# Addendum 1 to the 10-second flip paper brief (FW Analyst 2, 8 Oct 2026)

No change to the main rule (buy at +10 s, sell 60 s later). Please add the following LOG-ONLY items to the same paper positions.

## 1. Extra shadow exits: +180 s and +240 s after our fill
Backtest on the "F1" coins (standard launchpad token, factory 0x58daec3116aae6d93017baaea7749052e8a04fa7, fee 1%), priced from real trades, a pull or no real sell in the 10 s after the exit = 0:

| exit | 20-30 Sep mean | 1-7 Oct mean | median | zero |
|---|---|---|---|---|
| 60 s (main rule) | 1.35x | 1.34x | 1.43-1.47x | 15-16% |
| 120 s | 1.71x | 1.80x | 1.86-1.98x | 23-26% |
| 180 s | 1.96x | 2.19x | 2.19-2.37x | 29-34% |
| 240 s | 2.00x | 2.36x | 2.22-2.54x | 35-44% |

Why: this launchpad pulls every coin's liquidity itself (100% of 8,488 coins, Sep 20-Oct 7), median 7.5 min after the first swap, with a timer cluster at 576-595 s. Until the pull the price climbs steadily (about 1.46x of the +10 s price at 60 s, 2.3x at 3-5 min). Early pulls (27% by 150 s) hit coins with few buyers.

## 2. Tag each coin at entry
- f1 = factory 0x58daec3116aae6d93017baaea7749052e8a04fa7 AND fee 10000.
- buy ETH and number of buys in the first 10 s (from swaps). Backtest: buy ETH >= 0.5 lifts the 180 s exit to 2.14x / 2.39x on about half the coins.
- pull time (seconds after first swap) once it happens.

## 3. Regime gate (log only)
For each entry, the mean result of the last 20 flips that had CLOSED before it (60 s exit). Backtest: trading only when that mean is >= 1.0 cut the worst day from -$12,834 to about -$300 at $100 a trade, total profit unchanged.

## How the result is read
Report the five pass-bar numbers of the main brief for: all coins, f1 only, and f1 at the 180 s exit. A 180 s live version goes to the pipeline as a separate idea (A2-FLIP180) only if its paper result holds.