Tick vs MinuteNovember 8, 202614 min read

Tick vs Minute Data: Which Resolution Fits Your Backtest?

Tick data records every trade, minute data summarizes them in bars. Learn what each shows, what each costs, and which fits your strategy's time horizon.

Tick data records every individual trade or quote update with its own timestamp. Minute data compresses everything that happened in 60 seconds into one bar: open, high, low, close and volume. Neither is "better" in the abstract. The right choice is the coarsest resolution that still captures what your strategy actually depends on. For most strategies that hold positions for minutes, hours or days, that is minute bars or higher. For strategies whose edge lives inside a single bar, it is tick data.

This guide explains what each format contains, what a one-minute bar hides, what tick data costs in storage and money, and how to match the data to your strategy.

TL;DR

Tick data is every trade or quote change with a timestamp, price and size. Minute data is a summary: four prices plus volume per minute.
Bars can be built from ticks, but ticks cannot be rebuilt from bars. That is why ticks are the raw record.
A one-minute bar cannot show whether the high or the low came first, which matters for stops, targets and scalping logic.
Tick data is heavy: one day of US consolidated quotes alone runs to tens of gigabytes. Minute bars are small and easy to work with.
Pick the lowest resolution that captures the dynamics your strategy relies on, then model spread and slippage honestly.

What tick data and minute data actually contain
Why you can build bars from ticks but not the reverse
What a one-minute bar hides
How resolution affects backtest accuracy
The cost of tick data: volume, storage and processing
Which resolution fits which strategy
Tick data versus Level 1, 2 and 3 data
A practical workflow for choosing and validating your data
Testing strategies without managing a data pipeline
Common mistakes when choosing data resolution
FAQ
Sources

What tick data and minute data actually contain

Tick data is the most granular market data there is. It records every individual quote change and trade with price, size and timestamp, depending on the feed. A full tick record typically carries the price, volume, and the bid and ask at that moment, as described in this overview of market data types.

Minute data, also called period or bar data, takes all those ticks and condenses them into one candle per interval. Each candle holds an open, a high, a low, a close and a volume figure. Everything that happened between those numbers is gone.

The two formats also differ in rhythm. Tick data is irregular: during busy periods hundreds of ticks can arrive in a second, while in quiet periods seconds or minutes pass with none. Minute bars arrive on a fixed clock, no matter how active the market was.

	Tick data	One-minute bars
What a row is	One trade or quote update	One minute of activity
Fields	Timestamp, price, size, often bid and ask	Open, high, low, close, volume
Timing	Irregular, driven by activity	Uniform, every 60 seconds
Intrabar path	Fully preserved	Lost
Size	Very large	Small
Best for	Execution-sensitive and microstructure work	Most intraday systematic strategies

Why you can build bars from ticks but not the reverse

This is the core asymmetry. Ticks are the source record, and bars are a derived product. You can always build bars from ticks but not the reverse, which is why many traders who own tick data resample it into whatever bar size a given test needs.

Ticks can form bars but bars cannot restore ticks

Starting from bars, the full path of updates inside the interval is not preserved, so no amount of clever processing recovers it. If you later discover you need finer detail, you have to go back to the source.

That said, the reverse is also worth keeping in mind: most of the time you do not need the source. Having the option is not the same as needing it.

What a one-minute bar hides

Imagine a bar that says a stock traded between 100.00 and 100.20 and closed at 100.15. That is all you know. It cannot tell you whether the market moved smoothly upward, gapped around on thin liquidity, or briefly traded through several levels before snapping back.

The most practical consequence is the order of the high and the low. A one-minute bar by design cannot show whether the high or the low printed first, so two completely different paths can produce an identical bar. Two candles with identical OHLC values can follow very different routes, and a strategy reacting to intraminute moves may behave differently on ticks than on candles.

Here is where that bites in a backtest:

Stop and target in the same bar. If a bar's range covers both your stop and your profit target, the bar cannot say which was hit first. Any backtest on bars has to make an assumption, and the result flips depending on it.
Limit order fills. A bar that touched your limit price does not tell you whether your order would have been filled, or how much size traded there.
Fast markets. Around news, opens and liquidations, price can sweep through several levels in seconds. A bar shows the result, not the sweep.

These are not reasons to avoid minute data. They are the specific places where you should be careful about how a backtest resolves ambiguity, and where you should know the assumption in use.

How resolution affects backtest accuracy

The argument for tick data in testing is about execution realism. One data provider's analysis argues that tick-level data gives microsecond or nanosecond exchange timestamps and detailed slippage and queue modeling, while minute bars rely on broad percentage-based cost estimates. The same source notes that sampling too coarsely lets microstructure noise distort volatility and price-discovery estimates.

Another data provider makes a related point: even for lower-frequency strategies, OHLCV data often obscures when within a bar a trade would really have executed, which makes fill timing and slippage harder to model. Spreads also widen in volatile moments, so a constant-spread assumption can understate true costs.

The practical lesson is not "use ticks or your test is wrong." It is this: resolution is only one of the things that make a backtest honest. Whatever bars you use, the test still needs realistic assumptions about:

the bid-ask spread you would pay,
slippage, especially in fast moves,
fees and, for perpetual futures, funding,
how same-bar stop and target conflicts are resolved,
enough history across different market regimes.

A tick-level test with no spread and no slippage is less honest than a minute-bar test with conservative costs. Data detail and test discipline are separate things, and you need both. Our guide on how long your historical data should be covers the regime and sample-size side of that question, and the Backtestify methodology page explains how we simulate a strategy.

The cost of tick data: volume, storage and processing

Tick data is large because markets generate a huge number of events. NYSE's Daily TAQ specification, version 4.3 dated March 3, 2026, lists typical sizes for one day of consolidated US equities: about 2.4 GB for 69 million trade records, 38 GB for 1.9 billion quote records, and 11 GB for 330 million NBBO records, all gzipped.

That is one day, for the whole US equity market. Even a single instrument over a few years means billions of rows, which affects:

Storage. You need disk space and, realistically, a database or columnar file format rather than a spreadsheet.
Processing time. A test that runs in seconds on minute bars can take far longer on ticks, which slows down any optimization loop where you try many rule variations.
Data quality work. Raw ticks include bad prints, duplicates, out-of-order records and, for futures, contract-roll issues. A checklist for dataset quality should cover aggressor side, consistent high-resolution timestamps, documented roll handling, corrupt or duplicate records, and coverage across multiple regimes.
Licensing. Exchange data is often licensed, and fees vary by exchange, product and how you use the data. Check the exchange's or vendor's current terms for the instruments you trade.

Minute bars avoid nearly all of this. They are compact, easy to clean, quick to iterate on, and cover long histories cheaply. That speed matters, because the more variations you can test, the more you learn about whether an edge is robust.

Which resolution fits which strategy

A widely cited rule of thumb: use the lowest resolution that captures the dynamics your strategy depends on. A futures data guide frames it this way: daily bars for longer-horizon trend-following and swing strategies, intraday bars such as 1, 5 and 15 minutes for most systematic futures strategies, and tick data for execution-sensitive strategies, limit-order fill modeling and microstructure analysis.

Strategy type	Typical holding period	Sensible starting resolution
Position or trend following	Weeks to months	Daily bars
Swing trading	Days	Daily or hourly bars
Intraday breakout, trendline, moving-average systems	Minutes to hours	1 to 15 minute bars
Scalping with tight stops	Seconds to minutes	1-minute bars with care, or ticks
Order-book, queue or market-making logic	Seconds or less	Tick data plus Level 2

Two questions help you decide:

How big is your edge relative to the bar? If your average winner is many times the typical one-minute range, the order of events inside a bar rarely matters. If your target is a few ticks, it matters a great deal.
Does your rule act inside the bar? Rules evaluated at the bar close ("enter when the close is above the 20-period average") do not need intrabar data. Rules that rely on touching a precise price intrabar are the ones that benefit from ticks.

For traders moving from daily to intraday ideas, our walkthrough on intraday backtesting for day trading strategies shows how the details change as timeframes shrink, and the swing trading backtest guide covers the slower end. For a view on whether finer data is worth it for your time horizon and edge, this note on real-time exchange market data is a useful companion.

Tick data versus Level 1, 2 and 3 data

People often mix up tick data with market depth levels. They describe different things. Levels describe quotes; tick data describes trades.

Level 1 shows the top of the book: last trade price and size, best bid and ask with sizes, and the day's open, high, low and volume.
Level 2 shows several price levels of the order book with quantities at each.
Level 3 shows individual orders in order-by-order feeds.

Tick data lists every execution with price, size and time. Some strategies combine Level 2 depth with tick-by-tick trades to see both intent and action. Most retail strategies built on chart patterns, trendlines or indicators need neither the depth nor the sequence of orders, just a reliable price series.

A practical workflow for choosing and validating your data

You do not have to pick once and live with it. A sensible sequence:

Write the rules precisely. Entry, exit, stop, target, and what happens when two trigger in the same bar.
Start at the resolution that matches your decision timeframe. If you trade a 15-minute chart, test on 15-minute bars or finer, not on a coarser series.
Add realistic costs. Include spread, slippage and fees before you look at the results.
Stress the ambiguous bars. Test the strategy under the pessimistic assumption for same-bar conflicts. If the edge disappears, resolution is your problem and finer data deserves a look.
Hold out unseen data. Keep a period, ideally from a different market regime, that you never used while tuning. A good backtesting approach is to validate on out-of-sample data and reconcile live results with the test.
Compare live fills to the backtest. Early live trades show whether your cost assumptions held.

If you want a full process from idea to result, our step-by-step guide to backtesting a trading strategy walks through it.

Testing strategies without managing a data pipeline

If you would rather spend your time on rules than on downloading, cleaning and storing market data, Backtestify lets you describe a strategy in plain English and run it on real historical market data. You get the trade-by-trade detail and metrics such as win rate, profit factor and maximum drawdown, and you can compare an original strategy with an improved version side by side on recent data. You can also see how published strategies performed, wins and losses alike, in the strategy library.

The Free plan lets you get started, and the Pro plan is $29 per month or $190 per year for unlimited backtesting, improvement and forecasting. See the plans or go straight to Backtestify Pro.

Backtestify

Common mistakes when choosing data resolution

Assuming more detail means a better test. Ticks give you more information, not better assumptions. A test is only as honest as its costs and its sample.
Testing on a finer series than your decisions use. Your rules should act on the same information you would have live.
Ignoring same-bar conflicts. If stop and target both sit inside one bar's range, know how your tool resolves it.
Using too little history. A beautifully detailed month of ticks says less about robustness than several years of bars across varied conditions.
Skipping dataset checks. Bad prints, gaps and roll errors will distort results at any resolution.

FAQ

Is tick data more accurate than minute data?

Tick data is more detailed, because it keeps every trade and the order they happened in. Accuracy of a backtest also depends on spread, slippage, fees and sample size. For strategies that hold for minutes or longer, well-modeled minute bars are often enough.

Can I convert tick data into minute data?

Yes. Grouping ticks into one-minute windows and taking the first, highest, lowest and last price plus total volume produces standard minute bars. The reverse is not possible, because the sequence inside each bar is lost once ticks are summarized.

Do I need tick data for day trading strategies?

Not necessarily. Many intraday systems that decide on bar closes work well with 1 to 15 minute bars. Tick data becomes important when your targets are only a few ticks, when you model limit-order fills, or when the order of the high and low inside a bar changes your result.

How much storage does tick data need?

A lot more than bars. NYSE's Daily TAQ specification lists about 38 GB gzipped for one day of consolidated US equity quotes alone. Minute bars for the same instrument over years fit comfortably on a laptop.