How big is forex tick data?

Roughly 1–2 GB per instrument-year as CSV for an active major pair, at 40–50 bytes per quote — but it varies widely with market activity. A full multi-decade tick archive for one pair runs to tens of gigabytes.

How big is 1-minute historical data?

A 24×5 instrument produces about 360,000 one-minute candles a year — roughly 18 MB as CSV. Twenty years is around 7 million rows and 350 MB: large for a spreadsheet, trivial for pandas or a database.

How many bytes is a row of market data?

About 40–50 bytes as CSV for a timestamped OHLC or tick row. A measured 4.8-million-row export weighs 181 MB, almost exactly 40 bytes per row.

Does downloading more data cost more?

Not here. The software licence is a one-time purchase and retrieval runs on your own machine, so there is no per-download or per-gigabyte charge. Every licence tier reaches the full history of every instrument.

How Big Is Tick Data? File Sizes by Timeframe

The first tick download surprises everyone. A decade of daily candles is a file your email could carry; the same decade in ticks is tens of gigabytes. This guide puts real numbers on that gap — so you can size a download before you start it, not after it fills the disk.

Get the app — €5.60 →

The rule of thumb

A CSV row of market data — timestamp plus prices — weighs roughly 40–50 bytes. A real 4.8-million-row export from the MarketData Hub client measures 181 MB on disk, about 40 bytes a row. So the only question that matters is: how many rows does your timeframe produce? That is where the factor-of-a-million spread comes from: one instrument-year is a few hundred daily candles, a few hundred thousand 1-minute candles, or tens of millions of ticks.

Worked example: EUR/USD

EUR/USD trades roughly 24 hours, five days a week. Its full history, timeframe by timeframe:

Timeframe	History from	Approx. rows	Approx. CSV
Tick (every quote)	2003-05-04	very large	—
1 second	2003-05-04	very large	—
1 minute	2003-05-04	~8.4M	~421 MB
5 minutes	2003-05-04	~1.7M	~84 MB
15 minutes	2003-05-04	~562K	~28 MB
30 minutes	2003-05-04	~281K	~14 MB
1 hour	2003-05-04	~140K	~7 MB
4 hours	2003-05-04	~35,100	~2 MB
Daily	1973-03-01	~13,500	~673 KB
Monthly	1973-03-01	~640	~32 KB

Worked example: BTC/USD

Bitcoin never closes — 24 hours, seven days — so per year it produces about 40% more candles than a forex pair at the same timeframe, despite its shorter history:

Timeframe	History from	Approx. rows	Approx. CSV
Tick (every quote)	2017-05-07	very large	—
1 second	2017-05-07	very large	—
1 minute	2017-05-07	~4.8M	~242 MB
5 minutes	2017-05-07	~968K	~48 MB
15 minutes	2017-05-07	~323K	~16 MB
30 minutes	2017-05-07	~161K	~8 MB
1 hour	2017-05-07	~80,700	~4 MB
4 hours	2017-05-07	~20,200	~1 MB
Daily	2017-05-07	~3,400	~168 KB
Monthly	2017-05-07	~110	~6 KB

Why tick rows are "very large" instead of a number

Candle counts follow from session hours, so they can be estimated honestly. Tick counts can't: they follow from activity, which varies by orders of magnitude between a quiet Asian session and a payrolls release. As an order of magnitude, an active major produces tens of millions of quotes a year — roughly 1–2 GB per year as CSV, more in volatile years. Treat any precise-sounding tick estimate with suspicion, ours included; the honest answer is a range.

How to keep downloads manageable

Use the coarsest timeframe that answers your question. Most backtests don't need ticks — see tick data vs OHLC for when each is right. Dropping from tick to 1-minute cuts the size by roughly two orders of magnitude and loses nothing a candle-based strategy can use.
Slice tick ranges. Fetch a year at a time rather than the whole archive in one job. The client queues long runs and resumes them if interrupted, but smaller files are also simply easier to work with afterwards.
Prefer columnar formats for big series. CSV is the universal interchange; it is also the bulkiest. The Pro licence writes Parquet — typically several times smaller than the same data as CSV, and dramatically faster to query (see querying market data with SQL). JSON goes the other way: repeating the key names on every row makes it the largest option, so keep it for small extracts.
Know your tool's ceiling. Excel stops at 1,048,576 rows — decades of daily data, a couple of years of 1-minute, no ticks at all (see the spreadsheet guide). pandas wants the working set in RAM; count on several times the on-disk CSV size once parsed.

Nothing here is priced by size

One thing file size does not affect: cost. The MarketData Hub client fetches on your own machine, straight from the source — there is no metering, no per-gigabyte charge, and no tier that unlocks "more" data. A €5.60 licence downloads the same ticks a €40 one does; higher tiers change export formats and automation, never volume. Check any instrument's page — every one lists its own per-timeframe row and size estimates, like gold or the US 500.

Frequently asked questions

How big is forex tick data?: Roughly 1–2 GB per instrument-year as CSV for an active major pair, at 40–50 bytes per quote — but it varies widely with market activity. A full multi-decade tick archive for one pair runs to tens of gigabytes.
How big is 1-minute historical data?: A 24×5 instrument produces about 360,000 one-minute candles a year — roughly 18 MB as CSV. Twenty years is around 7 million rows and 350 MB: large for a spreadsheet, trivial for pandas or a database.
How many bytes is a row of market data?: About 40–50 bytes as CSV for a timestamped OHLC or tick row. A measured 4.8-million-row export weighs 181 MB, almost exactly 40 bytes per row.
Does downloading more data cost more?: Not here. The software licence is a one-time purchase and retrieval runs on your own machine, so there is no per-download or per-gigabyte charge. Every licence tier reaches the full history of every instrument.