Your Market Structure school taught that price is the last match in an auction. This chapter is about how that match reaches your screen — because the journey manufactures the thing you stare at all day, and knowing the manufacturing process dissolves a dozen daily confusions.
The raw material: ticks. A tick is one event at the exchange — a trade, or a change in the best bid/ask. Liquid instruments generate thousands of ticks per second in fast markets. No retail screen shows them all; every platform samples.
The two delivery methods — and why apps disagree. A streaming feed pushes updates continuously (websocket connections — a pipe held open between server and app). A snapshot feed asks "what's the price now?" every fixed interval — say once per second — and shows you that. Two apps, one streaming and one snapshotting (or snapshotting at different moments), will flicker differently and briefly disagree — both correctly reporting different samples of the same firehose. Add each app's own batching (grouping updates to save your phone's battery and data) and network latency (Market Structure school, Chapter 9), and the mystery inverts: the surprise isn't that apps disagree for a second — it's that they agree at all.
Delayed vs. real-time — read the fine print. Exchanges charge for real-time data; many websites and some free tools show prices delayed (often by several minutes, per exchange policy) or snapshot-refreshed. For an investor checking a portfolio, irrelevant. For anyone timing entries, a delayed feed is a map of where the market was. Know which one each tool in your stack serves — it's always disclosed, rarely read.
The manufactured product: candles. A candle is a summary of all ticks in a window — first (open), highest (high), lowest (low), last (close). Two facts most traders never learn: (1) The live candle is a lie until it closes. A forming 5-minute candle rewrites itself with every tick — the "hammer" at 10:03 may be a plain red bar by 10:05. Every rule in your Market Structure school that says close-based (BOS confirmation, sweep verdicts) exists partly because of this: the close is the only part of a candle that's a fact. (2) Candles can differ across platforms. Different data vendors, different tick sampling, different session-boundary handling — the same stock's daily candle can show slightly different highs/lows on two platforms. Small, usually harmless, but it means backtesting on one feed and trading on another (Module 3) has a built-in fuzz factor.
Depth data — the premium tier. The standard feed shows 5 levels of the order book (your Market Structure school's window). Exchanges also sell deeper views (20 levels, or full tick-by-tick data) — mostly consumed by algos and institutions. Retail rarely needs more than 5 levels; what retail does need is to remember that even 5-level depth is a sample refreshed at intervals, not a live X-ray — the queue you see is already slightly history (latency, again).
The practical rules this buys you: trade decisions off closed candles unless your strategy explicitly handles forming ones; know each tool's feed type (streaming/snapshot/delayed) before trusting its prices for timing; expect small cross-platform disagreements and stop investigating them; and when precision matters (tight stops, thin instruments), your broker's own feed — the one connected to your actual orders — is the reference, because it's the only one your fills will be judged against.
Key Takeaway
"The price" on any screen is a manufactured sample: ticks, filtered through feed type, batching, and latency. Apps disagree because they sample differently — both honestly. Trust closed candles over forming ones, know each tool's feed type, and treat your broker's own feed as the reference for anything precise.
Think About It
Have you ever entered on a beautiful pattern in a forming candle that vanished by the candle's close? In this chapter's language — what did you actually trade: a fact, or a provisional sample?
Tech Lab — The Feed Audit
Open the same liquid stock in three places simultaneously: your broker's app, TradingView, and one financial website. For two minutes, note: which updates fastest, which lags, whether any is marked delayed (check the fine print near the price). Then list every tool in your actual workflow and label its feed: streaming / snapshot / delayed. Anything mislabeled in your head until today goes in the journal — that's a timing assumption you've been trading on without knowing it.