/cto-entry-timing. Updated 2026-10-27. Inputs eb5cf72cf54d, formula 1.5, covered 44, missing 0. CC BY 4.0.
Buy now versus wait (entry price is the price at that moment)
| Entry | Trades | Share of signals | Reached +50% | Reached +100% | Median peak (x) | Average peak | With a later reading | Down 80% or more |
|---|---|---|---|---|---|---|---|---|
| buy at the alert | 44 | 100% | 41% | 32% | 1.42x | +515.1% | 44 | 11% |
| buy 30 minutes after the alert | 44 | 100% | 43% | 30% | 1.37x | +298.1% | 44 | 11% |
| buy 60 minutes after the alert | 44 | 100% | 34% | 23% | 1.22x | +157.2% | 44 | 11% |
| buy at 60 minutes only if the coin held within 15% | 35 | 80% | 43% | 29% | 1.41x | +195.7% | 35 | 5.7% |
By instruction (buying at the alert)
| Instruction | Trades | Share of signals | Reached +50% | Reached +100% | Median peak (x) | Average peak | With a later reading | Down 80% or more |
|---|---|---|---|---|---|---|---|---|
| admin_set_creator (to 11 Sep 2026) | none this week | |||||||
| admin_cto (from 11 Oct 2026) | 44 | 100% | 41% | 32% | 1.42x | +515.1% | 44 | 11% |
admin_set_creator is the instruction pump.fun removed on 12 September 2026; admin_cto is its replacement, counted here from 11 Oct 2026. A row with no takeover yet prints none; a share over fewer than 10 readings prints too few readings.
Resting a bid under the alert price (entry price is the bid)
| Bid | Fills | Fill rate | Reached +50% | Reached +100% | Median peak (x) | Average peak | With a later reading | Down 80% or more |
|---|---|---|---|---|---|---|---|---|
| bid 10% under, 1 hour window | 13 | 30% | 15% | 7.7% | 1.06x | +41.4% | 13 | 31% |
| bid 15% under, 1 hour window | 9 | 21% | too few readings | too few readings | too few readings | too few readings | 9 | too few readings |
| bid 25% under, 1 hour window | 6 | 14% | too few readings | too few readings | too few readings | too few readings | 6 | too few readings |
| bid 15% under, 3 hour window | 15 | 34% | 27% | 20% | 1.30x | +50.4% | 15 | 27% |
| bid 25% under, 3 hour window | 11 | 25% | 27% | 9.1% | 1.22x | +38.8% | 11 | 27% |
A bid fills the first time the coin trades at or under the bid price inside the window; peaks are measured after the fill. Unfilled bids cost nothing and are not counted.
What the first hour tells you (outcomes from the alert price)
| First hour | Takeovers | Share | Reached +50% | Reached +100% | Median peak (x) | With a later reading | Down 80% or more |
|---|---|---|---|---|---|---|---|
| held within 15% in the first hour | 35 | 80% | 51% | 40% | 1.56x | 35 | 0.0% |
| dipped more than 15% in the first hour | 9 | 21% | too few readings | too few readings | too few readings | 9 | too few readings |
How this is measured
Every pump.fun community takeover Belphor detected on-chain in the window, at least 37 hours old, with minute-level price snapshots for the 36 hours after the alert (at least 30 readings). Demo and test rows are excluded.
Delayed entries use the snapshot closest to 30 or 60 minutes after the alert as the entry price and only count peaks after that moment. The confirmation variant additionally requires the coin never to have traded more than 15% under the alert price during the first hour. Bids fill at the bid price the first time a snapshot is at or under it within the window. A share over fewer than 10 readings prints as too few readings: the row's trades or fills for the outcome columns, every takeover counted here for the share of signals and the fill rate.
"Reached +X%" is measured from the entry price of that variant. "Down 80% or more" means the last reading in the 36-hour window was 20% of the entry price or lower; every trade counted here has that reading, so the "With a later reading" column equals the trade count. The Belphor engine can rest exactly these bids (dip bid entries, engine v29); these numbers are the evidence behind that feature, published so they can be judged.
The minute-level readings come from Belphor's snapshot table for the last 45 days and from its archive for anything older, so a takeover's 36 hours are read in full whatever its age.
pump.fun replaced its takeover instruction on 12 September 2026. This page counts takeovers detected under the earlier instruction, admin_set_creator, the last on 11 September 2026, and, from 11 Oct 2026, takeovers made under the new instruction's new-creator form, admin_cto, which Belphor alerts on and counts as takeovers behind the liquidity, mint safety and rug registry gates. The "By instruction" table shows the two populations apart. Rows recorded under the new instruction before 11 Oct 2026, and its holder-reward form, are not takeovers here.
Frequently asked
Is it better to buy a CTO immediately or wait for confirmation?
This record does not separate them yet. Buying at the alert reached +50% 41% (n 44); waiting 60 minutes and buying only if the coin had held within 15% reached 43% (n 35). Down 80% or more at the end of the window: 11% (n 44) at the alert against 5.7% (n 35) with confirmation.
Does buying the dip after a CTO work?
Resting a bid 10 to 25% under the alert price filled on 14% to 34% of the 44 takeovers, depending on depth and window (fills n 6 to 15). On reaching +50%, the record does not separate any of them from buying at the alert (41%, n 44): they range from 15% to 27% on n 11 to 15. On the share down 80% or more, the record does not separate any of them from buying at the alert (11%, n 44): they range from 27% to 31% on n 11 to 15. The tables above give the exact numbers for each depth.
Does this page count pump.fun's new takeover instruction?
pump.fun replaced its takeover instruction on 12 September 2026. This page counts takeovers detected under the earlier instruction, admin_set_creator, the last on 11 September 2026, and, from 11 Oct 2026, takeovers made under the new instruction's new-creator form, admin_cto, which Belphor alerts on and counts as takeovers behind the liquidity, mint safety and rug registry gates. The "By instruction" table shows the two populations apart. Rows recorded under the new instruction before 11 Oct 2026, and its holder-reward form, are not takeovers here.
Where does this data come from?
From Belphor's own on-chain detection of pump.fun creator-change instructions and its minute-level price snapshots, recomputed nightly from the raw record. The JSON and CSV files on this page contain the same numbers.
Belphor records pump.fun takeovers on-chain, scores each signal 0 to 100 and publishes measured outcomes with their sample sizes, windows and missing readings. Reflex Bot is coming soon and priced separately.
Explore Belphor →