Alerts paused for calibration · bucket distribution stabilizing post-cutover
why? see status →
LAST 30 DAYS · ≥70% CALLS · LIVE—%
of mints we scored ≥70% in their first 60 seconds
actually graduated
Every prediction publicly hashed before the
outcome was known.
The ML keeps learning.
Verified Jun 2 on a 14-day pre-registered window:
74.7% of ACT calls peaked ≥5× within 24h
(verdict, n=561). Live receipts on
/accuracy.
starred contracts — pinned across reloads (stored locally in your browser).
★
contract
live
vSOL
buyers
odds*
creator
links
no contracts pinned. ★ a row on THE LIVE BOARD to track it.
SNIPER WATCH
wallets that consistently buy in the first 2 seconds of a mint's life and end up holding ruggers. avoid following these.
#
wallet
mints
fast snipes
rug %
good %
links
computing sniper stats…
PAPER P&L · LIVE PROOF OF SIGNAL
Three bots · 1 SOL each · thresholds 70/80/90% · real Jupiter prices · zero overrides.
Each bot enters every mint whose grad-probability clears its threshold (70 / 80 / 90 %) and exits when the rules fire. Prices come from Jupiter v3 — the same data the rest of crypto uses — routed through pump.fun's bonding curve pre-graduation and Raydium post-migration automatically. Positions ride through graduation just like a real trader's would.
ENTRY grad-probability ≥ threshold (signal-only · no price cap) EXIT+20% take-profit · 30 min timeout · 5 min stale-quote RULES bankroll compounds · no human intervention · no fudge factors
RECENT TRADES
latest 30 across all thresholds, newest first
strategy
contract
entry prob
entry mult
exit mult
P&L
hold
reason
state
paper trader warming up — waiting for first signal…