Market data
Real exchange candles and venue data. Synthetic fallback is fail-closed and visibly labeled.
Public research preview · paper only
Multi-algorithm trading, with an evidence gate.
MMS HUB is a trading research stack built to test many strategies without confusing a good-looking backtest for a deployable edge. Every candidate moves through the same path: real data, cost-aware simulation, paper evidence, risk review, then a deliberate promotion—or a documented kill.
9strategies adversarially audited
0investor-ready as of Aug 5, 2026
56.5%best verified existing win rate
Lockedlive execution by default
The system
One path from idea to capital.
The product is not a pile of bots. It is the promotion pipeline around them. Strategies share evidence rules, cost assumptions, hard caps, and one source-of-truth ledger so weak results cannot hide behind a different dashboard.
Strategy lab
Regime, momentum, arbitrage, catalyst, mean-reversion, and forecast candidates.
Risk gate
Capital limits, position caps, daily-loss stops, drawdown breakers, and kill switches.
Paper ledger
Timestamped trades, fees, slippage, fills, equity, and comparable forward evidence.
Promote or kill
No live path until sample size, confidence, net expectancy, and manual approval all pass.
Strategy roster
Different ideas. Same burden of proof.
The August audit rejected the “combine everything” shortcut. A portfolio of correlated, sub-threshold systems cannot mathematically manufacture a 70% win rate. The rebuild focuses first on instrumentation and risk truth, then on new strategy research.
Hybrid V6
Regime-aware crypto system and the strongest existing real sample. Still far below the target profile.
181 closed · $30,057.01 fleet equity · 2026-08-26
Rapid Increase
Momentum scanner. Fabricated fallback history remains void; the current-candle, exit-update, and stale-feed defects are repaired; the post-fix forward sample is real now — 34 closed at 55.9% WR.
34 closed post-fix · 55.9% WR · fresh-feed guard verified
Arbitrage cluster
Cross-venue and super-bot variants. Observed spreads did not survive realistic fees and execution costs.
36 real trades · 47.4% fee-exclusive WR
Equity catalyst stack
Schwab rumor, news, and leveraged-ETF research. Low frequency, weak win rate, and stale execution plumbing.
best grid run: 47.95% WR
BTC Brain
Append-only public forecast ledger with model × horizon scoring, baseline comparisons, and shadow promotion gates.
Loading current ledger…
OpenClaw
Maker-only, spot, long/flat execution design. It stays unbuilt until its signal contract proves a real edge.
fail-closed · no leverage · no live capital
Live evidence
The ledger is allowed to say no.
These figures load from the same public files the BTC Brain uses. A promising hit rate is not promoted on its headline alone: the lower confidence bound, sample size, baseline, and fee-adjusted expectancy all have to agree.
All resolved forecasts
3,503
Current candidate
Shadow — not tradeable
The candidate remains behind the promotion gate while the live ledger accumulates evidence.
Loading current gate…
Promotion gate
Every condition must pass.
This is the line between research and capital. A red gate stops the whole chain; no strategy gets to substitute a backtest, a different period, or a friendlier dashboard.
No silent synthetic candles or unlabeled fallback.
Fees, spread, slippage, and real fill quality.
Enough independent trades and regimes for confidence.
Per-trade stop, daily loss, drawdown, and kill switch.
Promotion remains a deliberate owner decision.
What this preview proves
- The system records negative findings instead of deleting them.
- The live BTC ledger can be independently inspected.
- Execution is separated from research and locked by default.
What it does not claim
- No strategy is represented as profitable or investor-ready.
- Backtests are not presented as live returns.
- No return, win rate, or trade frequency is guaranteed.