The real cost of an intraday trade in India

Most retail backtests of Indian strategies are wrong for one boring reason: they don't charge what trading actually costs. Here is what a round trip really involves, and what happens to a strategy once you include it.

The six charges on every trade

• Brokerage — typically flat per executed order at discount brokers, so it matters most on small positions.
• STT (Securities Transaction Tax) — the big one. Rates differ by segment, and for options it is charged on premium for sells and on settlement value for exercised options, which is where people get badly surprised.
• Exchange transaction charges — a percentage of turnover, set by the exchange and different for each segment.
• GST — charged on brokerage plus exchange charges.
• SEBI turnover fee — small, but it's there.
• Stamp duty — on the buy side, varies by segment.

Individually each looks negligible. Together, on an intraday options round trip, they routinely add up to more than a scalper's entire expected edge.

The rule of thumb we use: if your strategy's average win is under about 15 points on Bank Nifty options, costs are not a detail in your backtest — they are the result. Model them exactly or don't bother testing.

Slippage: the cost nobody puts in the spreadsheet

Charges are at least published. Slippage isn't, and it's often larger. Three sources:

• Spread. On liquid ATM strikes it might be 0.05–0.20. Move a few strikes out and it widens sharply.
• Impact. Your order moves the book. Irrelevant at 1 lot, very relevant at 20 in a thin strike.
• Latency. The gap between signal and fill. On expiry-day moves, prices can travel meaningfully in that window.

A workable default is to assume you cross the spread on both entry and exit, then add a tick. That is pessimistic on a calm day and roughly right on a fast one. If a strategy only survives with optimistic slippage, it does not survive.

A worked example

Take an intraday Bank Nifty options strategy: buy an ATM option, target 20 points, stop 10 points, roughly 250 trades a year, 45% win rate. On paper:

Gross expectancy per trade
  = (0.45 x 20) - (0.55 x 10)
  = 9.0 - 5.5
  = +3.5 points

Across 250 trades  ->  +875 points/year   ... looks fine

Now charge reality. Say all-in costs plus slippage work out to roughly 4 points per round trip on your lot size and broker:

Net expectancy per trade
  = 3.5 - 4.0
  = -0.5 points

Across 250 trades  ->  -125 points/year   ... it's a losing system

Nothing about the strategy changed. The only difference is that the second version is true. This flip — from modestly profitable to modestly losing — is the single most common finding when we re-test a strategy a client believed in.

What this means for how you build strategies

Costs are effectively a fixed toll per trade, so they punish frequency. Three consequences worth internalising:

1. Higher timeframes forgive more. A strategy averaging 60 points barely notices a 4-point toll. One averaging 8 points is destroyed by it.
2. Trade count is a cost, not a virtue. "More opportunities" is only good if each one clears the toll.
3. Test at your real lot size. Fixed charges hurt small positions proportionally more; slippage hurts large ones more. There's a size band where a strategy works, and it may not include yours.

Getting your own numbers

Rates change, and they differ by broker and segment, so don't trust a figure you read in a blog — including this one. Use your broker's official brokerage calculator, run your actual instrument and size through it, and take the all-in number from your own contract notes. Then put that into the backtest.

Every backtest we deliver models these explicitly and states the assumptions on the first page, so you can challenge them rather than take them on trust.