Strategy in, orders out, every step traced.
Automation across brokers, prop firms, and platforms, with the risk guards that keep it from running away.
Toggle the costs. Watch the promise shrink.
The gap between the raw signal and the net curve is the fees and slippage most backtests quietly leave out. We report the honest one, with the risk guards that keep a bad run survivable.
Strategy in, orders out, every step traced.
A signal doesn't become an order until it clears the risk gate. Size, stop, and the daily loss cap are checked before anything is sent, the order carries its own protection, and every step is logged, so a trade is always explainable and a runaway is stopped before it starts.
A rule you define fires on confirmed data, the same way every time. No discretion, no hesitation.
Position size, stop, and daily-loss checks run before an order exists. Fail one and it's blocked.
Sized to the risk, with the stop and target attached in the same request, not bolted on after.
Routed to MT5, a prop firm, or the exchange, and tracked to a real fill, not assumed.
Signal, decision, order, and fill are logged, so every trade is explainable after the fact.
Guards that keep it from running away.
The market outcome isn't ours to promise, but how much a bad run can cost you is ours to bound. Limits, stops, a daily loss cap, and a kill switch are declared up front and enforced before every order, so the worst case is survivable instead of an account-ender.
1# risk.yml — the guards that run before every order2max_position: 3 # contracts, hard cap3per_trade_stop: 300 # dollars, attached to the order4daily_loss_cap: 1500 # dollars, then flatten and halt5no_averaging_down: true # never add to a losing position6kill_switch: manual + auto # one command, or on breachEvery order is sized from the stop distance and your risk-per-trade, so one loss is a known number.
Cross it and the system flattens and stops for the day. No revenge trading, no 2am blow-up.
One command cancels every resting order and closes out, and breaches trip it automatically.
The single most common account-ender is disabled by policy, not by willpower.
A green backtest is a hypothesis, not a promise.
It's easy to make a backtest look incredible by leaving the costs out and fitting the parameters to the exact past you're testing on. That curve sells, and it doesn't survive contact with a live market. We build the honest one, and we're plain about what it can and can't tell you.
- ✓commissions, per contract
- ✓slippage on entry and exit
- ✓an out-of-sample split
- ✓realistic fills, no perfect entries
- ✕look-ahead bias
- ✕survivorship in the data
- ✕zero-cost, perfect fills
- ✕parameters curve-fit to the past
We scope the tool. We don't promise the market.
A trading system is software: it executes a rule you define, fast and the same way every time. What it can't do is guarantee a market outcome, and anyone who tells you otherwise is selling you something. So we build the execution, the risk guards, and the honest backtest, and we stay clear about the line: how the tool behaves is ours to guarantee. What the market does isn't.
What we build
10 buildsTell us the strategy, not the target.
Send the rule and the risk limits it has to respect. We come back with a scope, a plan, and where a demo proves it. We won't promise a return.