TL;DR: Slippage is the gap between the price your algo decided on and the price you actually got. Measure it on every fill — signal price, order price, fill price — in points and in rupees, per strategy and per order type, and look at the bad tail (the 90th percentile), not just the average. For most retail Indian option algos it is the biggest cost after statutory charges, it never appears on a contract note, and it is the main reason live results trail the backtest.
Last updated: 15 September 2026 · 10 min read
By Pulkit Mangal — F&O trader since 2017, founder of TradeDiary. Traded across Zerodha, Kotak and Dhan, ran rule-based strategies through broker APIs from 2020; built TradeDiary after losing ₹14L to positions I never reviewed properly.
The backtest sold the NIFTY straddle at the close of the 9:20 candle, every expiry, and bought it back at the close of the exit candle. The live algo sold it at whatever the bid was when the order reached the exchange a second later — about 1.5 points worse on each leg — and bought it back a point worse on each leg. At 300 units a leg that is ₹1,500 per round trip. Over 180 trades it is ₹2,70,000 that the backtest never showed and no contract note ever itemised.
That is algo trading slippage. It is not a bug and it is not bad luck; it is the price of turning a signal into a fill, and every live strategy pays it. The difference between traders whose algos survive and traders whose algos quietly bleed is mostly whether anyone measured it.
This post defines slippage precisely, gives you a seven-field log to capture it, a step-by-step way to measure it per strategy, a worked example with real-looking numbers, and the India-specific reasons it is larger than most retail traders assume.
💡 Short on time? You cannot measure slippage without a clean fill-by-fill ledger. Start a free TradeDiary account, import your broker's tradebook, tag each trade with its strategy, and you have the fills, times and charges in one place to set against your algo's signals.
What "algo trading slippage" actually means
Slippage is the executed price minus the price your strategy expected, signed so that a positive number is a cost. For a buy, it is fill price minus expected price; for a sell, it is expected price minus fill price. The expected price is the one your logic used to decide — the signal price (institutional desks call the whole gap implementation shortfall, measured from the arrival price).
Three prices describe every order, and the gaps between them are two different kinds of slippage:
- Signal price → order price: delay slippage. The market moved between the moment your condition was true and the moment the order was sent — candle-close logic, polling intervals, API round trips and broker risk checks all add delay.
- Order price → fill price: execution slippage. The spread you crossed and the depth you walked through. A market order pays the full spread plus impact; a limit order pays little or nothing here but may not fill at all.
- The unfilled order. A limit or stop-limit that never fills is also slippage — the cost is the move you missed, and it usually happens exactly when the move is large.
Three India-specific gotchas make retail slippage bigger than the textbook:
- Option spreads are wider than they look. The tick is ₹0.05, but outside the at-the-money NIFTY and BANKNIFTY strikes — far OTM strikes, stock options, next-month contracts — the bid-ask gap is often several ticks, and it widens further in the first minutes after 9:15 and in the last hour of an expiry.
- Market and SL-M orders are restricted on options by many brokers. Stops on option legs are usually SL-limit orders with a buffer. If price gaps through the buffer, the stop does not fill — the worst slippage of all, because the position stays open.
- Large orders are sliced. Orders above the exchange's freeze quantity are split into slices, and each slice walks further into the book. The average fill of a big order is worse than the first slice suggests.
Why slippage matters for Indian traders specifically
It is the cost nobody reports
Your contract note itemises brokerage, STT, exchange charges, SEBI fees, stamp duty and GST. Slippage is not on it — it is already inside the traded price. A trader who reconciles charges to the rupee can still have no idea that execution is costing more than brokerage.
Retail option edges are thin
SEBI's September 2024 study found that 93% of individual F&O traders lost money over FY22–FY24, with average losses of around ₹2 lakh per trader including transaction costs. Most retail option strategies that do work make a few points per trade. A point and a half of slippage per leg does not dent an edge like that — it replaces it.
Backtests fill at prices you cannot get
Most retail backtests fill at the candle's close or open, with no spread, no delay and no partial fills. That is fine for testing an idea and useless for sizing money. We covered the broader gap in backtesting vs journaling; slippage is the largest single line in it.
Retail latency is measured in hundreds of milliseconds
A retail algo cannot co-locate. Every order goes out over the internet, through the broker's API and risk checks, to the exchange — and under SEBI's 2025 retail algo framework, through a registered, static-IP route. None of that is a problem for a strategy that holds for an hour; all of it is a problem for one that trades the first minute of an expiry.
It compounds with frequency
Slippage is paid on every leg, entry and exit. An algo that trades twice a day pays it a few hundred times a year; one that re-enters on every signal pays it thousands of times. Frequency is where a small per-trade cost becomes the whole P&L — the same arithmetic as cost drag in an algo performance review.
The framework: seven fields to log on every fill
Slippage cannot be reconstructed after the fact from a P&L statement. It needs these fields captured per order — most from your algo's own log, the rest from the broker's tradebook.
| Field | Why it matters |
|---|---|
| Signal price and time | The price your logic acted on. The broker cannot give you this; only your algo can log it. Without it, you can only measure execution slippage, never delay. |
| Order type and limit price | Market, limit, marketable-limit or SL-limit. Slippage differs by an order of magnitude between types, so it has to be measured per type. |
| Order sent time | Separates delay (signal → sent) from exchange-side execution (sent → filled). |
| Average fill price and fill time | From the tradebook. Use the quantity-weighted average across partial fills and slices. |
| Ordered vs filled quantity | Partial fills and rejected slices are a slippage cost too — the unfilled part either misses the move or is chased at a worse price. |
| Bid-ask spread at the signal | Tells you whether a bad fill was the market (wide spread) or your order (market order into a thin book). |
| Slippage in points, ₹ and % of premium | The computed result. Points compare instruments, rupees compare against charges and P&L, % of premium shows how much of an option's value you give away. |
Optional fields once the basics are habitual: a liquidity bucket (ATM index / OTM index / stock option), a time bucket (open, midday, last hour, expiry last hour), the number of order modifications, and rejections with their reason codes.
How to measure algo slippage, step by step
- Make the algo log the signal price. One line per decision: timestamp, instrument, side, quantity, the price the condition fired on, and the order tag you sent. This is the only step that needs code, and nothing else works without it.
- Export the tradebook, not the P&L statement. The tradebook has every fill with time and price. Import it into your journal so each fill is a row with its charges.
- Match fills to signals. Use the order tag (Kite Connect and most broker APIs accept one) or the exchange order ID. Anything that will not match is a finding in itself — usually a retry, a modification or a manual intervention.
- Compute signed slippage per fill. Buy: fill − signal. Sell: signal − fill. Multiply points by filled quantity for rupees; divide by the option premium for a percentage.
- Split delay from execution. Delay = order price − signal (signed the same way); execution = fill − order price. If delay dominates, your logic or infrastructure is slow; if execution dominates, your order type or instrument choice is the problem.
- Aggregate per strategy, order type, time bucket and liquidity bucket — median and 90th percentile. The average hides the tail; the damage is in the tail. Put the numbers next to the backtest's assumption (usually zero).
- Feed it back. Add the measured median per leg as a fixed cost in the backtest and recompute expectancy — the expectancy formula and the expectancy calculator take a minute. If expectancy after slippage and charges is still positive, the strategy is real.
For how slippage fits with the other per-strategy numbers — win rate, profit factor, drawdown — see Trader Analytics 101; for option strategies specifically, options performance analysis covers the per-strategy view.
A real example with numbers
Rohan, 29, Bengaluru, a data engineer, runs a NIFTY weekly option-selling algo through his broker's API on ₹6,00,000 of capital, at 150 units a leg. The backtest promised ₹1,40,000 over the last quarter. The account made ₹61,000 after ₹27,000 of charges. The charges explain part of the gap; the slippage log he started that quarter explains the rest:
| Order type and time | Legs | Median slippage (pts) | 90th percentile (pts) | Slippage cost (₹) |
|---|---|---|---|---|
| Market orders, 9:15–9:20 | 60 | 1.8 | 4.5 | ₹19,800 |
| Market orders, rest of day | 110 | 0.6 | 1.4 | ₹11,550 |
| Marketable-limit orders | 180 | 0.15 | 0.5 | ₹5,400 |
| SL-limit stop exits | 30 | 2.5 | 9.0 | ₹15,300 |
| Total | 380 | ₹52,050 |
The arithmetic closes: ₹1,40,000 backtest − ₹27,000 charges − ₹52,050 slippage ≈ ₹61,000 live. Two rows — market orders in the first five minutes and stop exits — are 90 of the 380 legs but ₹35,100, or 67%, of the slippage. The marketable-limit orders, 180 legs, cost ₹5,400 in total. Two stops did not fill at all during a fast move and had to be closed by hand; their extra loss is inside the ₹15,300.
Rohan's decisions, written that evening: no entries before 9:20; every entry and planned exit as a marketable-limit order (limit a few ticks through the touch, cancelled and re-priced if unfilled in two seconds); stop buffers sized from the 90th-percentile slippage instead of a fixed two points; half size in the last hour of expiry. And the backtest now charges 0.5 points per leg, which cut its promised quarter to ₹98,000 — a number the live account can actually be judged against.
Common mistakes in slippage analysis
- Measuring against the price at fill time instead of the signal. That hides delay slippage entirely. Fix: log the signal price from the algo and always measure from it.
- Looking only at the average. A 0.4-point average can hide stop exits that slip 9 points. Fix: report the median and the 90th percentile per order type.
- Assuming limit orders have zero slippage. They trade price for fill risk; unfilled entries are missed trades and unfilled stops are open risk. Fix: track fill rate alongside slippage.
- Ignoring partial fills and slices. The first slice's price flatters the order. Fix: use the quantity-weighted average of every fill.
- Leaving slippage out of the backtest. Then every live comparison looks like the strategy decaying. Fix: add the measured median per leg as a cost and re-test.
- Mixing strategies and order types in one number. An account-level slippage average tells you nothing actionable. Fix: bucket by strategy, order type, time and liquidity.
How TradeDiary helps
TradeDiary gives you the ledger slippage analysis runs on. Import the tradebook from any Indian broker — every fill with its time, price and charges — tag each trade with its strategy, and keep the algo's signal price in the trade's notes or a tag. The journal analytics then show each strategy's expectancy after charges, win rate, profit factor and drawdown, so you can see what slippage is doing to each strategy rather than to the account as a whole. Zerodha users can start from the Zerodha journal guide.
Start tracking your fills free — import your tradebook and put every strategy's real, after-cost results next to its backtest.
You may also like: algo trading performance review — the seven numbers to review each strategy on, backtesting vs journaling — why the backtest and the account disagree, the trading expectancy formula, and expiry-day trading analytics for the session where slippage is worst.
Frequently asked questions
What is slippage in algo trading? Slippage is the difference between the price your algorithm decided to trade at — the signal price — and the average price you were actually filled at, counted as a cost when the fill is worse. It has two parts: delay slippage, from the market moving between signal and order, and execution slippage, from crossing the spread and walking the order book.
How much slippage is normal for NIFTY options algos? There is no universal number; it depends on the strike, the time of day, the order type and your size. Liquid at-the-money index options traded with limit orders in mid-session often slip a tick or two; far out-of-the-money strikes, stock options, the first minutes after 9:15 and the last hour of an expiry can slip several points. Measure your own median and 90th percentile per order type rather than borrowing someone else's.
Do limit orders eliminate slippage? They largely remove execution slippage but replace it with fill risk. A limit order that does not fill misses the trade, and a stop-limit that does not fill leaves the position open in a fast market. Track the fill rate of your limit orders alongside the slippage of the ones that did fill.
How do I add slippage to my backtest? Measure the median slippage per leg from your live fills — ideally per order type and time bucket — and subtract it from every simulated entry and exit, on top of brokerage and statutory charges. Then recompute expectancy. If the strategy is still profitable after that, the backtest is finally a fair benchmark for the live account.
Is slippage shown separately in my P&L or for tax? No. Slippage is already inside the traded prices, so it reduces your P&L automatically and appears nowhere as a separate line — unlike brokerage and statutory charges, which the contract note itemises. For tax, F&O and intraday results are business income under Section 43(5), computed on the actual traded prices, so slippage is reflected there without any adjustment.
This article is for educational purposes only and is not investment advice. Trading in equities and derivatives carries a risk of loss, including loss exceeding your capital in leveraged positions. Past performance and backtested results do not guarantee future results. Consult a SEBI-registered advisor before making investment decisions.
Pulkit Mangal writes about trading process and journaling for Indian traders at TradeDiary.
Last updated: 15 Sep 2026.