# Scout 18 Oct 2026
## 25. novamp scoring signals (https://github.com/bored2boar/novamp) - hypothesis, demo weights on synthetic data
Claim: original vs copy of a name is decidable from: birth order (copy 2 min late -22), >=3-4 proven wallets converging within 27-90s (+16), 18 distinct buyers in first minute, dev buy in a 1-6% band (+10), top-10 holders (ex curve/pool) <20% good / >30% bad, deployer graduated 2 of 4 prior launches (+14), 3 socials. TAINTED if first-born is bundle-loaded, serial deployer w/o graduations, or same funder as copies.
Test: on our runner/miss list, check whether runners were the ORIGINAL of their name, dev buy 1-6%, deployer prior graduation rate >=50%, top-10 <20% at +60s. Compare to non-runners (survivor bias: weights are author's examples, no sample size given). Data: pools+tokmeta, first-minute buyers, deployer history (have from rug gate).
Other scanners seen (not read): MEERKAT (kocer6), rh-newcoin-scanner (404), rabiq, FLETCH. The Block: Wazz $18.43M ring, not independently replicated (already covered #22/#9-10).