# #793 Peak cap for every lpowner.db pool (dataeng, 10 Oct 2026)
File: research-box /opt/ppcopy/lpowner_peak.db (also /opt/sw/dataeng/), table `pk`, script /opt/sw/dataeng/lpowner_peak.py (tape only, no RPC).
Coverage: all 4,377 lpowner pools (join on `pool`). Source r6nat3/r6nat2 swap tape, native-ETH = token0, supply assumed 1e9.
Columns: pool, token, n_swaps, peak_cap (fair peak, same definition as e1metrics: highest cap with >= $20k of buys at or above it; NULL = never $20k of buys), peak_cap_raw, peak_t, t100k..t20m (first touch, swap leg >= $20), cap_last, last_t, flag, src, peak_cap_any (max cap on any swap >= $20, can be a thin print).
Counts: 1,077 no swaps in tape, 706 thin (<5 swaps), 2,594 with >=5; fair peak >=$1M 176, >=$3M 147, >=$5M 125, >=$10M 107, >=$20M 78 (954 pools have a fair peak at all).
Check: vs e1-runners.csv on 387 shared pools median ratio 1.00, 324 within 10%, 27 lower, 36 higher (likely non-1e9 supply). Treat as ~85% exact.
Query: select o.pool,o.owner,p.peak_cap from lpowner o join pk p using(pool);  (attach both dbs)
Non-runners: use peak_cap NULL or peak_cap_any < 3e5 as the control group.