Tick Data vs OHLC: What Happens Inside the Candle

What a Tick Is (in OTC Markets)

In spot forex and CFDs, a “tick” is a quote update: a new bid/ask pair from the broker’s pricing engine, timestamped to the millisecond. It is not a trade record (there’s no central tape in OTC markets — see broker feed differences). An active hour on EUR/USD can print thousands of quote ticks; a quiet exotic hour, a handful. Tick data is the complete record of what prices you could actually have dealt at, in what order, with what spread — because each tick carries both sides of the quote.

The Information a Candle Destroys

Build a 15-minute candle from 600 ticks and you keep open, high, low, close — and destroy:

• Sequence — did the high print before the low, or after? Did price cross the midrange five times or once?
• Spread — one price series remains (usually bid); the bid/ask gap and its news-time widening vanish.
• Time-of-touch — when within the bar a level traded, which matters for session rules and time-based exits.
• Path texture — a slow grind to the high and a single spike to the same high look identical in OHLC.

The Ambiguity That Breaks Backtests

Concrete bar: EUR/USD M15 opens 1.0840, high 1.0855, low 1.0826, closes 1.0848. You were long from 1.0838 with stop 1.0828 and target 1.0853 — both levels are inside the bar’s range.

• If the path went down first (1.0826 sweep, then rally): you were stopped. −10 pips.
• If the path went up first (1.0855 tag, then dip): target hit. +15 pips.

The candle cannot say which happened. A 25-pip swing in the outcome of one trade — and over a 100-trade backtest, dozens of bars contain such conflicts. Candle-based engines resolve them with rules (“closer level first,” “optimistic,” “assume target”) that, across a test, accumulate the systematic optimism behind so many live-vs-backtest gaps. Tick data doesn’t resolve the ambiguity — it never has it: the path is in the data.

When OHLC Is Genuinely Fine

Candles are not the villain; mismatched granularity is. OHLC is adequate when the trade’s geometry is bigger than any one bar: daily-chart swing entries with 80-pip stops, position trades managed on closes, pattern studies on completed structures. The heuristic from the other side: count how often your backtest hits a same-bar stop/target conflict. Near zero → candles are fine; regularly → every such trade is a guess, and the test inherits the guessing.

Practicalities (Why People Settle for Candles)

Tick data is bulky — a year of EUR/USD runs to gigabytes raw — and the traditional pipeline (download from Dukascopy or similar, convert formats, import, patch gaps) costs hours before the first replay. That friction, not analytical preference, is why most manual backtesters end up on candles. It’s also increasingly avoidable: browser replay tools stream tick history on demand — StrategyTune replays recorded bid/ask ticks for 70+ instruments with nothing to download, which makes tick-level testing the lazier option rather than the harder one. (Disclosure: this site is run by StrategyTune’s team — about.)

Related: broker feed differences · spread modeling · how much history you need

Frequently Asked Questions

Is tick data the same as Level II / order book data?

No. Retail OTC tick data records top-of-book quote updates — best bid and ask over time. Level II adds market depth (size available at multiple price levels), which matters for order-flow trading in futures and equities but is generally unavailable and unnecessary for retail forex/CFD backtesting.

How accurate are 'tick' bars generated from M1 data?

Interpolated ticks (as in MT4's classic 99% modeling quality approach) reconstruct a plausible path inside each minute, but it remains a synthetic path — fine for rough intraday work, still a guess at the exact moments that decide tight stops. Real recorded ticks are categorically different: they're what was actually quoted.

Does tick replay make backtests slower?

Not meaningfully with modern tools — replay speed is adjustable (StrategyTune, for example, runs from 1× up to 50,000×), so you fast-forward quiet periods and slow down around entries. The bottleneck in manual backtesting is your decision time, not the data granularity.

Do I need tick data for a daily-timeframe strategy?

Almost never. If stops sit far outside any single bar's range and entries aren't intra-bar timing-sensitive, OHLC carries all the information your test uses. Spend the saved effort on regime coverage and sample size instead — those move daily-strategy validity far more.