- 2026-10-05
7 min Days 17–25. Silent failures, and a queue I do not win yetTwo bugs made the hunter silent for a day without any alarm, then the clock and the oracle forecast were fixed. The first win came on 2 October at $26.7, and the next two days showed that delivery, not the model, is what costs the places.
- 2026-09-25
8 min Days 13–16. The sniper is closed, the hunter meets the oracleThe sniper closed with its numbers — 90% of the loss was eight tokens that cannot be sold. The liquidation hunter moved next to the sequencer, learned the oracle's clock, fired about 250 times and has hit nothing yet. What every miss taught.
- 2026-09-21
6 min Queue, not speed. How I watched $16k of liquidations and took noneTwo liquidation clusters on Morpho in one week; my watcher saw all 40 positions in advance — and got zero. Dissecting the winner's blocks showed he isn't faster. He just stands in line every block.
- 2026-09-21
12 min Twelve days with $200. The diary, losses includedDay by day — from the first scanner to the first week of real money. What broke, which rules were born from losses, which directions were closed with numbers. Balance at the end of week two — $200 → $138.
- 2026-09-20
12 min 45,000 pools in four days. Anatomy of a meme market on a new L2What the scanner showed once I recorded every pool, swap and liquidity move on Robinhood Chain — before putting in the first dollar. Five different worlds instead of one market, traps, factories, and the single feature that predicts anything.
- 2026-09-19
1 min $200, one laptop, a new L2Public experiment — can automatic trading compound $200 toward $100k in a year. First nine days, with the losses left in.