Your first strategy in 10 minutes
Ten minutes, no code, no broker, no capital. At the end you will have a backtested strategy running in paper.
Before you start
You need: an account you have signed in to. That is all.
You do not need: a broker connection, an exchange API key, or any money. Building, backtesting and paper trading require none of them.
Minute 0-1: Open a new chat
Sidebar, then New Strategy. Or Chats > New Chat.
The welcome panel shows the six inputs the builder works from and the supported symbols. You do not have to supply all six. It will ask for anything missing.
Minute 1-3: Describe the strategy
Type this, exactly:
create a long-only strategy on RELIANCE, 5 minute candles, entry when RSI(14)
crosses above 55 and price is above the 20 EMA, exit when RSI drops below 45,
1.2% stop-loss, risk 1% per trade, max 3 trades a day
Why this one: it is complete, so nothing has to be defaulted except the take-profit; every threshold is numeric, so nothing is ambiguous; and RELIANCE on 5 minutes has enough liquidity and history to produce a meaningful test.
The builder assembles it directly, with no clarifying turn.
Instrument RELIANCE, 5m, NSE (09:15-15:30 IST)
Direction long_only
Entry RSI(14) crosses above 55 AND CLOSE > EMA(20)
Exit RSI(14) < 45
Stop-loss 1.2% from entry
Take-profit not stated, platform default applied
Per-trade risk 1.0%
Max trades per day 3
Everything except take-profit came from your message. Say the word and I will run a backtest.
Read it before you go on
This is the habit worth forming on your first strategy rather than your tenth.
| Check | Why |
|---|---|
| Every value you stated is present and correct | The builder reads your words, but read them back |
| Anything marked default is acceptable to you | A default is a reasonable value, not your value |
| The trading window matches the instrument's session | A window off the session means no trades |
| The direction is what you meant | long_only cannot short, whatever the exit says |
If something is wrong, say so: "make the stop 0.8%". Correcting the builder is the normal path.
Minute 3-5: Backtest it
Type:
run backtest
Or state a period:
backtest the last 6 months
With no period, the default window is 1 January 2024 to now. That floor is the same for every timeframe and asset class. See Choosing a test period
The result carries return, Sharpe, max drawdown, win rate, trade count and a grade, with brokerage, slippage, STT, stamp duty, exchange charges, the SEBI turnover fee and GST already deducted. See Fees & charges
Minute 5-8: Read the result properly
Four questions, in this order. The order matters, return last, not first.
1. How many trades?
Under 20 and the result means very little. The grading system knows this: sample size is one of five scoring components, and it is what stops a three-trade backtest scoring an A.
2. What was the max drawdown?
In money, on the capital you would actually allocate. 18% of ₹5 lakh is ₹90,000. Would you have held through that?
If the answer is no, the strategy is not deployable at that size. And reducing the allocation is a real fix.
3. What is the Sharpe?
Return per unit of volatility. Below 0.5 the volatility is not being paid for.
4. What is the return?
Last, because a good return with an intolerable drawdown is not a good strategy.
And the grade
Six grades: A, B+, B, C+, C, D, from a 0-100 score across five components, return (30), Sharpe (25), drawdown (25), consistency (15), sample size (5).
A grade is a threshold score over the tested window. Not a detection of edge, not a forecast. See Strategy grades.
Minute 8-9: Set your risk limits
The step most people skip, and the one with the largest consequence.
Seven of the nine execution risk gates are disabled until you configure a value. A strategy with none configured will place orders until it runs out of allocated capital.
Say this:
stop trading if I'm down 3% in a day, and stop after 4 losses in a row
Then verify on the strategy's risk panel. Not from the chat transcript. The configuration parser fails open, so a payload that did not parse leaves you with no limits and no warning.
See Which limits are on by default and The nine order gates.
Minute 9-10: Deploy to paper
- Open the strategy and click Deploy.
- Choose Paper.
- Set a capital allocation, use a realistic number, because the percentages only mean something against one.
- Confirm.
The strategy starts evaluating on the next candle close. No broker connection was involved and no real order was placed.
What to do over the next two weeks
| Watch for | What it tells you |
|---|---|
| Does it trade at all? | A strategy that never fires has a condition that is never true |
| Do fills look like the backtest? | Paper models slippage; live adds latency and queue position |
| Does a risk gate engage? | A gate you have never seen trip is one you trust on faith |
| Does the drawdown feel like the number looked? | This is the honest test, and only paper gives it to you free |
Then, and only then
Before any live capital:
- How risk is enforced. The whole section
- Risk disclosure. Every way this can cost you money
- Your responsibilities, what is yours to decide
- Connect a broker, including the permission you must never grant
A fortnight of paper in one market condition is a fortnight of paper in one market condition. Going live is a separate decision with a different downside, and Indian equity additionally requires per-strategy broker approval. See Paper vs live
If something did not work
| Symptom | Cause | Fix |
|---|---|---|
| It asked for something you thought you gave | The value did not parse as stated | Restate it on its own: "stop-loss 1.2%" |
| "I could not read near the 20 EMA" | No threshold in the phrase | Give it a number: "within 0.5% of the 20 EMA" |
| A question about what "unless" applies to | Ambiguous scope, never guessed | Restate positively |
| Zero trades in the backtest | The entry is never true | Loosen one clause; check the trading window |
| "Backtests are supported from 01 Jan 2024" | Window floor | Pick a later start |
| A safety boundary instead of an answer | You asked what to trade | The platform will not answer that, see safety boundary |
Next
- Creating with AI. The builder in full
- Strategy anatomy. Every field explained
- Indicators, the full vocabulary
- How risk is enforced