TTrendSentry

[ AI-AUGMENTED // DISCIPLINE LAYER ]

YOUR STRATEGY,
NO DEVIATION.

Connect your Bitget exchange. TrendSentry logs every trade you take, flags when you step outside your plan, and simulates what discipline would have earned you. No emotion, no FOMO, no deviation.

Track your discipline →See the proof

BITGET // READ-ONLY API // FREE TIER OPEN // NO SIGNAL SELLING

CLUSTER-A2 // PAPERLIVE

6Y RETURN

+152%

MAX DD

-26.45%

SHARPE

0.82

SIGNAL

LONG_ENTRY // BTC

STOP LOSS

ENTRY − 2×ATR

✦ AI CURSOR
CLUSTER-A2 // OUR OWN DOGFOOD — NOT FOR SALEFULL METHODOLOGY →

SHARPE (6Y)

0.82

MAX DRAWDOWN

-26.45%

TOTAL RETURN

+152%

TRADES

94

WIN RATE

36.17%

PROFIT FACTOR

2.27

Backtest 2020-08 → 2026-08, 10 pairs, Donchian 20/10 + ATR(14)×2, long-only, risk 1%. Reference figures contain survivorship bias and a wide confidence interval — they are an upper expectation from historical data, not a promise. The live paper dashboard at /papertrading is the honest, ongoing record.

[ WHAT IT DOES ]

The discipline layer for your strategy.

01 // DEVIATION DETECTION — LIVE NOW

Did you stick to the plan today?

Connect your Bitget account with a read-only API key. TrendSentry logs every trade, compares it against your rules, and flags the moment you deviate — before a small mistake becomes a blown account.

CONNECT YOUR EXCHANGE → /START

02 // AUTO-LOGGING

Every trade, on record.

ENTRYlogged with reason
EXITPnL + R-multiple
DEVIATIONflagged in real-time
HOLDsilence = data
ACCOUNTread-only API

03 // ALERTS

Real-time deviation alerts.

Premium: Telegram the second you step outside your plan. HOLD days stay silent.

04 // COMING SOON

Discipline Benchmark.

Simulate what your portfolio would look like if you followed your own rules 100% of the time. The gap between actual and benchmark is the cost of deviation.

05 // TEMPLATES

Start from a template.

Donchian breakout, SMA crossover, RSI mean-reversion — or define your own. Fill in your parameters, TrendSentry handles the rest.

06 // FASE 3

LLM risk filter.

AI sanity-checks valid signals for contradictory risk — never generates them.

DOGFOODING

Cluster-A2 is our own 6-year backtested system running live on this site. It's not for sale — it's our credibility. We eat our own cooking, and we publish the kitchen logs at GitHub.

[ METHODOLOGY // TURTLE-STYLE TREND FOLLOWING ]

Boring rules. Relentlessly followed.

  1. 01

    Execute the rule

    Donchian 20-day breakout for entry, 10-day counter-breakout or ATR stop for exit. Detected exactly one day after candle close — the same bar the backtest used. No look-ahead, no repainting.

  2. 02

    Size the risk

    1% of equity per trade, stop distance from 2×ATR(14), max 2 positions per correlation cluster. Idle cash earns simulated 5% APY in paper — tracked openly, platform risk disclosed.

  3. 03

    Account for everything

    Every signal, fill, slippage sample and yield credit lands in SQLite, committed daily by CI, rendered on /papertrading. Weekly reviews; if live trails the backtest 2–3 weeks running, we pause and investigate.

  4. Read the full proof →

FIELD NOTE // DECISION LOG

“Top-5 trades carried ~100% of net PnL. That's normal for trend-following — the edge is cutting losses fast and letting winners run, not precision.”

decision_log.md →

[ PRICING // PRIVATE BETA ]

Start free. Upgrade when you need alerts.

Watcher tier is open now — connect your exchange, set your rules, and TrendSentry logs your discipline. Premium unlocks real-time deviation alerts.

Start tracking →See full pricing →

[ FAQ ]

Is this financial advice or a signal-selling service?

No. TrendSentry sells discipline of execution, not signals. Cluster-A2 is our own system running live as proof — it is not for sale, and nothing here is financial advice. See /disclaimer.

Can I open an account today?

Yes — the Watcher tier is open now. Connect your Bitget read-only API key, set your strategy rules (or start from a template), and TrendSentry starts logging. Free, no time limit.

What does the live dashboard prove?

That the engine runs daily without human touch: every HOLD, entry, exit, slippage sample and yield credit is logged and committed. Compare live win rate and avg R against the backtest reference — after ≥10 closed trades, not before.

What happens in a losing streak?

Trend-following has long flat and losing stretches by design. Risk stays 1% per trade with a mandatory stop; if live trails expectations 2–3 weeks running, the system pauses for review instead of hoping it recovers.

Is the code open source?

Yes — AGPL-3.0, source-available. Read, learn, and self-host. Commercial SaaS use requires a separate license.

Still deciding? Start with the proof, then connect your exchange.