CRYPTO
BTC · ETH · SOL — spot, 24/7 · research-first lane
THE LAB
A hypothesis must survive discovery → walk-forward → sealed OOS → paper to earn a signal hereNo crypto signals are published yet — and that is the feature. Every candidate strategy is kill-tested with pre-registered gates on exchange-native data after real costs. Most hypotheses die. The ones that survive show up here with their full record.
ALPHAPULSE
Everything the system does, as it happens · the machine's own wire
PERFORMANCE
Mechanical triggers, untampered · updated daily
Recent Performance (Last 7 Days)
| Date | Signals | Win Rate | Avg Gain | Result |
|---|
What You Missed (Examples)
More opportunities
locked
QUANT REPORTS
Daily after-hours momentum watchlist · generated mechanically by the engine
METHODOLOGY
How the engine finds, ranks and risk-manages momentum
The Radar. Every 60 seconds we scan the full US equity universe on consolidated (SIP) market data — including the extended session from 4:00 AM ET. The scan is defended against the two classic sources of fake movers: tickers with no trades today (stale prints) are excluded, and recent reverse/forward splits are detected via corporate-actions data so a 1-for-5 split never shows up as a +400% "runner".
Runner Score (0–100). A mechanical profile of pre-explosion conditions: float size, relative volume, float rotation (turnover), session structure and the presence of a real catalyst. It is not a prediction — it measures how closely a stock matches the profile that historical +100% runners shared before they ran. Scores of 85+ are rare by design. Our public ledger shows roughly half of 85+ alerts touch +10% intraday after the alert.
The Day-0 model. A machine-learning classifier trained on the historical intraday behavior of this exact universe — small-cap movers, not the S&P. It estimates the probability of an intraday extension for each radar ticker (the "AI" number on each card). We evaluate it walk-forward: trained only on data available before each trading day, never on the answers. Our own measurements show the edge in this asset class is intraday; overnight holds are where momentum goes to die, and our published results reflect that honestly.
Breaking Catalysts. Press-release wires and Benzinga headlines stream in real time. Each headline is classified (FDA, contract, M&A, earnings — and dilution flagged in red, because offerings kill runners), and we freeze the price at the moment the headline hit. The "since news" figure you see is the stock's actual move from that frozen price — measured, not narrated.
A+ signals & risk. Setups (VWAP reclaim, pullback, breakout) come with a defined entry, a stop anchored below VWAP, and two targets. The stop is not decoration: in this asset class survival is position sizing plus a stop you actually take. Halts and spreads are real — size accordingly.
ML HISTORY
Raw daily dataset the models train on · every scanned ticker, archived
| Date | Symbol | Close | Change % | Vol | Rel Vol | RSI(14) | MACD | MA(20) | Vs MA(20) | VWAP | Market Cap | News | Sector |
|---|
AI PREDICTOR
Dual engine over 150+ historical variables · intraday and swing projections
DAY-0 WATCHLIST
The model's top-20 at the close — tomorrow's radar, tonight. Out-of-sample: on an average day ~3 of the next session's 20 biggest movers are on this list; the top-3 touch +20% intraday on 37.5% of days. DIL = dilution/reverse-split risk flag. Not financial advice.
Instant Predictor
Type a ticker, get the model's intraday and swing projection
ADMIN PANEL
Superuser access · registered users, live presence, engine feedback
SPY 0DTE · GAP SIGNAL
Signal window 09:45 ET · 08:45 CST — gap up/down confirmed by the first 15m candle
PROFITABILITY
—Broker Auto-Trader
Set capital limits and switch the paper-account auto-trader on.
AlphaPulse Push
Send a test notification to registered devices. "To me" targets devices logged in with your email.
Alerts-Pause Announcement
Signals are paused. Send the announcement: test lands only in your inbox; "everyone" emails all registered users and posts to the Telegram channel.
Quant Engine Feedback Loop
Supervised learning: daily predictions vs realized outcomes.
| Date | Ticker | Max Gain | Reason | Score |
|---|---|---|---|---|
| Click Refresh stats to load data | ||||
Registered Users
| ID | Email Address | Role | Created At |
|---|