Home / Market Data / Minute Data
Intraday barsBeginner

Minute Data

Minute data is intraday OHLC bars sampled at one-minute (or a few-minute) intervals, giving enough intraday resolution for most systematic strategies while staying far smaller and easier to work with than raw ticks.

Quick answer: Minute data is intraday OHLC bars sampled at one-minute (or a few-minute) intervals, giving enough intraday resolution for most systematic strategies while staying far smaller and easier to work with than raw ticks.

In simple words

Minute data breaks the trading day into one-minute slices, each with its own open, high, low, close and volume. It is the workhorse resolution for intraday algorithmic trading — detailed enough to see moves within the day, but compact enough to store years of it and backtest quickly. It sits between coarse daily bars and overwhelming tick data.

Purpose

Minute bars give intraday strategies a practical, uniform time grid to work on — fine enough to capture intraday structure and time-based rules, without the storage and compute burden of ticks.

Professional explanation

How minute bars are constructed from ticks

A one-minute bar aggregates every trade in a 60-second window: the open is the first trade's price, the close the last trade's price, and the high and low the extremes, with volume summed. The critical design choice is the window boundary and whether the timestamp marks the bar's open or its close — a bar labelled 09:20 might cover trades from 09:20:00 to 09:20:59 (open-stamped) or from 09:19:00 to 09:19:59 (close-stamped). Different data vendors choose differently, and mixing two conventions in one pipeline shifts every signal by a bar, an off-by-one error that quietly wrecks intraday backtests.

The NSE session and its exact boundaries

The NSE equity and F&O continuous session runs 9:15 to 15:30 IST, preceded by a pre-open auction from 9:00 to 9:15. This means a valid intraday day has 375 one-minute bars (9:15 up to 15:30). The first bar (09:15) is special — it opens on the auction-discovered price and often carries a large gap and volume; the last bar (15:29 / 15:30) sits inside the closing session. There is no lunch break on NSE, unlike some Asian exchanges, so bars are continuous through the day. Any minute pipeline must encode these exact boundaries or it will fabricate bars outside trading hours.

Alignment, missing minutes and thin instruments

A liquid instrument like Nifty futures trades every second, so all 375 minute bars exist. A thin option strike may not trade for several minutes, leaving genuine gaps in the minute series. How you handle these gaps — leave them absent, forward-fill the last price, or synthesise a flat bar — changes indicator values materially. Aligning multiple instruments onto a common minute grid (for pairs or portfolio logic) requires an explicit calendar and a chosen fill rule; naive joins silently drop or duplicate bars where one instrument traded and another did not.

Why minute data is the intraday sweet spot

Minute bars balance three competing pressures: resolution, storage and backtest speed. They resolve intraday structure — opening-range breakouts, time-of-day effects, VWAP behaviour — that daily bars cannot see, while a full year of one-minute Nifty data is only a few tens of megabytes versus terabytes of ticks. They are fast enough to run large parameter studies over. The trade-off is that, like all bars, they still hide the intrabar path, so same-minute stop-versus-target ambiguity remains and precise fill modelling still needs ticks.

How minute-data errors silently corrupt intraday systems

Intraday backtests are unusually sensitive to timestamp and boundary errors. If bars are close-stamped but your code treats them as open-stamped, every signal fires one minute early — effectively look-ahead bias, because you act on information from a bar that has not finished. A single missing 09:15 opening bar changes an opening-range calculation. Daylight or timezone confusion (storing IST bars but computing in UTC) shifts the whole session by 5 hours 30 minutes, placing 'trades' in the middle of the night. None of these throw an error; they just produce a plausible, wrong equity curve.

Practical example

Illustrative example (Indian market)

Consider an opening-range breakout on Nifty using the first 15 minutes. You take the high and low of the 09:15 to 09:29 bars, then go long if price breaks the range high after 09:30. If your feed is close-stamped (the 09:15 bar actually covers 09:14 to 09:15, i.e. pre-open) but your code assumes open-stamped, your 'first 15 minutes' silently include a pre-open bar and exclude 09:29 — the range is computed on the wrong window and every breakout level is off. On a day where Nifty opens at 25,000, ranges 24,960 to 25,050, the mislabelled range might read 24,940 to 25,030, flipping several signals. The code runs cleanly and the bug is invisible without checking bar timestamps against the known 375-bar session.

NSE's 375 one-minute bars per day is a useful validation constant: a clean intraday day for a liquid instrument should have exactly 375 bars from 9:15 to 15:30. A day with 376 or 300 bars signals a boundary bug, a holiday, a half-day (like a Muhurat session) or missing data.

Advantages

  • Enough intraday resolution for opening ranges, time-of-day and VWAP-style logic
  • Far smaller than ticks — years of minute bars fit in tens of megabytes per instrument
  • Fast to backtest, so it supports large parameter studies intraday strategies need
  • Uniform grid (375 bars/day on NSE) that is easy to validate and align across instruments

Limitations

  • Still hides the intrabar path — same-minute stop-versus-target remains an assumption
  • Open-stamp versus close-stamp ambiguity causes off-by-one-bar look-ahead bugs
  • Thin instruments have genuine missing minutes that fill rules can distort
  • Insufficient for microstructure or precise fill modelling, which need ticks
  • Session-boundary and timezone errors place trades outside real trading hours

Common mistakes

  • Mixing open-stamped and close-stamped minute data in one pipeline, shifting every signal by a bar
  • Assuming a full day has more or fewer than the 375 NSE minute bars, and not validating the count
  • Forward-filling missing minutes on thin option strikes, creating flat bars that distort indicators
  • Computing intraday signals in UTC while bars are stamped in IST, shifting the whole session 5.5 hours
  • Treating the 09:15 opening bar as an ordinary bar when it reflects the pre-open auction and a gap
  • Acting on the current minute's close and filling on that same close, which is not executable live

Professional usage

Serious intraday desks store minute bars as a validated derivative of clean ticks, with each bar carrying an explicit timestamp convention (open- or close-stamped), timezone (IST) and session calendar including holidays and half-days. Pipelines assert the expected bar count per session and reject days that fail, so boundary bugs surface immediately rather than in a mysterious backtest result. Where fill precision matters, the minute backtest is cross-checked against a tick-level simulation on a sample. The discipline is to make every assumption about a minute bar explicit and machine-checked rather than implicit.

Key takeaways

  • Minute bars are the intraday workhorse — resolute enough for most strategies, small enough to be practical
  • Always know whether bars are open-stamped or close-stamped; mixing them creates look-ahead bugs
  • A clean NSE intraday day has exactly 375 one-minute bars from 9:15 to 15:30 — use it as a check
  • Timezone and session-boundary errors silently place trades outside real hours; validate against IST

Frequently asked questions

What is minute data in trading?
Minute data is intraday OHLC bars sampled every minute (or every few minutes), each with an open, high, low, close and volume. It is the standard resolution for intraday systematic trading — detailed but far smaller than tick data.
How many one-minute bars are in an NSE trading day?
Exactly 375, covering the continuous session from 9:15 to 15:30 IST for a liquid instrument. This makes 375 a handy validation constant: a day with a different count signals a bug, holiday, half-day or missing data.
What is the difference between open-stamped and close-stamped bars?
It is which end of the interval the timestamp marks. An open-stamped 09:20 bar covers 09:20:00 to 09:20:59; a close-stamped one covers the minute ending at 09:20. Mixing the two conventions shifts every signal by one bar.
Why does the timestamp convention cause look-ahead bias?
If bars are close-stamped but your code treats them as open-stamped, it acts on a bar before that bar has actually finished forming, using information not yet available in real time. That is effectively look-ahead bias introduced by a labelling error.
Does NSE have a lunch break that affects minute bars?
No. Unlike some Asian exchanges, NSE runs a continuous session from 9:15 to 15:30 with no lunch break, so minute bars are unbroken through the day. Any gap in a liquid instrument's minute series indicates a data problem.
How should I handle missing minutes on thin instruments?
It depends on the strategy. You can leave them absent, forward-fill the last price, or synthesise flat bars — but each choice changes indicator values, so it must be deliberate and documented, not an accidental default in your data library.
Is minute data enough for intraday backtesting?
For most intraday strategies on liquid instruments, yes, provided it is clean and correctly aligned. It is not enough for microstructure work or precise fill modelling, which require tick or quote data.
What is special about the 09:15 opening bar?
It opens on the price discovered by NSE's pre-open auction (9:00 to 9:15), not on an ordinary first trade, and it often carries a large gap and heavy volume. Treating it as a normal bar can distort opening-range and gap logic.
How do timezone errors corrupt minute data?
NSE bars are in IST. If your code computes signals in UTC without converting, the whole session shifts by 5 hours 30 minutes, placing trades outside real trading hours. It never errors — it just produces a wrong result.
How do I align minute data across multiple instruments?
Use an explicit shared calendar and a chosen fill rule to place each instrument on a common minute grid. Naive joins silently drop or duplicate bars wherever one instrument traded in a minute and another did not.
Can I build minute bars from tick data myself?
Yes, and it is good practice: aggregate clean ticks into 60-second windows with a fixed boundary and timestamp convention. Building bars from your own clean ticks avoids inheriting a vendor's undocumented conventions.
Why is minute data preferred over tick data for most intraday work?
It captures the intraday structure most strategies need while being orders of magnitude smaller and faster to backtest. Tick data is reserved for the specific cases — microstructure, fills — where its extra detail genuinely matters.
What does a bar count other than 375 tell me?
It flags something to investigate: a full trading day should have 375 one-minute bars. Fewer may mean a half-day, holiday or missing data; more usually means a session-boundary bug fabricating bars outside 9:15 to 15:30.
Do minute bars still have the intrabar-path problem?
Yes. Like any OHLC bar, a minute bar does not record whether its high or low came first, so same-minute stop-versus-target outcomes remain an assumption the data cannot resolve.

Voice search & related questions

Natural-language questions people ask about Minute Data.

What is minute data?
It is price data broken into one-minute bars, each with an open, high, low and close. It is the usual choice for intraday trading because it shows enough detail without being as huge as tick data.
How many minute bars are in a trading day on NSE?
Exactly 375, from the 9:15 open to the 3:30 close. If you count a different number, something is wrong — a bug, a holiday, a half-day or missing data.
Why does the bar timestamp matter so much?
Because a bar can be labelled by its start or its end, and if your code assumes the wrong one, every signal shifts by a minute and you can act on data before it actually exists.
Does the Indian market have a lunch break?
No, NSE trades continuously from 9:15 to 3:30 with no lunch break, so a gap in a liquid instrument's minute data usually means a data problem, not a real pause.
Is minute data good enough for my strategy?
For most intraday strategies on liquid instruments, yes. You only need finer tick data for very short-term microstructure ideas or to model exactly how your orders fill.
What time zone is Indian minute data in?
Indian Standard Time. If you accidentally process it in UTC, the whole trading day shifts by five and a half hours and your trades land at the wrong time.

Sources & references

    Last reviewed 11 July 2026. Educational content only — not investment advice. Markets and rules change; verify current conventions with SEBI, NSE/BSE and your broker.

    ← Previous
    Tick Data
    Educational content only — not investment advice. Examples use illustrative numbers and simplified models. Algorithmic trading and derivatives involve substantial risk. See our Risk Disclosure and SEBI Disclaimer.