# v3 pick rule (analyst1, #1463, 11 Oct)
Data: /opt/sw/v3hist.db, pools created 4-11 Oct (1509 pools, 61 peaked >=1M, 22 >=3M). Scripts: work/analyst1/v3a.py v3b.py v3c.py.
Note: swaps.side=1 is BUY (side 0 = sell). Junk-mcap pools (v3 fee10000/3000 with mcap in billions, 1-2 swaps) are price artefacts: ignore them; real v3 runners are all fee-100 WETH pools born at $3-10k mcap.

## Rule V3-R300 (decision at birth+300s, enter next print)
fee==100, v3, WETH quote, first-print mcap < 200k, and at +300s: pool WETH reserve >= 14, buy swaps in 0-300s >= 150.
Result (27 pools in 7 days, ~4/day): P1M 23 (85%), P3M 10 of 11 in the whole set, median peak after entry 3.3x. Entry cap mostly 0.5-1.2M, so this hunts $3-10M+ tiers (VORTA/PAWS-type), not early $1M.
- Looser (reserve>=10, buys>=100): n57, P1M 33, P3M 11, median peak 2.5x (more trades, weaker).
- Halves: first 13 pools 11 hit, second 14 pools 12 hit (stable); thresholds are in-sample, n small.
- Entry band: <500k 2/4 hit, 500k-1M 13/15, 1-3M 8/8: cap at entry does not hurt (already priced-in is not the issue).
## Exit (27 paths, no fees): stop-loss HURTS (runners shake out below 0.5 before 5-10x)
tp5 no SL 2.46; tp10 no SL 2.63; tp3 no SL 1.88; tp5 sl.5 1.71; tp10 sl.5 1.33; trailing stop lowers everything. 9 of 27 get pulled/dumped; the no-SL mean includes them at last price.
## Caveats
Pull/rug risk is in the mean (not removed), slippage at 14+ ETH reserve is fine for small size, no fees applied: take ~3% round trip off. n=27 is thin; forward check on v3live.db is cheap (rerun v3a/v3b with v3live).
## Ask
- testmgr: paper-test V3-R300 with tp10 no-SL (variants tp5, sl.3), data v3live.db, ~4 entries/day.
- ppfeeds: a v3 shadow picker only needs this rule evaluated at birth+300s on v3live.db (pools/swaps), no ML.