LGB / RF / Linear / Adaptive each trade independently using their own model. ClaudeSpec is a meta-bot
that only fires on full-unanimous BUY consensus with DA ≥ 55%, no near-term earnings, and no FOMC week.
Performance by Horizon
▼
Horizon
Open
Closed
Target %
Avg P&L / trade
Loading...
Position Sizing — flat vs conviction-weighted
▼
What-if only — the live bot still trades a flat $1k. This keeps total capital the same and re-tilts it by conviction, to see whether sizing toward stronger signals would have helped. (Strong has historically earned far more per trade than Moderate.)
Weights:
Loading...
👀 Public view ▾
Paper trades, P&L, and the live scoreboard are real and updated automatically.
Manual scans and outcome resolution are admin-only.
Scanning...0%
Open Positions
Symbol
Bot
Signal
DTE
Entry
Stop
Target
Conviction
DA
Predicted
Opened
Loading...
Trade History
Symbol
Bot
Signal
DTE
Entry
Result %
P&L
Fill P&L
Outcome
Conviction
DA
Opened
Closed
Loading...
Scan Watchlist (62 symbols, three times daily Mon–Fri)
Scans run at 9:35 AM ET, 12:30 PM ET, and 3:55 PM ET on market days.
Criteria: 1-day requires DA ≥ 60% + Strong conviction; 3–5-day requires DA ≥ 55%; 10–21-day requires DA ≥ 52%.
* About the P&L figures.
The headline number is order-fill (capped) P&L: every stop-loss and take-profit is booked at its
order level — or at the day's open when the price gapped straight through it — the way a real bracket order would actually fill.
The smaller raw price figure books the market price at the moment the resolver happened to run, which can overshoot the
stop/target in either direction (a −6% stop resolving at −7.1%, say). Both are kept so nothing is hidden; the order-fill number is
the realistic "if you'd actually traded these levels" result and is what every panel on this page uses. Order-fill values for
stops/targets are recomputed nightly from each trade's daily high/low path.