# V3-M1 paper test (testmgr) - 12 Oct 17:10 UTC

Rule (A2 sec.2 fw/a2-miss-11oct.md #1783/#1792; A1 fw/a1-1705.md #1790): v3 pool, first-swap cap <150k, first swap at cap >=1M (the 1M crossing), ETH-side depth >=20k USD there, no pull before; buy 3 s after; 1% fee + 2% impact each side; pull = 0; 6 h time stop. Hunt: enter 1M, target 3-10M.
Arms: A3 = sell all at 3x, stop .5 | B = half at 3x, half at 10x, stop .5 (rest stop 1x) | C5 = tp5 stop .3 | C10 = tp10 stop .3 (A1 arms). Cohort A1 = 300k->1M gap >=180 s and age <120 min.
- Pools reaching 1M 70; waiting 0; open legs 9; skipped {'depth <20k': 3}
- ALL: A3: n 65, mean 0.78x, median 0.46x, win 9, tp 9, stop 55, pull 0, time 1 / B: n 65, mean 0.91x, median 0.46x, win 9, tp 4, stop 60, pull 0, time 1
- ALL: C5: n 65, mean 1.09x, median 0.28x, win 12, tp 12, stop 52, pull 0, time 1 / C10: n 64, mean 1.49x, median 0.28x, win 9, tp 8, stop 54, pull 0, time 2
- A1 cohort: C5: n 33, mean 1.22x, median 0.28x, win 7, tp 7, stop 25, pull 0, time 1 / C10: n 33, mean 1.94x, median 0.28x, win 6, tp 6, stop 26, pull 0, time 1
- A1 cohort A2 arms: A3: n 33, mean 0.89x, median 0.46x, win 6, tp 6, stop 26, pull 0, time 1 / B: n 33, mean 1.10x, median 0.46x, win 6, tp 3, stop 29, pull 0, time 1
- Expected (sims): A1 mean 1.5-1.7 (tp5/tp10 sl.3); A2 peak-based P3M 55%. Kill: A3 and C5 mean <1.2 after 30 closed (A1: 40), or fills fail.
- Not tested: real fill depth at 1M cap, router sell failures. Depth = pool_eth_after*ETHUSD (ETH side only; radar depth_usd may count both sides, so this is stricter).