NOVAMP TERMINAL USERS MADE $100K+ ON ROBINHOOD TOKENS (PONS ONLY) CA: 0xf219746e9fc4edd81e9c377b63df2a6446bcec3a here's all you need to know about it: robinhood chain seals a block every 100ms and ships tens of thousands of launches a day. a name catches. within a minute there are eleven tokens carrying it. but most of them are vamps. you open the chart you were linked, you buy, and the thing that moves is a different contract with the same name. most of what people lose here is not lost to the rug they bought. it is lost to buying number seven while reading number one's timeline. i built a terminal that answers one question. which of the eleven is real. HOW IT DECIDES: > it groups every launch fighting over a name, on the ticker AND the token name. a copy that cannot take the ticker takes the name instead, and a symbol-only matcher never sees that fight > before matching, every string is folded: homoglyphs, leetspeak, plurals, filler words, zero-width characters. a cyrillic T is not a T > order is by block position, not by timestamp. timestamps tie too often at 100ms > then five labels: ORIGINAL, TAINTED, CONTESTED, VAMP, DEAD > being first is not enough. a first launch loaded with a bundle, or funded by the same wallet that funded the copies, is TAINTED and the money is usually somewhere else in the cluster > the sell side. the deployer selling inside five minutes is the heaviest single penalty in the file > early money counted in size, not in wallet count. wallets are free. 25 wallets spending a dollar each is one person > a proven-wallet list built on what those wallets actually sold, not on what they are still holding > and when a name goes from nothing to nine launches in ten minutes from unrelated wallets, it catches the burst and checks whether a headline actually landed before it or after it every point of the score prints the reason it was given. you can argue with a specific line instead of with a black box. WHAT IS NEXT: > exempt wallets out of launch calldata, the strongest signal in the scoring and the live reader still reads 0 for it > fingerprint clustering past the shared funder, for operators who fund through a mixer > scoring the change over time, so it reads a launch losing momentum and not just one that started badly > backtesting every rule against what the chain actually did, and publicly deleting the ones that do not survive 177 tests. 2 runtime dependencies. MIT. every line public. this is just the beginning.