# a2 verdict: gMAK beta-play avoid rule (#333), 10 Oct
Data: 2,114 native pools (r6nat, 21 Sep-9 Oct = 18 days, not 30; no older native data), names for 1,764. Entry sim = FAST5 style (A1 tp3, H3000u tp5, +5 s entry, 10% stake, fees+3% slip). Script it28/b1.py.
Beta = launched within 24 h after a "runner" (peak >=5/10/20x vs first price) AND shares a name word (>=3 letters, stoplist) with it.
| runner def | family | beta n, mean x, zero-rate | rest n, mean x, zero-rate |
|---|---|---|---|
| >=10x | A1 | 229, 1.14, 61% | 1039, 1.08, 62% |
| >=10x | H3000u | 13, 3.96, 0% | 230, 3.27, 27% |
| >=20x | A1 | 128, 1.20, 59% | 1140, 1.08, 62% |
| >=20x | H3000u | 4, 1.63, 0% | 239, 3.34, 26% |
Verdict: claim does NOT hold on our chain at entry-minute horizon. Beta coins are as good or slightly better than the rest (A1 1.14 vs 1.08, rug rate equal). H3000u n too small (13) but not worse. No skip rule; do not send to testmgr.
Caveats: "runner" here is price peak (no $1M cap data), so many runners are tiny; name-word match is crude (no meme/image similarity); 18 days only. Without the word match every coin is "after a runner" (daily runners exist), so the time-only rule is meaningless.
Other claims (Tier1/2 social, DEX paid, template site, GitHub) need dataeng's table; not in my data. Second-wave/quality score belong to A1/later.