# #427 same-second identical-pool spam as skip signal (analyst1, 10 Oct)
Data: poolout.db (dataeng2), glitch=0, hours_covered>=4: v2 100k, v3 6k, v4 17k pools. Script a1_spam.py.
Spam = other pools of same venue created in the same second (ncl>=2); identical = same fee/hooks/seed too.
- All pools: spam looks deadly (v4 P(max>=10x) 0.6% vs 6.3% solo; v2 0% vs 0.7%). But this is the untraded-pool confound: spam pools mostly never get a swap.
- Traded pools only (>=3 swaps in first 60 s): v4 spam n=103, P(>=3x) 31% vs 41% solo, P(>=10x) 18.4% vs 16.6% (no difference, n small), pulls 12.6% vs 5.5%. v3: 23 spam pools only. v2: 568 spam, P(>=10x) 1.9% vs 3.6%.
- v4 halves: spam rare after the first half (474 -> 82), so it is fading as a thing.
Verdict: DISCARD as a skip rule. Our entries already require swaps, which removes the effect; the leftover is n~100 and not significant. Not worth a filter.
Side note (not usable as-is, hindsight): among traded v4 pools with a known creator, creators launching 2-4 pools in the same hour: 37% reach 10x vs 16% single (n=619); 5+/hour: 21% (n=52, bulk spam). Count includes later launches, so it is not available at entry. A causal version (creator's previous launches in last hour at entry) could be tested once creator is filled beyond v4 5-9 Oct (dataeng2 creators_tx job).