Crypto chart pattern recognition

What a market that never closes changes

On a stock chart the trading day starts and stops. On a crypto spot or perpetual chart, every bar is a full slice of continuous trading, so 80 one-hour bars always mean 80 hours. Bar counts translate cleanly into clock time, and the session gaps that shape so many stock patterns, where price jumps between one day's close and the next day's open, rarely appear. Gaps on crypto-related charts come mostly from products that do close, such as US-listed bitcoin and ether ETFs, which the stock chart pattern guide covers.

A "daily close" in crypto is a convention. Each exchange ends its daily bar at a fixed time of its choosing, commonly 00:00 UTC though not everywhere, so a daily breakout is a snapshot taken at an arbitrary hour, not the end of a trading day. A move that looks decisive on the daily chart may have happened in the last few minutes before the rollover.

Weekend liquidity and thin books

Trading never stops, but participation varies. Weekends, holidays and the quiet hours between the main trading regions tend to have thinner order books. In a thin book a modest market order moves price further, so a breakout can be produced by a small amount of buying or selling, and it can fade when weekday volume returns.

Thin periods also distort relative volume. A volume test compares the breakout bar with recent bars; on a quiet Sunday the recent bars are quiet too, so a bar with 1.5 times the recent average can still be small in absolute terms. An absolute liquidity floor is a useful second check.

Before trusting a weekend breakout, ask:

• Is 24-hour volume in line with the coin's usual weekday level, or a fraction of it?
• Is the spread wider than normal?
• Did the move come from one large bar on an otherwise quiet chart?
• Would you take the same trade, at the same size, on a Tuesday?

Wicks, stop runs and liquidation spikes

Crypto charts are known for long wicks: price spikes through a level and snaps back within minutes. Three things feed them: thin books, clusters of stop orders resting just beyond obvious levels, and forced liquidations on perpetual futures, which hit the market as price-insensitive orders and can cascade. The most visible levels therefore attract false breakouts: the neckline everyone can see is also where everyone's stops sit.

Two practical consequences follow. First, a breakout has to hold, not just touch: a wick through resistance is not the same as trading above it. If a tool evaluates the latest bar while it is still forming, as PatternX does, a breakout can fade back inside the pattern before the bar is finished, so look at the bar before approving a signal. Second, a stop placed exactly at an obvious level is exposed to the same wicks. How the exchange measures price for stops and liquidation matters too; the perpetual futures guide explains mark price.

One coin, many charts: exchange fragmentation

The same coin trades on many exchanges at once, each with its own order book. Arbitrage keeps prices close but not identical: highs and lows differ slightly, a wick can print on one exchange and not another, and volume is split unevenly. Pairs may be quoted in USDT, USDC, USD or a local currency, which adds small differences of its own, and reported volume is not equally reliable everywhere.

For a pattern defined by precise levels, small differences matter. For example, PatternX's double top detector requires the two peaks to be within 2.5% of each other. If the peaks are 2.3% apart on one exchange and 2.7% apart on another, the first chart shows a double top and the second shows none.

The practical rule is to read the pattern from the exchange and instrument you will trade. PatternX works this way: it fetches 200 bars separately for each exchange, symbol and timeframe, measures liquidity, spread and the order book on that exchange, and its historical win-rate input uses PatternX's live results on that exchange and symbol once there are at least 10 closed trades.

Perpetual futures versus spot charts

A perpetual future tracks its coin's spot price through funding payments, but it trades on its own order book, so the two charts are close relatives rather than copies. In fast markets the perpetual can lead and trade at a premium or discount to spot. Perpetuals also carry data that spot markets lack, the funding rate and open interest, and they make short selling straightforward, so bearish patterns such as a head and shoulders or a bear flag can be traded directly. On a spot market you can only sell what you hold.

Liquidations are the other difference. They add forced orders to perpetual markets, one reason perpetual wicks can be longer than spot wicks on the same coin.

Which timeframes suit crypto

No timeframe is right for everyone; each trades signal count against noise and costs. Detectors look back a fixed number of bars, so the timeframe decides how much clock time a pattern can span. For example, an 80-bar window, the length PatternX's cup and handle and ascending triangle detectors read, covers:

Costs scale differently too. For example, with fees of 0.05% per side and a 0.05% spread, a round trip costs about 0.15%: 15% of a 1% target, but 2.5% of a 6% target. Lower timeframes produce smaller targets, so costs take a bigger share.

PatternX supports eight timeframes from 1m to 1d and starts new users on 5m, 15m and 1h. It also checks the next timeframe up, but that check only changes the score and never rejects a signal. Formations larger than a detector's window are not seen at all, so a months-long base belongs on the 4h or daily chart. For the method, see multi-timeframe analysis.

How automated recognition handles crypto

An automated scanner applies the same checks to every bar of every coin it tracks, including at hours when nobody is watching. PatternX's approach to crypto in brief:

• Coverage: 40 crypto exchanges. PatternX is built for perpetual futures; 8 spot-only exchanges are available for data, paper and manual use, and Aster, Hyperliquid and dYdX are decentralised exchanges connected with wallet authentication.
• Universe: you choose the coins to track, and the coin picker offers the top 20 by 24-hour volume on each exchange. Scans repeat every 10 to 120 seconds depending on plan, so this is not tick-by-tick monitoring.
• Filters: every pattern must pass universal checks, including breakout-bar volume of at least 1.2 times the 20-bar average, a spread of at most 0.25%, a liquidity score of at least 30 and order-book imbalance against the trade no worse than −0.5. Crypto adds a 24-hour volume floor of $3 million.
• Score inputs: the PatternX Score weighs order-book imbalance (6%), funding (3%) and open interest (2%) alongside pattern geometry, volume and trend.
• AI context: on Pro and Elite, PatternX AI evaluates setups the engine has already detected and can veto them or re-weight their confidence. Its sentiment factor is the perpetual-futures funding rate and nothing else.

Limits and common mistakes

• Treating a weekend breakout like a weekday one without checking absolute volume.
• Reading a pattern on one exchange and trading it on another.
• Assuming spot and perpetual charts are interchangeable.
• Assuming a scanner sees everything: detectors have fixed look-back windows, scans run at intervals and only tracked coins are scanned. See limitations.
• Forgetting that filters reduce the risk of thin markets but do not remove it: a coin above the volume floor can still move straight through a stop during a liquidation cascade.