Strategy builder

Describe it. Tune it. Test it honestly.

Write your idea in plain English, then refine every parameter with no-code blocks. FSP compiles it to look-ahead-safe Python and runs the full honesty suite on the result.

Plain English

The AI parses this into the blocks below — edit either side. (Compilation runs when the engine is live; the blocks are pre-filled here as a preview.)

Name

Symbol

Timeframe

Entry — all must be true

Exit — any can trigger

Risk

Risk / trade %

Target (R)

Stop

Compiled rule-treeJSON → Python

{
  "strategy": "RSI mean-reversion",
  "universe": {
    "symbol": "EUR/USD",
    "timeframe": "1h"
  },
  "entry": {
    "all": [
      {
        "RSI(14)": {
          "<": 30
        }
      },
      {
        "Price": {
          ">": "EMA(200)"
        }
      }
    ]
  },
  "exit": {
    "any": [
      {
        "RSI(14)": {
          ">": 60
        }
      }
    ]
  },
  "risk": {
    "perTradePct": 1,
    "targetR": 2,
    "stop": "ATR(14) × 1.5"
  }
}

Tests run on every result

In-sample / out-of-sample split (70 / 30)Reserve unseen data to catch curve-fitting.
Walk-forward optimizationRe-fit on a rolling window, test forward.
Monte Carlo trade-order shuffle (1,000 runs)Confidence band on the equity curve.
Deflated Sharpe + probability-of-overfitPenalize the number of variations tried.
Slippage + commission modelCosts applied to every fill.
Survivorship-free, point-in-time dataNo peeking, no vanished losers.

Start free to run it

Preview — the live engine + candle charts are landing next.