Best Options Backtesting Software in India 2026

Two platforms can run the exact same NIFTY iron condor definition and hand you two completely different win rates. One used captured option-chain data with per-leg slippage and the full Indian charge stack. The other filled every leg at the last traded price with a single flat cost. Only one of those numbers is worth acting on, and the screen does not tell you which.

So this guide does two things at once. It ranks the options backtesting platforms Indian traders actually use, with pricing read off each vendor's own page in October 2026 — no estimates, no stale figures carried over from last year. And it gives you the checklist to test any of them yourself, including ours, because a ranked list is only as good as the method behind it.

We have traded NIFTY and SENSEX option-selling strategies since 2006. Most of what follows exists because a backtest once told us something the live market later disagreed with.

The Short Answer

This guide is the purchase decision — which platform to buy. If you want the method instead, see how to backtest an option-selling strategy for the end-to-end sequence, and NSE historical options data for what the dataset underneath it has to contain.

There is no single best options backtesting software in India, because "best" depends on whether you need depth, cheapness, unlimited runs, or a path to live execution. Here is the honest split:

• Best all-round options backtester: AlgoTest — deepest options-specific workflow, transparent credit pricing, and 25 free backtests every week that never expire as an offer.
• Best value for heavy testing: Optest by Definedge — unlimited backtesting at ₹499 + 18% GST a month, on 5+ years of data.
• Cheapest paid entry: uTrade Algos at ₹99 + GST a month, and Tradetron's free-for-life plan if you want to pay nothing at all.
• Best for heavy, high-volume testing — and the deepest multi-leg structures: mt Quant — multiple backtests running at the same time across different scripts, up to 24 legs per portfolio, and strike selection by Greeks or nearest premium. It publishes no price, so judge it on capability rather than on cost per run.
• Best for coders: AlgoBulls' Python route, or building your own on open-source libraries.
• Not a backtester at all, despite how often it appears on these lists: Sensibull. It is an excellent options analytics and strategy-building platform. It does not run historical backtests, and its own site does not claim to.

Whichever you pick, run it through the 12-point checklist and the data accuracy test before you size a position on its output.

---

How We Evaluated These Platforms (Our Methodology)

Most "best backtesting software" lists are assembled from vendor marketing pages. This one was assembled the other way round: we started from the questions that break a backtest, then checked which platforms publish an answer.

Here is exactly what we did, and — more importantly — exactly what we did not do, so you can judge how much weight to give each verdict.

What we checked, platform by platform

1. Published pricing, read off the vendor's own page. Every rupee figure in this guide comes from the vendor's own pricing page or official documentation, read in October 2026. Where a vendor hides pricing behind a login or a sales call, we say so instead of guessing.
2. Data provenance. Does the platform state where its historical options prices come from — captured exchange trades, or a theoretical price series generated from spot? This single answer moves backtest results more than any other variable.
3. Data granularity and depth. Tick, one-second, or one-minute. And from which year, because a strategy tested only against the last two years has never met a volatility shock.
4. Instrument coverage. NIFTY, BANKNIFTY, FINNIFTY, MIDCPNIFTY, SENSEX, stock options, MCX. SENSEX coverage in particular separates the current platforms from the stale ones.
5. Multi-leg depth. How many legs per strategy, and whether each leg can carry its own stop-loss, target, trailing stop and re-entry rule.
6. Cost and slippage modelling. Whether slippage is configurable per leg, and whether the cost model includes the full Indian charge stack rather than a single flat brokerage number.
7. Expiry-day mechanics. Correct expiry weekday per index, correct lot size per contract period, and settlement handled against the exchange's closing-window calculation rather than the last traded price.
8. Validation tooling. Whether you can split a window into training and validation data, or whether the platform only hands you one headline number over one window.
9. Auditability. Whether the output can be inspected per day and per leg, and exported.
10. Path to live. Whether a backtested strategy can be deployed, and through which brokers — because a strategy you cannot execute is an academic exercise.

What we did NOT do, and why it matters

We did not run a controlled head-to-head where the same strategy definition is executed on every platform and the outputs are compared line by line. That test is the only one that would settle the data-quality question definitively, and we are not going to pretend we ran it.

We also did not buy every paid tier of every platform. Where a feature sits behind a tier we did not purchase, this guide reports what the vendor documents and labels it as the vendor's claim, not as something we observed.

And we have no affiliate arrangement, reseller agreement or commercial relationship with any platform named here. EliteAlgo's own engine is in the comparison. That is a conflict of interest. So we have put it in a clearly marked section with its limits stated, rather than handing it a ranked slot it has not earned in a controlled test.

What we can bring that no vendor comparison can is the other half of the problem: what a backtest number does and does not predict once the strategy is live. That comes from 163 real NIFTY and SENSEX expiry sessions of our own option-chain data, and it has its own section below.

---

All 11 Platforms at a Glance

Every price below was read off the vendor's own live pricing page in October 2026. Where a vendor does not publish a price, this table says so rather than guessing.

#	Platform	Entry price (published)	Free tier	Options-specific depth	Live deployment	Best for
1	AlgoTest	₹499 for 500 credits (+18% GST), credits never expire	25 backtests/week	High — intraday + EOD, leg-level, adjustments, re-entries	Yes	Serious options traders wanting one test-to-deploy workflow
2	Optest (Definedge)	₹499 + 18% GST / 30 days	No free tier published	High — unlimited runs, intraday, positional and custom, 5+ years data	No	Heavy testers who want unlimited runs for a flat fee
3	Stockmock	₹449 + 18% GST / 7 days (backtesting plan)	Not published on the pricing page	High — expiry-cycle focus across six indices	No	Fast visual expiry-cycle backtests
4	Quantman	₹1,300 / month excl. GST	No free tier published	High — rule-based, 6+ years of data, adjustments	Yes	No-code systematic rules with a live path
5	mt Quant	Not published	Not published	Very high on structure — up to 24 legs, Greeks or nearest-premium strike selection	Platform-linked	Heavy, high-volume testing and complex multi-leg portfolios
6	Tradetron	₹0 free for life; paid from ₹250 / month	Yes — backtesting plus paper trading, free for life	Medium — depends heavily on how the strategy is authored	Yes	Paying nothing to start, and the widest automation reach
7	uTrade Algos	₹99 + 18% GST / month	First month free on one tier	Medium — templates and an AI builder; backtest window capped	Yes	The cheapest paid entry point
8	AlgoBulls	₹299 / week (Starter Phoenix)	No free tier published	Low on the no-code tier, unlimited if you code it	Yes	Python developers who want India and US markets
9	Streak	Free	Yes — the whole platform	Medium — indicator-first, options strategies supported	Yes	Free no-code testing, if you bank with the right broker
10	SpeedBot	USD 30 / month publicly; INR not published	One backtest per month	Medium	Yes	Cloud-hosted bots, if you can get an INR quote
11	AmiBroker + DIY Python	One-off licence, or free libraries plus paid data	Open-source libraries are free	Whatever you build	You build it	Researchers who want no ceiling and accept the work

Two things worth saying about this table before you use it. First, "options-specific depth" is our read against the 12-point checklist, not a vendor rating. Second, EliteAlgo's own engine is deliberately not in it. We built it, so we cannot rank it against these without the comparison being worthless. It has its own section further down, with its limits stated.

---

The 11 Platforms, Reviewed

1. AlgoTest — the most complete options backtesting workflow

What it is. An options-first platform built around backtesting, forward testing and live algo execution for Indian index and stock options. Of everything on this list, it is the one most Indian options traders meet first, and that is largely deserved.

Pricing, exactly as published. AlgoTest does not sell a monthly subscription. It sells credit packs, and you spend credits on activities. The packs, read off its live pricing page in October 2026, all exclusive of 18% GST, with credits that carry no expiry:

Credit pack	Price	Credits	Effective cost per credit
Starter	₹499	500	₹1.00
Explorer	₹999	1,100	₹0.91
Pro	₹2,499	3,000	₹0.83
Advanced	₹5,999	7,500	₹0.80
Ultimate	₹14,999	20,000	₹0.75

What the credits buy, per its own documentation: 100 limited-mode backtests for 100 credits with no expiry; unlimited backtesting for 7 days; unlimited backtesting for 30 days at 1,599 credits; one live algo execution for 100 credits over 28 days, which works out at about ₹4 a day or less; and one forward test for 50 credits over 28 days, about ₹2 a day. Algo execution and forward testing are priced pro rata, so a strategy started on day 28 of a cycle costs a fraction of one started on day 1.

Note one inconsistency we found and could not resolve: AlgoTest's own documentation quotes the 7-day unlimited backtest pack at 499 credits in one place and 599 in another. Check the figure in-app before you buy.

The free tier is the real story. Every user gets 25 free backtests a week, resetting on Monday. They do not accumulate, so use them or lose them — but they are a standing allowance, not a trial. That is enough to properly evaluate a strategy idea without paying anything, and it is the most generous permanent free allowance of any dedicated options backtester here.

There is also a genuinely free path worth knowing about: AlgoTest publishes zero-cost access to its full feature set — backtesting, forward testing and live execution — for BTC and ETH on one crypto exchange. Irrelevant if you only trade NIFTY, but it is a real way to learn the platform's mechanics for nothing.

Pros

• Options-native rather than an equity backtester with options bolted on — leg-level logic, adjustments and re-entries are first-class, not workarounds.
• 25 free backtests a week, permanently, is the best free allowance in the category.
• Credits never expire, so occasional testers are not forced onto a monthly treadmill.
• Covers both intraday and end-of-day backtesting, across indices and stock options.
• One continuous path from backtest to forward test to live, so you are not re-implementing the strategy in a second tool.

Cons

• The credit model makes real monthly cost hard to predict. Heavy testing plus two live strategies is a very different bill from light testing, and you only find out by running it.
• Its own documentation contradicts itself on one pack price, which is a small but real transparency wobble.
• Published documentation does not state the historical data depth or the data capture method — the two things that matter most. You have to ask.
• Indian-market focused, so it is not the tool if you also want commodity or currency options.

Verdict. If you are backtesting Indian index options seriously and want one tool for the whole lifecycle, start here. Use the 25 weekly free backtests to decide before you spend a rupee.

2. Optest by Definedge — the best value if you test a lot

What it is. Definedge's dedicated options backtesting product, sold separately from Opstra, its better-known options analytics suite. This distinction matters, and most comparison articles miss it entirely — they list Opstra and describe its limited backtesting, without mentioning that Definedge sells a purpose-built backtester for less.

Pricing, exactly as published. Read off its own pricing page in October 2026, under a stated introductory offer:

Plan	Price	Access	What's included
Monthly	₹499 + 18% GST	30 days	Unlimited backtesting, all backtesting methods, intraday + positional + custom, filters and trade navigation, 5+ years of data
Annual	₹4,999 + 18% GST (listed ₹5,999, 16.5% off)	365 days	Same as monthly
Opstra + Optest bundle	₹7,999 + 18% GST (listed ₹11,000, 27% off)	—	Everything in Optest plus everything in Opstra

Pros

• Unlimited backtesting for a flat fee. On a credit-based rival, the research phase of a single strategy can burn through a whole pack. Here the marginal cost of the four-hundredth run is zero, which changes how you work — you stop rationing experiments.
• 5+ years of data stated up front, which is deeper than several credit-based platforms allow you to reach at any price.
• Intraday, positional and custom backtesting in the same plan, with no tier gating.
• Clearly the cheapest route to heavy testing: at ₹4,999 a year it costs less than four months of the cheapest no-code rival with a live path.
• The bundle gives you analytics and backtesting in one place at a stated 27% discount.

Cons

• No live deployment. It is a backtester, full stop. You validate here and execute somewhere else, which means re-implementing the strategy in a second system — the exact point where backtest and live behaviour quietly diverge.
• The prices are explicitly labelled an introductory offer, so they are not guaranteed to hold.
• Data granularity and capture method are not published. "5+ years" tells you depth, not fidelity.
• No free tier is published, so you pay before you can judge it.

Verdict. The best rupees-per-backtest on this list by a wide margin. If your bottleneck is how many experiments you can afford to run, this is the answer — just plan for executing elsewhere.

3. Stockmock — fast, visual, expiry-cycle backtests

What it is. An options backtesting and simulation platform built around the expiry cycle, covering NIFTY, BANKNIFTY, FINNIFTY, MIDCPNIFTY, SENSEX and BANKEX. Strategy building is genuinely quick: index, futures or options, call or put, buy or sell, strike, premium range, lots, stop-loss and target, entry and exit times. Results can be grouped by expiry, by day, by date, or filtered by days — which is exactly the slicing an expiry-day trader wants and most general-purpose tools make awkward.

Pricing, exactly as published — and read this carefully. Its own pricing page sells two separate plan families, and neither includes the other.

Family	Term	Price (+18% GST)	Effective rate
Analytics Simulator + Builder + Charts + Option Chain	30 days	₹349	₹349/month
	180 days	₹1,599	₹267/month
	360 days	₹2,599	₹217/month
Backtesting Multi-year single and multiple backtests	7 days	₹449	₹449/week
	28 days	₹1,299	₹325/week
	84 days	₹2,999	₹250/week

The pricing page states this plainly: the Analytics plans do not include multi-year single-click backtesting, and the Backtesting plans do not include the Simulator, Builder or Charts. So the realistic cost of using Stockmock as both an analysis and a backtesting tool for a year is the sum of two plans, not one — roughly ₹2,599 plus ₹2,999 across comparable terms, before GST. Comparison articles that quote "from ₹349" are quoting the half that does not backtest.

Pros

• Six index families covered, including SENSEX and BANKEX — broader Indian index coverage than most rivals at this price.
• Grouping by expiry, day and date is purpose-built for expiry-cycle analysis and genuinely fast to use.
• Strategy definition is about as frictionless as no-code gets.
• Three term lengths on each family, including a one-week backtesting pass for a short burst of research.
• F&O stock coverage in the Simulator, on its own stated beta terms.

Cons

• The two-plan split is the real cost. Buying both families is unavoidable if you want builder and backtests, and that roughly doubles the headline price.
• No live deployment. Validate here, execute elsewhere.
• API access is explicitly prohibited. Its pricing page states that plans cover website use only, that script or tool access is blocked, and that no refunds are entertained for it. If your workflow involves pulling results out programmatically, this platform is closed to you.
• No free tier is published on the pricing page.
• Data granularity and capture method are not published.

Verdict. The quickest way to see how an index-option structure behaved across many expiries. Budget for both plan families, and do not plan on automating anything around it.

4. Quantman — rule-based no-code with a real live path

What it is. A no-code backtesting and automated-trading platform from Simply Algo Fintech Pvt. Ltd., based in Coimbatore. Its own site states 6+ years of backtesting data, 150,000+ users, 40+ broker integrations and over two crore backtests run. Simply Algo also appears on the exchange's list of empanelled algo providers, which is a meaningful signal in the current regulatory environment and one most platforms on this list cannot claim.

Pricing, exactly as published. All figures excluding GST, read off its own plans page in October 2026.

Plan	Billed monthly	Billed yearly (per month)	Backtest credits	Live deployment credits
Basic	₹1,300	₹1,084	250	4
Premium	₹3,300	₹2,417	1,000	10
Pro	Yearly only	₹5,417	5,000	25

Basic and Premium include calendar spreads, strike multipliers, "add cases", adjustment features and a library of sample strategies. Pro adds options basket features and is capped at a stated 1,000 users.

Pros

• 6+ years of backtesting data is among the deepest stated depths here, and depth across regimes is what stops a backtest being a regime artefact.
• Adjustment features and calendar spreads are included from the entry tier, not gated upward.
• Live deployment credits are bundled with every tier, so backtest and execution live in one system.
• 40+ broker integrations means you are unlikely to be forced to switch accounts.
• Appearing on the exchange's empanelled-provider list is a genuine compliance signal.

Cons

• The most expensive entry point of any no-code platform here. ₹1,300 a month is over two and a half times Optest's unlimited plan, and Optest gives you unlimited runs where Basic gives you 250 credits.
• No free tier is published, so the first look costs ₹1,300.
• Pro is yearly-only, so testing the top tier means committing for a year.
• Data granularity and capture method are not published.
• The footer on its own plans page still reads 2025, which is a small freshness signal worth noting on a platform you are trusting with historical data.

Verdict. Strong if you want deep data, adjustments and live execution in one no-code system and the monthly fee is not your constraint. If budget is the constraint, it is beaten on both price and backtest volume.

5. mt Quant — the deepest multi-leg structures on this list

What it is. A browser-based strategy backtesting platform, published by mastertrust at mastertrust.co.in/mtquant as part of its masterWeb platform. It is built for backtesting, performance and risk evaluation, and strategy optimisation and comparison.

Best for: heavy, high-volume testing. It is the only platform on this page that states it runs multiple backtests at the same time across different scripts, and it also carries the highest published leg limit here at 24. If your constraint is how many variants you can put through in a session rather than what a run costs, this is the capability to look at. It publishes no price, so the cost side of that trade cannot be stated.

What it does, in its own words. These are the platform's own stated capabilities, and on structural depth they are the strongest claims on this page:

• Complex multi-leg strategies and portfolios of up to 24 legs, with a wide range of combinations.
• Multiple backtests at the same time across different scripts.
• Strike selection by Greeks or nearest premium, rather than by fixed strike or simple offset.
• Multi-level risk management, with stop-losses and targets specifiable at portfolio level, not only per leg.
• Actual, non-hypothetical stop-loss and target values — the platform's stated position is that these come from real values rather than assumed ones.
• Export to Excel or HTML for analysis outside the platform.
• A scripting mode for custom logic beyond the standard builder, alongside no-code multi-condition rule building.

Twenty-four legs is well beyond what most platforms here support, and matters for anyone running layered baskets rather than single structures. Strike selection by Greeks or nearest premium is the other standout: selecting by delta or by a premium target is how option sellers actually choose strikes, and platforms that only offer fixed-offset selection force you to approximate it.

Pros

• Up to 24 legs per portfolio — the deepest structural capacity on this list.
• Greeks-based and nearest-premium strike selection matches how sellers really pick strikes.
• Portfolio-level stop-loss and target, not just leg-level, which is how a real basket is risk-managed.
• Multiple concurrent backtests across different instruments speeds up comparative research.
• Excel and HTML export means the output is auditable outside the vendor's own report.
• Browser-based, so no Windows desktop install to maintain.

Cons

• No pricing is published anywhere on the product page. You cannot compare its cost against anything on this list without signing up, and the page's main call to action is opening a trading account.
• Data granularity, historical depth and exchange coverage are not stated on the product page. These are the three facts that most determine whether a backtest is trustworthy, and all three are absent. Widely-repeated claims about tick-level data or specific exchange coverage for this product are not supported by anything we could verify on its own page — ask the vendor directly and get it in writing.
• It sits inside one broker's platform, so your backtesting tool and your execution account are coupled.
• No free tier or trial is published.
• Its own FAQ is thin on specifics — it confirms what backtesting is and that it does not guarantee future returns, which is honest but not informative.

Verdict. On paper, the most capable structural backtester here, and 24 legs with Greeks-based strike selection is a genuinely differentiated combination. The problem is that none of the questions our checklist cares about most — data source, granularity, depth — are answered publicly, and neither is the price. Worth a hard look if you run complex baskets, but go in with the checklist and get written answers first.

6. Tradetron — the best way to start at zero cost

What it is. A strategy automation platform with a marketplace, covering NSE, NFO and BFO, plus MCX and several crypto venues. Backtesting is one module inside a system whose centre of gravity is automation and strategy distribution.

Pricing, exactly as published. Read off its own pricing page in October 2026. Note that what changes between tiers is how many live bots you can run and whether you can list on the marketplace — not whether you can backtest.

Plan	Per month	Billed yearly	Live trading bots	Marketplace listings
Free	₹0 — free for life	₹0	Live Offline only	—
Starter	₹250	₹3,000	1	—
Starter+	₹708	₹8,500	3	—
Retail	₹1,000	₹12,000	5	—
Retail+	₹2,083	₹25,000	12	—
Creator	₹4,167	₹50,000	25	5
Creator+	₹7,500	₹90,000	50	10

Yearly billing saves a stated 16%. Every plan, free one included, ships three things: unlimited strategy templates, "Live Offline", and error alerts by WhatsApp and phone call. Live Offline runs every trade and sends no orders, so it is paper trading by another name.

Worth flagging for accuracy: several current comparison articles still quote Tradetron's entry tier at ₹300 a month. It is ₹250. A small thing, but it tells you how often these lists are actually re-checked.

Pros

• Genuinely free to start, for life. You can build, backtest and paper-trade a strategy for ₹0, indefinitely. You only pay when you want real orders sent.
• ₹250 a month for the first live bot is the cheapest paid live deployment here.
• Broad segment coverage in one account: NSE, NFO, BFO, MCX and crypto.
• Paper trading that runs the full trade logic and withholds only the order is the right design for the gap between backtest and live.
• Error alerts by phone call is an unusual and genuinely useful operational feature.
• Funds stay in your own broking account — the platform never custodies money.

Cons

• Options depth depends on who wrote the strategy, not on the platform. A marketplace strategy is only as rigorous as its author, and you usually cannot inspect their assumptions.
• The credit and plan structure is built around live bot count, so a researcher who runs many backtests and few live strategies is paying for the wrong axis.
• Not purpose-built around expiry-day option mechanics, so expect more manual verification if that is your strategy class.
• Data granularity, depth and capture method are not published.

Verdict. The correct first stop if you want to spend nothing. Build it, backtest it, run it offline for an expiry cycle, and only then decide whether ₹250 a month is worth it.

7. uTrade Algos — cheapest paid entry, with a backtest ceiling

What it is. An AI-assisted algo platform from uTrade Solutions Pvt. Ltd., Mohali, built around a strategy builder, backtesting with day replay, forward testing and automated execution.

Pricing, exactly as published. All figures from its own pricing page, excluding 18% GST. Monthly credits refresh each billing cycle and do not roll over; purchased top-up credits never expire.

Plan	Per month	Credits/month	Concurrent live deployments	Support
Beginner	₹99	75	2	Standard
Essentials	₹499 (first month free)	500	5	Priority
Pro	₹1,999	2,000	10	Premium
Elite	₹4,999	5,000	20	Dedicated

This is the most transparent credit model on the list — the vendor publishes exactly what each action costs:

Action	Credit cost
Backtest, up to 30 days of history	1 credit
Backtest, 30 to 90 days	3 credits
Backtest, 90 days to 1 year	5 credits
Backtest, 1 to 3 years	10 credits
Forward test run	5 credits
AI interaction	2 credits
Live deployment, within your plan's limit	Free

Top-up packs run from ₹99 for 100 credits down to ₹3,199 for 5,000 credits, which is ₹0.64 a credit at the top pack.

Read the credit table carefully, because it contains the catch. The deepest published backtest band is "1 to 3 years" at 10 credits. There is no band beyond three years. So on the ₹99 Beginner plan — 75 credits a month — you get at most seven three-year backtests in a month, and no published way to test a longer window at any price. Set against Optest's unlimited runs on 5+ years of data for ₹499, the cheap entry price buys you a shallower research horizon, which is exactly the trade-off a cheap tier usually hides.

Pros

• ₹99 a month is the lowest paid entry here, and it still includes two concurrent live deployments.
• Live deployment consumes no credits at all — you pay for research, not for running.
• The most transparent published credit costing of any platform on this list. You can forecast your bill before you spend.
• One tier gives the first month free, so you can evaluate the real thing rather than a demo.
• Top-up credits never expire, which softens the monthly-credit reset.
• Forward testing is built in at a known 5 credits, so the paper-trading step is not an afterthought.

Cons

• The backtest horizon appears capped at three years. That is not enough regime coverage for an options strategy you intend to trade with size, and it is the single biggest limitation on this platform.
• Monthly credits do not roll over, so an unused month is money gone.
• Credits are shared across backtesting, forward testing and AI interactions, so chatting to the assistant spends your research budget.
• Data granularity and capture method are not published.
• Heavy research gets expensive fast: 2,000 credits at ₹1,999 is 200 three-year backtests, where a flat-fee rival charges ₹499 for unlimited.

Verdict. Excellent value if you want a live algo running cheaply and your research needs are modest. A poor fit if your research needs are deep — the three-year ceiling will bind long before the price does.

8. AlgoBulls — built for people who write code

What it is. A development and execution platform spanning Indian and US markets, with a Python library, Jupyter notebook support and IDE integration alongside a no-code "Classic Build" mode. AlgoBulls also appears on the exchange's empanelled algo-provider list.

Pricing, as published. AlgoBulls publishes a long feature-by-feature plan comparison. The entry tier we could read directly is Starter Phoenix, ₹299 for a week, and its documented limits are unusually specific — and unusually revealing:

Starter Phoenix (weekly, ₹299) — documented limits	Value
Strategies allowed to be saved	20
Legs allowed for an options strategy in Classic Build	1
Child legs allowed for an options strategy in Classic Build	1
Legs allowed for an equity strategy in Classic Build	2
Backtesting execution time	2,000 minutes
Historical data for backtesting	3 years
Strategies running in parallel	5
AI copilot prompts	50
Support tickets	10

Higher tiers raise these allowances. Credit to AlgoBulls for publishing leg limits at all — nobody else on this list does, and it is exactly the kind of number that should be in the open.

Pros

• 100% coding flexibility through its Python library and notebooks, so there is no ceiling on strategy logic if you can write it.
• Indian equity, F&O, commodity and currency plus US markets in one platform — the broadest market coverage here.
• Publishes hard per-tier limits including leg counts, which is more transparency than any rival offers.
• ₹299 for a week is a cheap, genuinely time-boxed way to evaluate it properly.
• Appearing on the exchange's empanelled-provider list is a real compliance signal.
• Detailed execution logs and trade analytics, with intellectual-property protection options for your strategy code.

Cons

• One option leg in Classic Build on the entry tier. A short strangle needs two. An iron condor needs four. So the no-code mode on the cheap tier cannot express the structures this guide is about — you either pay up or you write Python.
• 3 years of history on the entry tier, which is thin for regime coverage.
• Backtesting is metered in execution minutes, an unusual unit that is hard to budget against until you have run a few.
• The Python route has a real learning curve. "No ceiling" and "no effort" are not the same thing.
• Data granularity and capture method are not published.

Verdict. The right choice if you write Python and want one platform across India and the US. The wrong choice if you want no-code multi-leg options structures cheaply — the entry tier's one-leg limit rules that out.

9. Streak — free, if you bank with the right broker

What it is. A no-code systematic trading platform — scanners, a scalping module, backtesting, and both equity and options strategies built from pre-built logic, assisted conditions or fully custom rules. It is free, and has been free for all of the large discount broker's users since January 2024, when its earlier paid tiers were retired.

Pricing. Free. Its own site describes it as a free systematic trading platform, and access is through that broker's login. There is no published price to compare.

One current caveat. The platform's own pricing page shows "V 1.0.0 (In Beta Phase)" at the time of writing. A self-declared beta is not a reason to avoid it, but it is a reason not to make it the only validation step before committing capital.

Pros

• Free, with no credit metering and no plan to choose. For a large population of Indian traders this is the zero-friction starting point.
• Covers backtesting, live scanners, and both equity and options strategies in one no-code interface.
• Custom rule-based frameworks are available, not just templates.
• Deploys live through the same broker account, so there is no second system to re-implement in.

Cons

• Tied to one broker's login. If you do not have that account, this platform is not available to you at the free tier.
• Indicator-first by design, not options-structure-first. Greek-aware, delta-neutral, multi-leg selling logic is not what it was built around.
• Self-declared beta on the current version.
• Data granularity, depth and capture method are not published.

Verdict. If you already have the account, there is no reason not to use it as a first-pass filter — it costs nothing. Do not treat it as final validation for a multi-leg selling strategy.

10. SpeedBot — cloud bots, with pricing you cannot see in rupees

What it is. A cloud-hosted platform for developing custom strategies, backtesting them and running them as live or paper "bots", across Indian and US markets.

Pricing, exactly as published — and this is the problem. SpeedBot's public pricing page quotes subscriptions in US dollars only, under a stated 50% discount: Basic USD 360 a year or USD 30 a month; Starter USD 840 a year or USD 70 a month; Professional USD 2,400 a year or USD 200 a month; Ultimate USD 6,000 a year or USD 500 a month. Those tiers carry 1, 3, 10 and 35 live bots respectively. For Indian pricing, the page says only this: Indian traders can find the INR price in the app once they sign up.

There is also a wallet-credit, pay-as-you-go option. Its published terms include one free backtest per month, and live bot deployment at USD 2 per bot per day.

Pros

• Pay-as-you-go wallet credits mean you are not locked into a monthly platform fee during a research phase.
• A free trial is offered, and referral credits are available.
• Paper bots are included on every tier — three on the cheapest — so the paper-trading step is well provided for.
• Cloud-hosted, so there is no machine of yours that has to stay up during market hours.
• Covers both Indian and US markets.

Cons

• No INR price is published. You must create an account to find out what it costs in your own currency, which makes an apples-to-apples comparison impossible before you commit.
• One free backtest per month on the credit plan. Against a rival's 25 free backtests a week, that is not a usable evaluation allowance.
• Dollar list prices are heavily discounted from a higher list figure, a pattern that makes the real steady-state price unclear.
• Options-specific depth is not documented to the level this guide's checklist needs.
• Data granularity, depth and capture method are not published.

Verdict. Possibly fine, genuinely unassessable on published information. If you are interested, get the INR pricing in writing and run the checklist before you pay.

11. AmiBroker and the do-it-yourself Python route

What it is. Two versions of the same answer: build it yourself. AmiBroker is a long-established desktop analysis and backtesting platform with its own formula language, sold as a one-off licence. The open-source route means Python libraries — all free — plus a historical Indian options data set you buy separately, plus the code to glue them together.

Pricing. The libraries cost nothing. Everything else does. The real budget line is data: a multi-year, minute-level Indian option-chain data set is a paid product, and it is normally the largest single cost in this approach. AmiBroker is licensed one-off rather than by subscription, which is cheaper over many years than any monthly plan here — but it carries no Indian option-chain data, so you are buying that separately either way.

Pros

• No ceiling. Any strike-selection rule, any re-entry logic, any cost model, any validation scheme. Nothing on this list can match that.
• You control the cost and slippage model completely, which means you can be as pessimistic as you like — the single most valuable freedom in backtesting.
• You own your data and your results. No API prohibition, no credit meter, no vendor that can change its terms.
• A one-off licence or free libraries beat every subscription here on a multi-year view.
• Full auditability by construction — you wrote it, so you can inspect every assumption.

Cons

• The work is substantial and the failure modes are silent. An expiry calendar off by a day, a lot size applied across a revision, a settlement price taken as closing LTP — each produces a plausible, confident, wrong number. The failure modes of automation are rarely loud.
• Data cost is real and recurring, and good Indian option-chain history is not cheap.
• Nothing is pre-validated. Every mechanic that a vendor platform handles for you is now yours to get right.
• No live execution comes with it. You build or buy that separately too.
• The time cost is usually underestimated by a wide margin.

Verdict. The right answer if backtesting is your work and you need no ceiling. The wrong answer if you want to test an idea this month — the first honest number is months away, and the first dishonest one is hours away.

---

Our Own Engine, and Why It Is Not in the Ranking

EliteAlgo runs a backtesting engine of its own. It is not in the table above, and it should not be — we built it, so any rank we awarded it would be worth nothing to you.

Here is what it is, stated plainly so you can judge it against the same checklist. It is narrow by design: built around intraday expiry-day option selling on NIFTY and SENSEX, with multi-leg structures, per-leg stop-loss and trailing stop-loss, re-entry logic, and per-leg slippage under full India-standard cost modelling. It runs on minute-level option-chain and spot data — the same data our own desk trades on — with a train-and-validation split built into the workflow rather than bolted on. Every strategy that clears backtesting then goes through a separate strategy grading step before it is considered production-ready.

What it is not. It does not backtest equities, commodities, currencies or US markets. It is not a general-purpose platform, it has no marketplace, and it is not a self-serve product you sign up for and explore on a credit pack. If your strategy is not intraday index-option selling, the platforms ranked above will serve you better, and we would rather say so than pretend otherwise.

Our conflict of interest, stated. We have no affiliate arrangement, reseller agreement or commercial relationship with any platform named in this guide, and we earn nothing if you choose one. We do have an obvious interest in our own engine. That is precisely why it is in its own section with its limits stated, rather than holding a ranked slot it has not earned in a controlled test.

EliteAlgo is a proprietary trading desk. We publish analytical software and research — not investment advice — and we hold no SEBI or NSE registration as an investment adviser or research analyst. Nothing on this page is a recommendation to buy, sell or hold any security.

---

Platforms We Considered and Excluded, With Reasons

A ranking is only honest if it says what it left out. Every exclusion below is a category decision, not an oversight.

Platform or category	Why it is not ranked
Sensibull	It is not a historical backtester. Its own site presents it as an options strategy builder and analytics platform — 25+ pre-built structures, payoff curves, Greeks, advanced option chain, open-interest and FII/DII data, position analysis. Excellent at understanding a position you are about to take. It does not run a strategy over years of history, and does not claim to. We note that a rival's comparison list reaches the same conclusion, marking its backtesting capability as "No" — so this is not a contrarian read. Worth knowing: the company behind it holds SEBI Research Analyst registration, which is more than most platforms on this page can say.
Opstra	Ranked indirectly, through Optest. See also our alternatives round-up for option sellers. Opstra is Definedge's options analytics suite; Optest is its dedicated backtesting product, sold separately and cheaper. Lists that rank "Opstra" for backtesting and quote no price are describing the wrong product. If you want both, the bundle is the answer.
Backinzo	A serious-looking desktop options backtester — up to 24 legs, leg action types including OnStart, OnSL and OnTarget, Greeks and nearest-premium selection, dynamic hedges, order slicing, three levels of risk management, and one-click export to an execution bridge. We left it unranked for one reason: no pricing, data granularity or historical depth is published, and its own notes indicate users may need to arrange their own data through broker APIs. Promising, unassessable on public information.
Overseas platforms	MetaTrader 4 and 5, TradeStation, NinjaTrader and similar have no retail NSE or BSE index-options access from India. They appear on India "best of" lists regularly, and they should not. A platform that cannot reach the instrument you trade is not a candidate, however good it is.
Quantopian	It shut down in 2020. It still appears on currently-published Indian "best free backtesting software" lists, which tells you how much of this category is recycled rather than re-checked.
QuantConnect, Blueshift and research-platform peers	Capable quantitative research environments, but not built around Indian index-option structures, and the Indian option-chain data you would need is not the thing they are set up to give you. If you want this route, the DIY section is the honest version of it.
Investment product firms	Several names that rank for "options backtesting" are not software vendors at all — they sell managed or structured investment products. They are not comparable to a tool you run your own strategy on, so they are not in the table.
Broker platforms and APIs generally	We do not rank brokers here. We have run no expiry-day load testing on broker infrastructure, and we hold no broker relationships to promote. A broker's free API can be a perfectly good foundation for the DIY route — that is a category statement, not a recommendation of any one of them.
Decommissioned broker algo add-ons	At least one 2013-era broker-terminal algo add-on is still listed as current on live comparison pages. It no longer exists. Check that any platform you read about still has a working product page before you evaluate it.

---

What a Year Actually Costs: Real Cost of Ownership

Headline prices are not comparable, because the platforms meter completely different things. One sells credits, one sells unlimited runs, one sells live bot slots, one splits analytics and backtesting into separate purchases. So here is the same question asked of all of them: what does a year cost for a moderately active options trader who runs regular research and keeps two strategies live?

All figures below are built only from each vendor's own published prices, and all are stated before GST unless noted. Treat them as an order-of-magnitude comparison, not a quote.

Platform	Published basis	Approximate annual cost	What binds you first
Streak	Free	₹0	Needing the right broker account; options-structure depth
Tradetron	Free plan; Starter+ at ₹8,500/yr for 3 live bots	₹0 to ₹8,500	Live bot count, not research volume
uTrade Algos	₹99/mo Beginner, or ₹499/mo Essentials	₹1,188 to ₹5,988	The three-year backtest ceiling, and credits that do not roll over
AlgoTest	Credit packs; two live strategies ≈ 200 credits/28 days, plus research	₹2,600 to ₹6,000+	Credit burn — genuinely hard to predict in advance
Optest	₹4,999/yr, unlimited backtesting	₹4,999	No live deployment at all
Stockmock	Analytics ₹2,599/360 days plus Backtesting ₹2,999/84 days	₹5,598 and up (more if you need backtesting year-round)	Two separate plan families; no live deployment; no API access
AlgoBulls	₹299/week entry tier	₹15,548 at 52 weeks on the entry tier	One option leg in no-code; 3 years of data
Quantman	₹1,084/mo billed yearly (Basic)	₹13,008	250 backtest credits a month
SpeedBot	USD 360/yr published; INR not published	Cannot be stated	Not knowing the price until after you sign up
mt Quant	Not published	Cannot be stated	Not knowing the price, the data depth, or the granularity
DIY / AmiBroker	One-off licence or free libraries, plus paid data	Data cost dominates	Your own time, and silent modelling errors

Three things fall out of this table that no headline price shows you.

The cheapest entry is rarely the cheapest year. A ₹299-a-week plan looks like the bargain next to ₹1,084 a month. Over a year it costs more. Weekly pricing is a pricing strategy, not a discount.

Unlimited beats metered the moment you get serious. Research is iterative — you do not run one backtest, you run two hundred while you work out whether an idea is real. Every credit-metered platform quietly taxes exactly the behaviour that produces a good strategy. A flat unlimited fee removes that tax, which is why Optest's ₹4,999 a year is the strongest value here despite not being the cheapest number on the page.

"No live deployment" has a cost that never appears on an invoice. If you validate on one platform and execute on another, you have two implementations of the same strategy, and they will differ. That difference is the single most common reason live results depart from backtested ones, and it is not in any price table.

---

Feature Comparison: What Each Platform Will and Will Not Do

"Published" below means the vendor states it on its own site. "Not published" means exactly that — not that the feature is absent, but that you cannot confirm it without asking. We think the distinction matters more than a tick-box, because the questions that decide a backtest's honesty are the ones vendors are quietest about.

Platform	Historical depth	Granularity	Max legs	Strike selection	Live deploy	Export / API
AlgoTest	Not published	Not published	Not published	Leg-level with adjustments and re-entries	Yes	Not published
Optest	5+ years	Not published	Not published	All backtesting methods; intraday, positional, custom	No	Not published
Stockmock	"Multi-year"	Not published	Not published	Strike price or premium range	No	API use prohibited
Quantman	6+ years	Not published	Not published	Rule-based; strike multiplier, calendar spreads, adjustments	Yes	Not published
mt Quant	Not published	Not published	24	Greeks or nearest premium	Platform-linked	Excel and HTML export
Tradetron	Not published	Not published	Not published	Depends on the strategy author	Yes	API available
uTrade Algos	Capped at 1-3 years per its credit table	Not published	Not published	Templates and AI builder	Yes	Not published
AlgoBulls	3 years on the entry tier	Not published	1 option leg in no-code on the entry tier; unlimited in Python	Whatever you code	Yes	Python library, notebooks, IDE
Streak	Not published	Not published	Not published	Pre-built, assisted or custom rules	Yes	Not published
SpeedBot	Not published	Not published	Not published	Custom strategies	Yes	Not published
DIY / AmiBroker	Whatever data you buy	Whatever data you buy	Unlimited	Whatever you code	You build it	Total

Look at the granularity column. Eleven platforms, and not one publishes whether its historical option prices are tick-level, one-second or one-minute — or whether they were captured from real trades at all. That is the single most important variable in options backtesting, and it is the one the entire category is silent about. It is also why the twenty-minute data test further down is the most valuable thing on this page.

---

Free vs Paid: What You Actually Lose

You can do real work for nothing. Three genuinely free routes exist, and they are not trials:

• 25 backtests a week, permanently on the deepest options-specific platform here — enough to test one strategy idea properly every week, forever.
• A free-for-life plan with unlimited strategy templates, backtesting and full paper trading, where you only pay when you want real orders sent.
• A completely free no-code platform, if you hold the right broker account.

So "you need to pay to backtest" is false. What is true is that five specific things tend to sit behind the paywall, and each of them is a thing that changes your results rather than your convenience.

What paid tiers add	Why it changes the number, not just the experience
Research volume	Free allowances are rationed. Rationing pushes you towards confirming an idea rather than trying to break it — which is the opposite of what validation is for.
Historical depth	Cheap and free tiers cap the window. A three-year window may not contain a volatility shock, and our own VIX-band data shows premium available roughly doubles from the calmest band to the most stressed. A shallow window measures the regime, not the strategy.
Per-leg cost and slippage control	Free tiers tend to apply one flat assumption. A short ATM leg and a far-OTM hedge have nothing like the same spread, so a single number is wrong on at least one leg of every multi-leg strategy.
Structural depth	One platform here documents a one-option-leg limit in no-code mode on its entry tier. A short strangle needs two legs; an iron condor needs four. A tier that cannot express your structure cannot test it.
Out-of-sample validation	The step that separates a strategy from a curve fit is usually the first thing a free tier drops, and the only one whose absence is invisible — a fitted backtest looks better, not worse.

The honest sequence. Use free tiers to kill bad ideas quickly, which is most of what research is. Pay when an idea survives that filter and you need depth, per-leg costs and a validation split to find out whether it is real. Do not pay before you have an idea worth paying to test.

---

The 12-Point Checklist: Run It Against Any Platform, Including Ours

A feature list tells you what a platform has. This checklist tells you whether the number on its screen is worth acting on. Twelve questions, each with the answer you want and the answer that should worry you.

1. Is the historical options data captured from real trades, or reconstructed from a model?

Ask the vendor directly. Captured data comes from actual exchange trades. Reconstructed data is generated by running a spot price series through a theoretical pricing model. What you want to hear: captured exchange data, with the capture method stated. What should worry you: a vague answer about "market data", or no answer at all. Reconstructed prices look fine on quiet days and diverge badly on expiry days, when real option prices are driven by order flow and gamma positioning rather than clean theoretical value.

2. What granularity — tick, one-second, or one-minute?

One-minute data is adequate for most time-triggered strategies. It is not adequate for anything that reacts within the minute, and it will hide the intrasecond gaps that eat a far-OTM fill. What should worry you: a platform that will not state its granularity, or states it only for the index and not for the option chain.

3. How many years of data, and across how many volatility regimes?

Length alone is not the point. A five-year window that sits entirely in a low-volatility band has tested one regime five times. You want coverage of at least one high-volatility stretch and one range-bound stretch, so the edge is not a regime artefact. Our own VIX-band table above shows why: premium available at VIX below 12 is roughly half what is available above 20.

4. Is slippage configurable per leg, or one flat number for the whole strategy?

Multi-leg option strategies have wildly different liquidity across legs. Per-leg slippage configuration is the minimum requirement for anything beyond single-leg directional testing. What should worry you: a single global slippage field, or slippage expressed only in points rather than as a percentage of premium.

5. Does the cost model include the full Indian charge stack?

Brokerage, STT, exchange transaction charges, SEBI turnover fees, stamp duty and GST all apply, and the STT rate on option sales changed on 1 April 2026. A platform that nets off a single generic "brokerage" figure is understating your real cost, and understating it most on exactly the high-frequency multi-leg strategies where it matters most.

6. Is the expiry calendar current, per index?

NIFTY weekly options expire Tuesday. SENSEX weekly options expire Thursday. Bank Nifty has no weekly expiry. Check an actual backtest output and confirm the trade dates fall on the right weekday. This is a thirty-second test that catches a surprising number of stale platforms.

7. Is the lot size correct for the contract period, not just for today?

NIFTY lots are 65 units for contracts expiring from January 2026, and were 75 before that. A platform applying one lot size across a multi-year window misstates every rupee figure in the earlier part of it.

8. Is settlement handled against the exchange's closing-window calculation?

Index options settle against a weighted average of the underlying across the closing window, not the final traded price. Strategies held to expiry and backtested against closing LTP carry a systematic bias.

9. Can you build multi-leg, re-entry, and trailing stop-loss logic — per leg?

Real option-selling baskets rarely stop at "enter once, hold to expiry". Re-entry after a stop and trailing stop adjustment are core to how they are managed. If the platform only supports static entry and exit, it cannot honestly backtest the strategy you will actually trade. Ask specifically whether each leg can carry its own stop, target, trail and re-entry count.

10. Does it support out-of-sample validation, not just one big window?

You want to select parameters on one slice of data and test on a slice you never looked at. A platform that only produces one headline number over one window makes overfitting invisible. What should worry you: an optimiser that reports only the best result from a parameter sweep, with no validation step.

11. Is the report auditable at per-day and per-leg level, and exportable?

You should be able to open any single day, see each leg's entry price, exit price, exit reason and cost, and export the whole thing to CSV or Excel. A summary-only report cannot be checked, and an unchecked backtest is a belief, not evidence.

12. Can the same strategy actually be deployed live, and through which broker?

A backtest you cannot execute is an academic exercise. Check whether the platform deploys to a broker you can realistically open an account with, and whether the live order type matches the one the backtest assumed. One thing to note about the retail algo framework that took full effect on 1 April 2026. A plain market order is not passed through as such. It is converted to a market-price-protection order with a price band around it. A backtest that assumed clean market-order fills is therefore modelling an order type you will not actually get.

---

The Data Accuracy Test You Can Run in Twenty Minutes

You do not have to take a vendor's word on data quality. Run this before you pay for anything.

1. Pick a violent expiry day from the last twelve months — one where the index moved more than 1% intraday. Quiet days hide data problems; that is the whole point.
2. Pick a far-OTM strike on that expiry, roughly 2% away from where spot opened, and pick a time in the last ninety minutes of the session.
3. Pull the platform's historical price for that strike at that minute, both bid and ask if it offers them.
4. Compare against an independent source — your broker's historical chain viewer, or an exchange bhavcopy for the day's traded range on that contract.
5. Read the shape, not just the number. Real far-OTM prices near expiry are jumpy and carry wide spreads. If the platform shows a smooth, well-behaved curve where the market showed gaps and gaping spreads, you are looking at reconstructed data.
6. Repeat on a second expiry. One match can be luck.

A second test takes two minutes and catches stale platforms. Run any expiry-day strategy and export the trade log. Then check two things: that the trade dates land on the correct expiry weekday for that index, and that lot sizes change at the right point in the window.

---

What No Backtest Report Will Tell You: 163 Expiry Sessions of Our Own Data

Every platform in this guide will hand you a win rate, a profit factor and a maximum drawdown. None of them will tell you how an option-selling position actually behaves between entry and the close — and that behaviour is where most live results diverge from backtests.

So we measured it. The figures below come from 163 NIFTY and SENSEX weekly expiry sessions between 2 January 2025 and 30 July 2026 (83 NIFTY, 80 SENSEX), computed from one-minute option-chain and spot data on our own desk. The ATM strike is fixed at the 09:20 entry snapshot. Nothing here is simulated or back-filled.

1. The median expiry day gives back 42.8% of its best mark

For a short ATM straddle entered at 09:20, we tracked peak mark-to-market profit across the session in five-minute steps and compared it with the close.

Metric	Value
Sessions with a usable series	159 of 163
Median give-back, as % of peak profit	42.8%
Sessions giving back more than 30% of peak	86 of 159 (54%)
Sessions giving back more than 50% of peak	78 of 159 (49%)
Sessions where a profitable peak closed as a loss	48 of 159 (30%)

Why this matters for your backtest. A backtest that exits at the close reports the close number. It does not report that the position was sitting on roughly twice that profit earlier in the session, nor that roughly one session in three turned a winning position into a losing one before 15:29. Two strategies can post the same end-of-day P&L and be completely different things to hold. If your platform cannot show you the intraday mark-to-market path per day, you are flying blind on exactly the risk that causes traders to abandon a mathematically sound strategy.

2. Premium does not bleed smoothly to zero — it front-loads and then stalls

Averaged across all 163 sessions, here is ATM straddle premium at each 15-minute bucket as a percentage of the 09:16 entry premium.

Time	% of entry premium	Time	% of entry premium
09:16	100.0	12:31	89.2
09:31	95.4	12:46	88.8
09:46	96.0	13:01	89.1
10:01	95.3	13:16	90.1
10:16	93.1	13:31	92.0
10:31	92.1	13:46	93.3
10:46	90.6	14:01	93.5
11:01	89.0	14:16	91.9
11:16	88.6	14:31	91.3
11:31	89.0	14:46	88.7
11:46	88.1	15:01	90.9
12:01	89.0	15:16	91.3
12:16	90.1

Why this matters for your backtest. The popular story is that expiry-day premium decays steadily towards zero, so the longer you hold, the more you keep. The data says something else: premium falls hard in the first two hours, reaching 88.6% of entry by 11:16, and then chops sideways in an 88-93% band for the rest of the session. It never grinds down towards zero by 15:30.

The reason is mechanical, and it is the reason this curve is not a theta curve at all. The strike is fixed at the 09:20 ATM level while spot keeps walking away from it. So the series nets time decay against gamma re-pricing as the position drifts off that strike. Any backtest whose premium model assumes smooth decay — rather than reading real traded prices — will systematically overstate the profit available in the afternoon session.

3. Premium budget roughly doubles across VIX bands

ATM straddle premium at the 09:20 entry snapshot, as a percentage of spot, grouped by that session's opening India VIX.

India VIX band	Sessions	Mean premium (% of spot)	Median
Below 12	46	0.38%	0.38%
12 to 14	43	0.47%	0.44%
14 to 17	35	0.58%	0.51%
17 to 20	24	0.64%	0.60%
Above 20	11	0.83%	0.76%

Why this matters for your backtest. The relationship is monotonic across all five bands, with at least eleven sessions in every bucket. Premium collected at VIX below 12 is roughly half what is available above 20. A backtest window that happens to sit mostly in one VIX band is not measuring your strategy — it is measuring that band. This is the single most common reason a backtest "stops working": the regime changed, not the strategy. Ask your platform whether it lets you segment results by volatility regime. Most do not.

4. The close sits within about 0.1% of max pain, and SENSEX pins tighter

Max pain computed from the end-of-day open-interest snapshot across the full strike chain, using the standard minimise-writer-payout method.

Distance between expiry close and max pain strike	All (n=163)	NIFTY (n=83)	SENSEX (n=80)
Mean distance (% of spot)	0.122%	0.160%	0.083%
Median distance (% of spot)	0.083%	0.130%	0.060%
Maximum distance observed	0.81% of spot	—	—

Why this matters for your backtest. If you backtest a pin-risk or max-pain strategy on NIFTY and then deploy it on SENSEX, or the reverse, you are changing the instrument's behaviour, not just its ticker. SENSEX pinned roughly twice as tightly as NIFTY over this sample. A platform that treats the two indices as interchangeable test targets is hiding a real difference from you.

---

Five India-Specific Traps That Quietly Invalidate Backtests

These are not platform bugs. They are facts about the Indian market that a backtest will cheerfully ignore unless you force it not to.

1. The 1 April 2026 STT increase changed the arithmetic for sellers

From 1 April 2026, securities transaction tax on the sale of options rose from 0.10% to 0.15% of premium value. STT on exercised options rose from 0.125% to 0.15% of intrinsic value, and futures STT rose from 0.02% to 0.05%. Because STT on options is charged on the sell side, an option seller pays it at entry, on the full premium, on every leg, on every re-entry.

Any backtest run over a window that straddles 1 April 2026 using a single flat cost assumption is wrong on one side of that date. On a high-frequency multi-leg strategy with re-entries, a 50% increase in one charge line is not a rounding error.

2. Lot sizes changed, so older backtests are not comparable in rupee terms

NIFTY's lot size is 65 units, applying to contracts expiring from January 2026 onward; it was 75 for contracts expiring up to 30 December 2025. SENSEX trades in lots of 20. A platform that applies today's lot size across a multi-year window will misstate every rupee figure in the earlier part of it. Percentage returns survive; absolute rupee P&L, margin and drawdown do not.

3. Expiry weekday is per-index, and it moved

NIFTY weekly options expire on Tuesday. SENSEX weekly options expire on Thursday. Bank Nifty no longer has a weekly expiry at all. If your platform's expiry calendar is stale, every entry and exit timestamp in an expiry-day strategy lands on the wrong day. The strategy will then look either brilliant or broken, for reasons that have nothing to do with its logic.

4. Settlement is a closing-window calculation, not the last traded price

Index options settle against a weighted average of the underlying across the closing window, not against the final tick. A strategy held to expiry and backtested against closing LTP carries a systematic bias — small on most days, and largest precisely on the volatile days that determine whether a strategy survives.

5. The regime break of 2024-2025 makes long windows misleading

Three things changed at once: the derivatives lot-size revisions, the withdrawal of several weekly expiries, and the STT increase. So the Indian index-options market a strategy traded in 2023 is not the market it trades today. A six-year backtest is not six years of evidence about today's market. Treat the pre-revision period as a different instrument and check whether your edge exists in the post-revision window on its own.

---

Reading a Backtest Report Without Fooling Yourself

Four numbers get quoted. Here is what each one hides.

Win rate hides the shape of the losses

An option-selling strategy can post an 85% win rate and still lose money, because the 15% of losing sessions are several times the size of the winners. Win rate without average-win-to-average-loss is a vanity metric. Ask for both, every time.

Maximum drawdown hides its duration

A 20% drawdown that recovers in a week and a 20% drawdown that takes nine months are the same number and completely different experiences. The second one is what makes traders abandon a sound strategy. Look for drawdown duration, not just depth.

Total return hides the path

Our own data makes this concrete. The median expiry session in our 163-session sample surrendered 42.8% of its best intraday profit by the close, and 30% of sessions turned a profitable peak into a losing close. An end-of-day equity curve shows none of that. If you will be watching the position live, the intraday path is the number that determines whether you can hold it.

A parameter sweep's best result hides the sweep

If a platform tested 400 parameter combinations and shows you the best one, the headline number is the maximum of 400 draws, not an estimate of future performance. Ask what the median combination produced. If the best result towers over the median, the edge is probably in the fitting, not in the market.

---

A Step-by-Step Validation Walkthrough

This is the sequence we run before any strategy is allowed near live capital. It works on any platform that clears the checklist above.

1. Write the rules down first, in full. Entry time, strike selection rule, stop-loss, target, trail, re-entry count, exit time, position size. If a rule is not written before the first backtest, it will get tuned afterwards and you will never know it.
2. Split the window before you look at it. Reserve the most recent third as validation data and do not open it. Everything that follows happens on the older two-thirds.
3. Run the base case with honest costs. Full Indian charge stack, per-leg slippage, no optimism. If the base case is marginal here, stop. Optimisation will not save it; it will only hide it.
4. Stress the slippage. Double your slippage assumption and re-run. A strategy whose edge disappears when slippage doubles has an execution-quality edge, not a market edge, and you will not win that fight.
5. Segment by regime. Split results by VIX band. If the edge exists only in one band, you have a conditional strategy, not a general one — which is fine, as long as you know it and trade it accordingly.
6. Check the intraday path, not just the close. Pull the per-day mark-to-market series. Look at peak-to-close give-back and at how many sessions flipped from profit to loss. This is the number that predicts whether you will still be running the strategy in three months.
7. Now open the validation slice. Run once, with the parameters already frozen. If performance falls off a cliff, the earlier result was a curve fit. Do not re-tune and re-test — that spends the validation data and you cannot get it back.
8. Paper trade before capital. Run the strategy live on paper for at least one full expiry cycle, and compare the paper fills against the backtest's assumed fills. This is the only step that measures the gap between the model and your actual execution. Paper trade first — every time, without exception.
9. Size from the drawdown, not from the return. Position size should come from the worst drawdown you are prepared to sit through, not from the return you hope for.

---

Seven Mistakes We See Most Often in Backtested Option Strategies

1. Testing on the close and trading intraday. If you will manage the position live, an end-of-day backtest is measuring a different strategy from the one you will run.
2. One flat slippage number across all legs. The far-OTM hedge leg is where slippage actually lives, and it is nothing like the ATM leg.
3. Treating a long window as a strong window. Six years spanning two lot-size revisions and an STT change is not six years of evidence about today's market.
4. Optimising and reporting the maximum. The best of 400 parameter sets is a statistic about the sweep, not a forecast.
5. Ignoring margin and the capital it blocks. A strategy's return means nothing until it is divided by the margin the position actually blocks through the session.
6. Assuming clean market-order fills. Since 1 April 2026, a market order routed through the retail algo framework is converted to a market-price-protection order with a price band. A backtest that assumed an unbanded market fill is modelling an order type you will not get.
7. Skipping the paper-trading stage. It is the only step that measures your real execution against the model's assumed execution. Paper trade first.

---

Start Here: Pick by What You Are Actually Doing

Find the row that describes you, not the platform with the most features.

If this is you	Start with	Because
I want to test an idea this week and spend nothing	The 25-free-backtests-a-week platform, or the free-for-life plan	Both are standing allowances, not trials. You can kill a bad idea for ₹0.
I run a lot of experiments and cost per run is my constraint	Optest — ₹499 + GST a month, unlimited	The marginal cost of the four-hundredth backtest is zero, so you stop rationing experiments.
I want the deepest options-specific workflow in one tool	AlgoTest	Options-native leg logic, adjustments and re-entries, with one path from backtest to live.
I need to put a lot of variants through in one session	mt Quant — multiple backtests at the same time across different scripts	It is the only platform here that states concurrent runs, so throughput is the constraint it removes. The price is unpublished.
I need 6+ legs, or strike selection by delta or premium target	mt Quant — but get the data specs in writing first	24 legs and Greeks-based strike selection are not matched here. The unpublished data facts are the risk.
I want a live algo running as cheaply as possible	Tradetron at ₹250/mo, or uTrade Algos at ₹99/mo	Cheapest published routes to a live deployment. Both have shallow research horizons.
I want to slice results by expiry across six Indian indices	Stockmock — budget for both plan families	Group-by-expiry and group-by-day are purpose-built for this. Just know it is two purchases.
I write Python and want no ceiling	AlgoBulls, or the full DIY route	Any logic you can express, you can test. The work and the data cost are yours.
I mainly want to understand a position before I take it	An options analytics platform, not a backtester	Payoffs, Greeks and open interest answer a different question from "did this work over five years".
My strategy is intraday expiry-day option selling on NIFTY or SENSEX	Run the checklist hard, whatever you pick — and read our 163-session data first	This is the narrowest, least well-served case on the list, and the one where a generic backtest misleads most.
I have no strategy yet, just curiosity	A free tier, and the validation walkthrough	The method matters more than the tool at this stage. Paying now buys you nothing.

---

Glossary: 14 Terms You Need to Read a Backtest Report

Term	What it actually means
Backtest	A simulation of a strategy's rules against historical data, under stated assumptions about fills, costs and settlement. It is a model, not a record.
Tick data	Every traded price, as it happened. The highest-fidelity input, and the most expensive to store and serve.
Reconstructed data	Option prices generated by feeding a spot price series through a pricing model rather than captured from real trades. Looks plausible on quiet days; diverges badly on expiry days.
Slippage	The gap between the price your rule assumed and the price you would actually have been filled at. For far-OTM options near expiry this can be a large fraction of the premium.
Per-leg slippage	A separate slippage assumption for each leg of a multi-leg strategy, because a short ATM leg and a far-OTM hedge do not trade with the same spread.
Profit factor	Gross profit divided by gross loss. Above 1 means the strategy made money before costs; it says nothing about how it got there.
Maximum drawdown	The largest peak-to-trough fall in the equity curve. The number that determines whether you can actually hold the strategy.
Peak-to-close give-back	How much of a session's best intraday profit is surrendered by the close. Rarely reported, and on our sample the median was 42.8%.
Out-of-sample validation	Selecting a strategy on one slice of data and testing it on a slice you never looked at. Without it, a good backtest is just a well-fitted curve.
Overfitting	Tuning parameters until the strategy fits past noise. Shows up as a beautiful backtest that fails immediately in live trading.
Re-entry	Taking the position again after a stop-loss within the same session. Core to how real option-selling baskets are managed, and absent from many backtesters.
Trailing stop-loss	A stop that moves in your favour as the position gains. Materially changes results on expiry-day strategies, and materially harder to backtest correctly.
Max pain	The strike at which option writers' aggregate payout is smallest, computed from the open-interest chain. On our sample the close landed a median 0.083% of spot away from it.
Regime	A persistent market condition — typically a volatility band. A strategy's edge is often specific to one regime, which is why a backtest window's regime mix matters as much as its length.

---

Frequently Asked Questions

---

The Bottom Line

Once you have picked a platform, the remaining two questions are method and data: how to backtest an option-selling strategy is the full procedure, and NSE historical options data covers what to check before you trust the dataset you are testing on.

If you want the short version: AlgoTest is the most complete options backtesting workflow in India and its 25 free backtests a week make it the obvious place to start. Optest at ₹499 + GST a month is the best value on this page by a clear margin if you test heavily and can execute elsewhere. Tradetron's free-for-life plan and uTrade Algos at ₹99 a month are the cheapest routes to a running live algo. mt Quant is the one to look at for heavy, high-volume testing, and it has the deepest structural capability here: multiple backtests at the same time across different scripts, 24 legs, and Greeks-based strike selection. It also publishes the least about its data and its price. That is a trade you should make consciously rather than by default.

But the ranking is the less useful half of this guide, and we would rather be straight about that. Across eleven platforms, not one publishes whether its historical option prices are tick-level, one-second or one-minute, or whether they were captured from real exchange trades at all. That is the variable that most determines whether the number on your screen means anything, and the entire category is silent about it. No feature table can fix that. The twenty-minute data test can, and it is free.

So the sequence that actually protects your capital is this. Pick a platform from the table by what you are doing, not by feature count. Run the 12-point checklist against it and get written answers on data provenance. Run the data accuracy test on a violent expiry day before you trust a single output. Then follow the validation walkthrough — split before you look, stress the slippage, segment by regime, check the intraday path, and open the validation slice only once. Paper trade first, for at least one full expiry cycle, and compare the paper fills against what the backtest assumed.

And keep the 163-session figures in view while you read any backtest report. The median expiry session in our sample gave back 42.8% of its best intraday profit by the close, and 30% of sessions turned a profitable peak into a losing close. Premium did not bleed smoothly to zero — it fell to 88.6% of entry by 11:16 and then stalled in an 88-93% band. None of that appears in a win rate, a profit factor or an equity curve. It is, however, what you will actually be sitting through.

Your strategy may be intraday expiry-day option selling on NIFTY or SENSEX, multi-leg, with re-entry and trailing stop-loss logic. That narrow problem is what our backtesting engine was built around, on a desk that has traded this strategy class since 2006. You can read our methodology and how strategies are graded after backtesting on our services page, or more about how we work on About EliteAlgo. And if one of the platforms above suits you better, use it — we have no stake in which one you choose.

---

This article is educational and reflects EliteAlgo's own backtesting methodology and market observations. It is not investment advice. EliteAlgo is a proprietary trading desk and holds no SEBI or NSE registration as an investment adviser or research analyst; we publish analytical software and research, not recommendations. All platform pricing and features above were read from each vendor's own published pages in October 2026 and can change without notice — verify current terms with the vendor before purchasing. Algorithmic trading and options strategies carry a substantial risk of loss, and past backtested or live performance is not indicative of future results. Paper trade any strategy first. Review current NSE, BSE and SEBI guidelines and consult a SEBI-registered investment adviser before deploying capital.