Skip to content
Ask · beta

Test an idea in plain English

Describe a setup or ask a data question. The assistant turns it into exact rules, writes the backtest, runs it on every NIFTY 50 stock and reports what happened after costs, with the code it ran.

Universe
NIFTY 50 constituents
Period
Jan 2016 – Aug 2026
Resolution
1-minute bars
Costs
Brokerage, taxes, slippage
New analysis
Sandboxed code⌘↵ to run0 / 2000

Example prompts

How it works

  1. 01
    Exact rules

    Your idea is turned into precise entry, exit and sizing rules, with every assumption stated.

  2. 02
    Code written

    The assistant writes a backtest against the open dpx_bt engine.

  3. 03
    Run in a sandbox

    It runs on the full minute history, with variants and a random-pick benchmark.

  4. 04
    Report after costs

    A plain verdict, the numbers by period, warnings and the code it ran.

Example answer

What you get back

A real run, start to finish. Ask returns a plain-English verdict, a full report, insights, every trade and the code.

Question

Does trading the 15-minute opening range breakout on NIFTY 50 stocks with a 1% stop and a 2% target make money?

1,20,643 trades 2016 – Aug 2026

Verdict: no. Over 120,643 trades from January 2016 to August 2026 this setup lost ₹2.49 crore after costs at ₹5 lakh per trade, and it lost money in all three periods (2016–20, 2021–23 and 2024–26).

  • Before costs the average trade made only +2.4 bps; a round trip costs about 6.5 bps.
  • Win rate 43.6%, profit factor 0.91. 11,833 trades hit the 2% target, 38,681 hit the 1% stop and 70,129 were closed at 15:15.
  • With 3 extra bps of slippage per fill the loss grows to ₹6.10 crore.
  • Random benchmark: random NIFTY 50 stocks traded at the same minutes in the same direction averaged 2.84 bps before costs; this strategy beat 3.5% of 200 random runs.
  • The least bad entry hour was 12:00 (-0.9 bps per trade after costs), still negative.

Assumptions: 5-minute bars; range = 09:15–09:30; entry at the next bar's open after the first 5-minute close beyond the range; one trade per stock per day; entries 09:30–14:30; everything closed at 15:15; stop and target measured from the fill price. Variants run: 1.

  • Edge smaller than costs: 2.37 bps gross vs ~6.5 bps of costs per trade.

Answers are generated by an AI model from backtests on historical data. They can contain mistakes: check the assumptions and the code shown with each run. Research and education only, not investment advice. Disclaimer.