A coin can close above its recent high on one exchange and remain below its recent high on another. Before deciding that a chart is wrong, check whether both charts are answering the same question with the same market data.
This Read the Reaction lesson keeps the timeframe fixed and changes the venue. Our previous lesson compared fifteen-minute and hourly candles. Today, two fifteen-minute charts will disagree for a different reason: their local trading histories differ.
One coin does not mean one trade history
An exchange chart built from trades describes transactions in a particular market. Kraken explains that exchange prices reflect client activity, which can differ between venues. There is no requirement for every venue to print identical highs and closes.
A shared announcement can attract attention to the same token without producing identical orders everywhere. That is a possible explanation for different reactions, not evidence that a particular headline caused a particular candle. Our news-versus-technical lesson explains that distinction.
The comparison below concerns ordinary traded-price candles on two fictional spot exchanges. It does not compare a spot chart with a perpetual contract or an index.
Define the comparison before reading the result
Imagine the same fictional token quoted in USD on Venue A and Venue B, on an invented trading day. Both charts use fifteen-minute intervals with identical UTC boundaries and complete data. A hypothetical announcement arrives at 12:05 UTC; it provides context, not proof of causation.
We assess the candle starting at 12:00 and ending just before 12:15 UTC, after both venues have delivered its completed data. The four preceding candles start at 11:00, 11:15, 11:30 and 11:45 UTC.
Write this teaching rule before inspecting the outcome:
A breakout qualifies only when the completed candle closes strictly above the highest high of the previous four completed candles on that same venue.
The four-candle lookback is chosen to keep the arithmetic short. It is not a recommended setting or a product specification. A wick above the boundary is insufficient, equality fails, and the candle being assessed is excluded from the reference window. See range-breakout basics for the underlying concept.
Same rule, different local boundaries
Every price, venue and event in this example is hypothetical. In chronological order, the four prior highs are:
- Venue A: 99.20, 99.80, 100.00, 99.60. The maximum is 100.00 USD.
- Venue B: 99.30, 100.30, 99.90, 100.10. The maximum is 100.30 USD.
In the table, R is that prior range high, H the assessed candle's high, and C its completed close. All values are USD prices.
| Venue | R | H | C |
|---|---|---|---|
| A | 100.00 | 100.50 | 100.20 |
| B | 100.30 | 100.40 | 100.10 |
Venue A qualifies: 100.20 > 100.00. Its close is 0.20% above its own reference high: (100.20 − 100.00) ÷ 100.00 × 100.
Venue B does not qualify: 100.10 < 100.30. Its close is approximately 0.20% below its own reference high, rounded to two decimal places. Its high of 100.40 crossed the boundary, but that does not satisfy a closing-price rule.
Neither calculation establishes a better exchange or a profitable trade. The two results describe different local observations under the same rule.
The reference level matters as much as the latest price
Now change only Venue B's close to 100.20, leaving its previous highs unchanged. Both venues would have the same close, yet only A would qualify: B would still be below 100.30. Matching the latest price does not make their histories identical.
Applying A's 100.00 boundary to B's original 100.10 close would produce a positive result. But that is an external-reference rule, not the venue-specific rule we wrote. Using another venue as a reference can be a deliberate research choice; record it explicitly and test that exact process.
Before treating a mismatch as economic information, check for mundane data differences. Are both candles complete? Do the windows contain the same timestamps? Is either feed stale or missing intervals? Coinbase's candle documentation notes that historical candles can be incomplete and that no-trade intervals may be absent. Missing data should not silently become a different lookback.
Use underlying numerical values, not rounded chart labels, for a strict inequality. Also confirm the actual asset and quote currency: a USD market and a stablecoin-quoted market are not automatically identical comparisons.
A chart gap is not an executable profit
The original closing prices differ by 0.10 USD. That does not establish an opportunity to buy at B's close and sell at A's close. Kraken distinguishes a historical last-traded price from the bid and ask available to a new order.
For a separate hypothetical snapshot after those closes, suppose A's best bid is 100.00 and B's best ask is 100.15. Buying on B and selling on A would start with a quoted difference of 100.00 − 100.15 = −0.15 USD per token, before fees. The historical 0.10 gap does not survive even this first quote check. Available size, price changes and slippage still matter; displayed quotes are not promised fills.
Likewise, A's qualifying close could be followed by a close at 99.90, below its original 100.00 boundary. A valid observation can lead to a failed continuation. Qualification is not a success forecast.
A checklist for comparing venues
Before explaining a missing breakout, save:
- Market identity: the same asset, quote currency and instrument type, with each venue named.
- Data basis: traded-price candles, matching boundaries, complete windows and receipt times.
- Rule: lookback, strict comparison, close versus wick, and whether the reference comes from the trading venue or elsewhere.
- Evidence: each venue's prior highs and candidate close, before changing any settings.
- Execution plan: current quotes, order size, costs, invalidation and a separate risk budget.
If you require both venues to qualify, define that condition in advance. Agreement is not automatically independent confirmation, and disagreement alone does not prove manipulation or tell you which venue will move next.
Educational content, not financial advice. All numerical examples and events are hypothetical. A breakout condition does not guarantee continuation or profit.
