Slippage in options and futures backtests: what to assume
Why backtests fill at the next candle's open, how tick and percentage slippage change NIFTY option and gold futures results, and how to pick a value to test.
- By
- Veloce team
- Published
- Reading time
- 9 min read
On this page
- Where slippage comes from
- Why fills happen at the next candle’s open
- How Veloce enforces it
- Two ways to model slippage
- Ticks
- A share of the price
- What slippage does to real trades
- Gold futures: small targets meet fixed costs
- How to choose a slippage value
- Special cases
- What a slippage allowance cannot capture
- Try it in Veloce
Slippage is the gap between the price a backtest assumes and the price you would really get. For NIFTY and SENSEX options it comes mostly from the bid-ask spread. For MCX gold futures it also comes from how far the price moves between your decision and your fill. There is no official figure to assume. A sound starting point is fill at the next candle’s open, because that is the first price that exists after your strategy has seen a candle’s close. Then add at least one tick per fill and test two or three times that. On a NIFTY option, one tick is ₹0.05 per unit, or ₹3.25 per lot of 65. This article shows why the next-open rule matters, how tick and percentage slippage work out on real trades, and how to choose a value.
Where slippage comes from
Three things separate a model price from a real fill:
- The bid-ask spread. A market buy pays the ask and a market sell receives the bid. Candle prices are trades, which happen at either side, so a candle’s open is not a price you can always get. Crossing half the spread per fill is the minimum realistic cost.
- Latency. The market keeps moving while your system reads a candle, decides and sends the order. On one-minute data, that is the gap from one candle’s close to the next candle’s open, and sometimes more.
- Impact. A large order eats through several price levels. For one to a few lots of NIFTY near the money this is usually small; for far out-of-the-money strikes or far-month gold contracts it can be the largest of the three.
A backtest can model the second with its fill rule and approximate the first and third with a slippage allowance. That is what the rest of this article is about.
Why fills happen at the next candle’s open
A strategy that uses a candle’s close to decide can only trade after that candle has closed. There are three common ways a backtest fills that order. Only one is safe:
| Fill rule | Price | Problem |
|---|---|---|
| Same candle’s open | The open of the candle you used to decide | Lookahead bias: you traded at 09:20 on information from 09:21 |
| Same candle’s close | The exact price you decided on | Assumes zero latency and no spread; optimistic in fast markets |
| Next candle’s open | The first price after the decision | Realistic timing; still needs slippage for the spread |
How Veloce enforces it
Veloce’s engine processes the market one moment at a time. Every candle that starts at the same moment, for the index and for every option or future the strategy holds, goes to the simulated broker before any of them reaches the strategy. The broker fills pending orders at those candles’ opens. Only then does the strategy see the candles’ closes and decide. An order placed at 09:21 can therefore only fill at a 09:21 open or later, and never at a price from a candle the strategy has already seen or one still in progress.
Two consequences are worth knowing:
- Exits work the same way. A stop loss that triggers on a candle’s close fills at the next open. If the market gaps, the fill is worse than the stop level. The exits article has an example where a ₹1,300 stop fills at a ₹1,820 loss.
- Illiquid contracts fill late. The broker fills an order on the next candle of that contract. If a far out-of-the-money option does not trade for seven minutes, the order waits seven minutes and fills at whatever price it opens at. Veloce reports these: the summary lists how many fills waited for a candle after their order and the longest wait. That is a real effect in live trading too: you cannot trade a contract that nobody is quoting.
Two ways to model slippage
On top of the next-open fill, Veloce can make every fill worse by a fixed number of ticks or by a share of the price. Buys fill higher and sells fill lower. Results are rounded to the contract’s tick against you: buys round up, sells round down. A sell can never fill below one tick.
Ticks
Each fill moves a fixed number of ticks against you. Index options tick in ₹0.05 and MCX gold in ₹1 per 10 grams.
| Contract | Tick | 1 tick per fill, per lot | Round trip (2 fills) |
|---|---|---|---|
| NIFTY option, 65 units | ₹0.05 | ₹3.25 | ₹6.50 |
| SENSEX option, 20 units | ₹0.05 | ₹1.00 | ₹2.00 |
| GOLD MINI, 10× the quote | ₹1 | ₹10 | ₹20 |
| GOLD, 100× the quote | ₹1 | ₹100 | ₹200 |
Ticks are the right model when the spread is a few ticks wide whatever the price, which is true of futures and of liquid, near-the-money options. Veloce’s default is one tick per fill.
A share of the price
Each fill moves a fraction of the price against you, rounded to the tick. This model fits options across a wide range of premiums, where expensive options have wider spreads in rupees.
Here is what each model does to a single open price, using Veloce’s own rounding:
| Open price | Model | Buy fills at | Sell fills at |
|---|---|---|---|
| NIFTY option ₹120.00 | 1 tick | 120.05 | 119.95 |
| 0.1% | 120.15 | 119.85 | |
| 0.5% | 120.60 | 119.40 | |
| 1% | 121.20 | 118.80 | |
| NIFTY option ₹0.10 | 1 tick | 0.15 | 0.05 |
| 1% | 0.15 | 0.05 | |
| GOLD MINI ₹1,28,000 | 1 tick | 1,28,001 | 1,27,999 |
| 0.1% | 1,28,128 | 1,27,872 |
Two rows deserve attention:
- Cheap options. On a ₹0.10 option, 1% is a tenth of a paisa, but rounding to the tick makes it a full ₹0.05, which is 50% of the price. That is roughly right in reality: a far out-of-the-money option quoted 0.05 bid and 0.15 ask really does cost that much to trade. Strategies that buy cheap wings for protection pay more for them than the premium suggests.
- Percentages on futures. 0.1% sounds small, but on GOLD MINI it is ₹128 per 10 grams, 128 ticks, or ₹1,280 per lot per fill. Gold’s spread is nowhere near that wide in normal trading. Use ticks for futures.
What slippage does to real trades
Take two option trades from the charges article. The first sells one NIFTY call at ₹120 and buys it back at ₹80, for ₹2,600 gross and ₹64.52 of charges. The second is an iron condor with eight fills, for ₹2,340 gross and ₹208.24 of charges.
| Slippage per fill | Single leg: slippage | Single leg: share of gross | Condor: slippage | Condor: share of gross |
|---|---|---|---|---|
| none | ₹0.00 | 0.0% | ₹0.00 | 0.0% |
| 1 tick | ₹6.50 | 0.3% | ₹26.00 | 1.1% |
| 2 ticks | ₹13.00 | 0.5% | ₹52.00 | 2.2% |
| 0.5% of price | ₹65.00 | 2.5% | ₹81.25 | 3.5% |
| 1% of price | ₹130.00 | 5.0% | ₹152.75 | 6.5% |
At 0.5% of price, slippage on the single leg costs about as much as all the charges put together. On the condor, 1% of price plus charges takes ₹360.99, or 15.4% of the gross. The condor suffers more because its eight fills include cheap wings, where rounding to the tick makes a percentage worse.
Gold futures: small targets meet fixed costs
Futures make the point even more sharply, because charges apply to the full contract value. Buy one GOLD MINI lot at ₹1,28,000 and sell it ₹30 higher, a move of about 0.02%:
| Amount (₹) | |
|---|---|
| Gross P&L, 30 × 10 | 300.00 |
| Charges (brokerage example ₹20 per order) | 267.30 |
| Net before slippage | 32.70 |
| Net with 1 tick per fill (₹20) | 12.70 |
| Net with 2 ticks per fill (₹40) | −7.30 |
| Net with 3 ticks per fill (₹60) | −27.30 |
A ₹30 target is a losing strategy at two ticks of slippage, even though every trade “wins” on price. Widen the target to ₹100 and the same costs leave ₹712.62 at one tick and ₹672.62 at three. The lesson carries over to options: the smaller the move a strategy tries to capture, the more its result depends on the slippage assumption. The gold futures article breaks the charges down line by line.
The real question is not the cost of one trade, but the cost relative to the average trade. If a strategy’s average net trade is ₹200 and slippage plus charges is ₹100, a small change in either assumption decides whether the strategy works at all.
How to choose a slippage value
There is no exchange-published slippage number, and the figures in circulation are rules of thumb. One Indian backtesting vendor suggests about 0.5% on sells and 1% on buys for liquid NIFTY options. A US options data provider models slippage as a share of the bid-ask width, starting at 75% of the width for single legs. Treat both as starting points, not facts. A more defensible process:
- Look at the spread you would actually cross. Note the bid and ask of the strikes your strategy trades, at the times it trades: the open, mid-session, and expiry-day afternoons. Half the spread is the floor for a market order.
- Pick the model that fits the spread’s shape. Ticks if the spread is a few ticks wide at any premium; a percentage if it scales with premium.
- Run the strategy at one, two and three times that value. If the ranking of parameter sets changes between them, the strategy’s edge is about the same size as its costs.
- Find the break-even. Increase slippage until net P&L reaches zero. A strategy that survives three ticks per fill is robust. One that dies at one tick was never tradable.
- Check the late fills. If many orders waited minutes for a candle, the strategy is trading contracts that are too illiquid for its timing. Move closer to the money or use a longer candle.
The engine page describes fill handling in full, and a sweep can vary slippage alongside strategy parameters.
Special cases
- The first minutes of the session have the widest spreads. A strategy that enters at 09:15 on NSE or 09:00 on MCX deserves a larger allowance than one that enters at 10:30.
- Expiry-day afternoons move fast in near-the-money weeklies. Exits there slip more than entries earlier in the week.
- Multi-leg entries are not atomic. In a backtest every leg fills at its own next open. Live, legs fill one after another, and the market can move between them. Allow an extra tick on the second and later legs if you want to be conservative.
What a slippage allowance cannot capture
A slippage allowance makes every fill a little worse. It cannot reproduce events where the order does not fill as placed:
- Partial fills. A backtest fills the whole order at one price. Live, a larger order can fill in pieces at several prices, or only partly.
- Rejections and freeze limits. Exchanges cap the quantity of a single order in index options, and brokers reject orders that breach their own checks. Large orders have to be split into several, each of which crosses the spread.
- Price bands. Options and futures have daily price limits. In a limit move, there may be no fill at all until the band is relaxed.
For the one to a few lots most individual strategies trade, these rarely bind. If you size up, check them before trusting a backtest.
Try it in Veloce
Run any strategy with no slippage, then with one, two and three ticks per fill, and compare net P&L and drawdown. Every fill in the trade list shows the price it got and its charges. Create a free account and start from the strategy library.
Frequently asked questions
- What is slippage in trading?
- The difference between the price your model assumes and the price you actually get. It comes from the bid-ask spread, from your order moving the price, and from the market moving between your decision and your fill.
- How much slippage should I assume for NIFTY options?
- There is no official figure. Start from the bid-ask spread of the strikes you trade at the times you trade, assume you pay about half of it per fill in calm markets, and test with two or three times that. Veloce's default is one tick, Rs 0.05 per unit, on every fill.
- Why do backtests fill at the next candle's open?
- Because a strategy that decides on a candle's close cannot trade at a price it has already seen. The first price available after the close is the next candle's open. Filling earlier, at the same candle's open or close, uses information you did not have when the order was placed.
- Is percentage slippage or tick slippage better?
- Ticks suit futures and liquid options, where the spread is a few ticks wide regardless of price. A percentage suits options across a wide range of premiums. Avoid percentages on futures: 0.1% of a gold price of Rs 1,28,000 is 128 ticks.
- What is lookahead bias in backtesting?
- Using information in a decision that would not have been available at that moment, such as a candle's close to trade at its open. It makes backtests look better than any live result can be.
Sources
- NSE: equity derivatives contract specifications (index options price step Rs 0.05)nseindia.com, accessed
- MCX: GOLD MINI contract specificationsmcxindia.com, accessed
- Algotest: what is the impact of slippage on an algo?algotest.in, accessed
- TradingQnA: how much slippage should we set up for backtesting NIFTY futures?tradingqna.com, accessed
- ORATS: backtesting methodology (slippage as a share of the bid-ask width)orats.com, accessed