What tick-by-tick market data actually means (and why your chart depends on it)

Open any chart and you are looking at candles. Each one seems like a fact: this is where the market opened, this is how high and low it went, this is where it closed. But a candle is not a fact. It is a summary of facts — sometimes a summary of thousands of them — and the facts themselves are individual trades. Tick-by-tick data is simply the practice of keeping every one of those trades. This article explains what a tick actually contains, how tick data differs from the bars you chart with, and why the difference decides whether the tools on your screen are measuring the market or guessing at it.

What one tick actually is

A tick is one trade. Not a price update, not a quote, not a refresh of your screen — one completed transaction between a real buyer and a real seller.

Each tick carries a small, precise set of facts:

• The price the trade happened at — the last traded price at that instant.
• The quantity that changed hands in that single trade.
• The time, stamped to the second or finer, in exchange time.
• The running totals that came with it — cumulative traded volume for the day and, on derivative contracts, the open interest at that point.

That is the entire raw material of a trading day. Every chart, every indicator, every scanner and every backtest is a transformation of this sequence. There is no deeper layer underneath it — ticks are where market data bottoms out.

A candle is a compression of ticks

Now watch what happens when those trades become a chart. A one-minute candle takes every tick inside its sixty seconds and keeps exactly four prices: the first (open), the highest, the lowest, and the last (close), plus the summed volume. Everything else is thrown away.

On a liquid index future, one minute can contain several thousand individual trades. The candle keeps four numbers out of thousands. That is not a criticism — compression is the whole point of a chart, and for many styles of trading the four numbers are enough. But it is worth being clear-eyed about what was discarded: the order the trades happened in, the size of each one, where the volume clustered, whether the move was a hundred small trades or three large ones.

The resolution ladder

It helps to think of market data as one ladder with the same trades at every rung, summarised harder as you climb:

• Tick — every trade, in sequence. Nothing summarised.
• One-minute bars — each bar compresses the ticks of one minute into open, high, low, close and volume.
• Hourly and daily bars — the same compression applied over longer windows.

The direction of travel matters. You can always build minute bars from ticks, and daily bars from minute bars — summarising downward is just arithmetic. You can never go the other way. Once the ticks are gone, no amount of processing recovers the sequence from the summary, for the same reason a monthly bank statement cannot tell you the order of your transactions.

This is why the question “does my vendor keep ticks?” matters even if you personally chart on five-minute candles: the tick record is the source everything else is checked against. When a printed bar looks wrong, the only honest way to verify it is to look at the trades inside it.

Recorded versus reconstructed

Here is the part most traders never get told. Two vendors can both sell you “tick data” and mean very different things.

Recorded tick data is captured live from the exchange feed as the trades happen, written down in sequence, and kept. What you download later is what was seen at the time.

Reconstructed tick data is manufactured after the fact — typically by taking bar data and interpolating plausible trades inside each bar, or by sampling the feed a few times a second and calling each sample a tick. The output looks like tick data. It has timestamps and prices. But the sequence, the trade sizes and the clustering are invented.

For a plain price chart the difference is invisible — both versions draw the same candles, which is precisely why reconstruction survives as a practice. The difference appears the moment a tool tries to read inside the bar. A volume profile built on invented ticks puts the volume at invented prices. An order-flow chart classifies invented trades as buying or selling. A backtest fills orders against a sequence of prices that never traded. None of these tools fail loudly on reconstructed data; they simply report things that did not happen, with full confidence.

Volume and open interest ride on the same ticks

Ticks do not only carry price. Two other series travel with them, and both are only as good as the tick record underneath.

Traded volume is the running count of quantity changing hands. When it arrives tick by tick, your chart can say not just how much traded but at which prices it traded — which is the entire basis of volume profile analysis. Summarised volume can only tell you the total.

Open interest, on futures and options, is the count of contracts currently open. It updates through the session, and reading it against price — is this rally new positions being built, or old positions being unwound? — requires that the OI series line up with the price series in time. That alignment is a property of the tick record. It also has to survive contract transitions correctly: when analysis moves from one contract to the next, a vendor that adjusts price history without treating open interest consistently leaves the two series telling different stories. We covered why that matters for a derivatives chart in more depth in the order flow article.

Who actually needs tick data

Honest answer: not everyone, not directly. If you trade end-of-day from daily charts, bars serve you well and the tick record matters to you only as the source your bars were built from.

You need the ticks themselves the moment you use any of the following:

• Volume profile and market profile tools, which distribute the day’s trade across price levels — a distribution that only exists in the tick record.
• Order flow and footprint charts, which classify each trade by which side initiated it. Sampled data does not weaken this analysis; it fabricates it.
• Intraday backtests, where fills, slippage and stop behaviour are only realistic if tested against prices that genuinely printed, in the order they printed.
• Your own software, if you are building tools over an API and computing anything — VWAP, imbalance, custom bars — from raw events rather than finished candles.

How to check what your vendor gives you

You do not need to take any vendor’s word for this, ours included. Three checks, in increasing order of effort:

• Ask the direct question: is intraday history recorded from the live feed, or built later from bars? A vendor that records will answer in one sentence.
• Watch trade sizes: pull up a time-and-sales or footprint view. Real tick data shows irregular, lumpy quantities — the market is made of odd-sized trades. Suspiciously uniform sizes and evenly spaced timestamps are the signature of sampling.
• Compare a bar to its contents: take one busy minute, sum the tick volumes inside it, and check the total against the minute bar’s printed volume. On recorded data the arithmetic closes exactly, because the bar was built from those very ticks.

That last check describes how Pix Connect is built: bars at every resolution are constructed from the same recorded tick archive they are checked against, so the chart you trade from and the ticks underneath it can never disagree. If you would rather work with the raw events yourself, the same recorded feed is available to your own code through the Pix APIs.

Key takeaways