2026-11-02
Trench lessons — 2026-11-02
Key takeaways
- Airdrop-style "holder payments" coins create hundreds of wallets that sold without ever buying: exclude those coins from any wallet's profit math
- The same farm deployer posted 21 graduations on Nov 1 after 11 on Oct 31: use a rolling multi-day graduation count as a skip signal
- Same-second sell waves and throwaway re-buy hops are cheap launch-bundle tells
- All-in swap cost on popular terminals runs about 4% per round trip, so a copy edge has to clear that
$uni: a "holder payments" coin where 846 wallets sold without buyingevidence: B: pumpdata.fun recorder; C: X posts
- $uni launched and graduated in the same second on Nov 1 (2:53 PM CT) and was the day's busiest coin: 92,285 SOL traded by 14,670 wallets, with a thin-float peak of about $203M. The deployer started with 99.2% of supply and sent 98.9% out by plain transfer, sold 0.30% for 2,091 SOL in about a minute, and now holds none.
- 846 wallets sold $uni without ever buying it, taking out 2,747 SOL. On X, one recap counted a six-KOL swarm putting 135 SOL in before the coin fell 96%.
- Lesson: if you rank wallets by "sold for more SOL than they bought", airdrop recipients look like infinite-ROI geniuses. Rejecting sell-only wallets isn't enough, because a real trader who also cashed an airdrop gets inflated numbers. Exclude, per coin, any coin a wallet sold without buying from its round-trip profit and repeat-win count. Wallets that bought these coins late on a KOL call should score as losers, not signal.
Deployer graduation jumps, day threeevidence: B: pumpdata.fun recorder
- The farm deployer that had 11 graduations on Oct 31 had 21 on Nov 1 from 703 launches. Others had 12 graduations from 468 launches and 11 from 241, and one launched 704 coins for 5 graduations.
- Lesson: a one-day trigger misses repeat farms. Skip coins from deployers with 10 or more graduations on any recent day, or 20 or more over three days, and don't count wins on those coins as evidence of smart money.
Launch-bundle tells from rug timelines: same-second sell waves and throwaway re-buy hopsevidence: C: GORZ recorder pages
- Nov 1 timelines. $BALLS: 8 wallets bought 68% in the first 3 seconds, then 6–7 wallets sold within the same second every couple of minutes, ending −97% in an hour. $HALLOWEEN: 6 wallets took 79.3% in the creation slot, then repeated same-second sells of 6–9 wallets, with about 900 trades but only $941 of volume in 5 minutes. $Amazon: 84 and then 145 wallets sold in the same second. $TON: 86 times, sale money went through a throwaway account to a fresh wallet that bought the coin back, and 5 coins named $TON launched within 30 minutes. $Mincho: a deployer flagged with 260 launches in a few days, 89 of 108 dead within an hour.
- Lesson: two cheap link tests. If a wallet's exits keep landing in the same second as five or more other wallets on the same coin, it's part of a bundle, not an independent trader. And sale proceeds routed through a one-use account into a new wallet that re-buys the same coin mark the whole chain as one actor. A "two different wallets bought it" confirmation can be fooled by two bundle wallets, so check pairs for co-selling overlap.
Venue deploy check: no changesevidence: A: mainnet RPC read of program data
- No new deploys since Nov 1 for the pump.fun curve and PumpSwap (Oct 19), Raydium CPMM (Oct 30), Raydium CLMM (Oct 22), Raydium AMM v4, Jupiter v6 (Oct 29), Raydium LaunchLab or Meteora DBC. Raydium's changelog is unchanged since Oct 30 and Jupiter has no November changelog yet.
Smaller itemsevidence: B/C, no action
- pump.fun's 4x referral boost ended Nov 1; referrers now get 37.5% of referred users' fees. The app platform fee applies only to pump.fun's own web and mobile app, not third-party interfaces.
- From Nov 30 at 9:00 AM CT, pump.fun retires web sign-in with external wallets like Phantom, and linking asks you to paste a wallet's private key into its dialog. Expect phishing sites that copy this flow, and never paste a trading bot's key anywhere.
- First day after Binance's Solana reopening, flow was still soft: 36,904 launches (−22% vs the 30-day average), 893 graduations (−18%), launching wallets −41%, trades −34%.
- OpenChainBench (24 hours to Nov 2) measured all-in cost per swap at 414 bps on Axiom and 412 bps on Photon. Any copy strategy needs well over about 4% per round trip to stay profitable.
Sources
- https://pumpdata.fun/report/2026-11-01
- https://pumpdata.fun/stats/2026-11-01.json
- https://x.com/TRENCHlE/status/2117251413998735618
- https://x.com/TrueGemHunter/status/2117174342966534611
- https://gorz.app/coin/FfXpLJbqpkKR6wNv6eWCVTHx9ku3k9tJRETP9Ypypump
- https://gorz.app/coin/aLvJ4bQXvZR1JhhnH6TSwyHaFL8ptp8ijvDMmyCpump
- https://docs.raydium.io/reference/changelog
- https://developers.jup.ag/changelog
- https://dexsheet.com/pump-fun/news/
- https://thedefiant.io/news/defi/pump-fun-cuts-referral-share-to-37-5-as-4x-boost-ends
- https://pump.fun/docs/wallet-login-changes
- https://openchainbench.com/compare/axiom-vs-photon
Evidence grades: A = large dataset, B = official docs, C = vendor/thread/small sample. Research notes for education only. Not financial advice.