2026-10-13

Trench lessons — 2026-10-13

Key takeaways
  • Bitquery, 1.75M Mayhem coins: graduations are hollow, the Mayhem bot leaves in ~39 s, outside wallets lose net
  • $HIRE (Trenchbook, Oct 12–13): creator bought 22% in the create tx and sold it 2 s later, then two linked wallets cashed out to one Kraken deposit address
  • Bitquery, failed swaps (Oct 10): at least 22% of Solana swaps fail; for small wallets it's mostly memecoin sells on PumpSwap hitting their own slippage
  • Smaller notes
  • Platform watch: nothing breaks

Bitquery, 1.75M Mayhem coins: graduations are hollow, the Mayhem bot leaves in ~39 s, outside wallets lose netevidence: A dataset

  • What the study found (Nov 12 2025 – Oct 7 2026):
  • The Mayhem bot traded 1.75M coins (19% of pump launches); Mayhem has been 20–30% of launches every month since April, mostly from serial devs (100+ coins).
  • Only 1 in 50 Mayhem graduations filled the PumpSwap pool with the normal 85 SOL. The bot's trades reset the curve price far more than their size (median finished curve quotes 28% of a fresh coin's price), so the last buy graduates the coin for a few SOL. 26 of 31 checked Jul–Sep graduations put <1 SOL in the new pool.
  • The bot trades a coin for a median 39 s (90% done within ~4 min), not 24 h. Since July it also trades graduated coins on PumpSwap in bulk. It burns leftovers, so supply ends near the normal 1B.
  • Wallets other than devs and the bot put ~61,600 SOL more in than they took out on Mayhem coins in week one. Sandwich bots took 6% of outside SOL on them.

$HIRE (Trenchbook, Oct 12–13): creator bought 22% in the create tx and sold it 2 s later, then two linked wallets cashed out to one Kraken deposit addressevidence: C trace

  • What happened (times CT):
  • The project's "central" fee wallet sold fee tokens on PumpSwap and routed the SOL in the same tx to the launch wallet (15.88 SOL in 7 of 8 sells). 37.2% of claimed fees became SOL vs the 25% the docs promise.
  • Same launch wallet did the same 2-second dump on $HIVE (Oct 9), a "launchpad" whose coins are paired with $HIVE.

Bitquery, failed swaps (Oct 10): at least 22% of Solana swaps fail; for small wallets it's mostly memecoin sells on PumpSwap hitting their own slippageevidence: A dataset

  • 2.4B of 11B swap attempts failed (true share nearer 27–39%, since pool-rejected fails are missing from records, mostly PumpSwap). Failed swaps still paid ~$11.3M in network fees; a tip inside the tx is not charged on failure.

Smaller notesevidence: C

  • Fomo leaderboard caution (unverified): an X post claims top Fomo-leaderboard traders coordinated with the app's team to pump a couple of tokens. If Fomo or Solana Tracker's FOMO leaderboard is ever used as a wallet source, run the same-second cofire check on those wallets first. ,

Platform watch: nothing breaks

  • Solana: Agave v4.3.1 and v4.4.0-beta.1 tagged Oct 13 ~1–2 AM CT (no release notes posted yet); nothing found about a transaction-format or RPC change.
  • Meteora: DBC 0.2.2 still due Oct 15 ~10 PM CT (covered 10-11).

Sources

  1. https://bitquery.io/investigations/pump-fun-mayhem-mode
  2. https://pump.fun/docs/mayhem-mode
  3. https://trenchbook.app/dispatch/hire-launch-wallet
  4. https://bitquery.io/investigations/solana-failed-swaps-arbitrage-bots
  5. https://solanacompass.com/projects/dedge-solana-bot
  6. https://x.com/Sybiliooo/status/2109609651939725691
  7. https://www.solanatracker.io/leaderboard/fomo
  8. https://t.me/s/pump_tech_updates
  9. https://github.com/pump-fun/pump-public-docs
  10. https://github.com/anza-xyz/agave/releases
  11. https://github.com/helius-labs/core-ai/commit/92fe708c0ae377d732e690bd9e9af8d745245ac0

Evidence grades: A = large dataset, B = official docs, C = vendor/thread/small sample. Research notes for education only. Not financial advice.