41 Signals Later: +2.7% at One Hour, −31.8% at One Day

Every ignition signal our scanner fires gets paper-traded: entry price snapshotted, then the return measured at +1 hour and +24 hours. No cherry-picking, no "the good ones count." As of this morning, 41 signals have resolved at the 1-hour mark and 33 at the 24-hour mark. The gap between those two numbers is the entire story of trading new tokens.

Speed is the whole edge

A +2.7% average at one hour and a −31.8% average at twenty-four hours describe the same tokens. The difference isn't luck — it's gravity. New-token liquidity spikes, attention moves on, and what was a momentum trade at the one-hour mark becomes a decaying asset by the next morning. Twenty of the 40 one-hour windows were losers too, but they were small, fast losers. The 24-hour holds are what crater.

That is why our High-tier paper bot runs a universal 1-hour fast exit. It's not a clever optimization; it's the only posture the data supports. Exit at one hour and you keep a modest edge. Hold for a day and you donate −31.8%.

The strategy bake-off scoreboard

• Ignition momentum (score jump 15+ or new-and-hot): 29 resolved at 24h, average −29.6%, 2 winners of 50%+. The flagship strategy is losing.
• Phoenix mean-reversion (crashed 60%+, then buy): 4 resolved at 24h, average −47.5%, zero winners. Still the worst seat in the house.
• Boost attention arbitrage (paid DexScreener boosts vs return): 39 resolved at 24h, average −41.3%, 2 winners of 50%+.

All three strategies are underwater at the 24-hour horizon. The only part of the system that ever looked healthy was the 1-hour window — and the margin there is +2.7%, a thin edge after fees, not a money machine.

The scanner keeps firing, the auto-analyst keeps logging, and the next strategy tweak goes into the same public record as every loss before it.