A breakout crosses an established chart boundary. A retest brings price back toward that area. Define what counts as a break and a hold before the return happens, then compare the entry available after it with your original levels.
View full-size diagram
Scroll sideways to explore the diagram.
Mark the area before the exciting candle arrives
A breakout is easier to judge when the boundary was already meaningful. Start with an area where the chart has previously changed direction, rather than drawing a line under the newest rally after it happens.
Fidelity explains how broken resistance can become support, and how broken support can become resistance. A retest examines that possible role change. Crossing the old line does not, on its own, show that the new role will hold.
In this hypothetical hourly example, earlier highs cluster between 100 and 101. Mark the whole band. A wick to 100.70 is still inside it; an hourly close at 102.20 is outside. Choosing the band first makes those two observations easy to distinguish.
Separate the break, the return and the response
After closing at 102.20, price continues to 104. It then pulls back to 101.20, near the former ceiling. That return is the retest attempt. You still need to see how the chart behaves there.
| Stage | Hypothetical observation | What it establishes |
|---|---|---|
| Break | Hourly close at 102.20. | Price closed above the 100–101 band. |
| Return | Pullback reaches 101.20. | Price revisited the boundary area. |
| Response | Next hourly candle closes at 102.50. | A completed candle recovered from that test. |
For this example, you might define a hold as that recovery close without an intervening hourly close below 100. Another rule would produce different observations. Record yours before the pullback instead of switching from a close to a wick when the chart becomes inconvenient.
The confirmation changes the price you can use
Suppose the proposed target is 108 and the hypothetical protective stop is 99.50. An assumed entry at 101.50 risks 2 to seek 6.50, giving 3.25 R before costs.
But that entry is not automatically available after the recovery candle closes at 102.50. Using 102.50 instead means planned risk of 3 and potential reward of 5.50, or about 1.83 R. The setup’s outline is similar; the arithmetic is materially different.
That is the tradeoff to examine when waiting for more evidence. Use the risk-reward calculator to compare the prices and costs. Do not move the target farther away just to recover the earlier ratio.
Sometimes the return fails. Sometimes it never comes.
If the hourly chart closes back at 99.70, the example’s hold condition has failed. Price is back below the original band. Keep that observation in your notes rather than continuing to describe 100–101 as confirmed support.
Alternatively, price might move directly from 104 toward 108 without returning. A plan that required a retest then has no qualifying entry. That is a distinct outcome, not a reason to pretend the missing pullback happened.
For a bearish version, reverse the directions: a support band breaks, price returns from below and the recovery stalls. The double-top guide shows one way that sequence can arise around a neckline.
Show the report the part that matters
Include the original range, breakout candle and return in the screenshot. Keep the interval and timestamps visible. A crop showing only the bounce can hide whether the level had any prior significance.
Check execution assumptions as well. A stop level is a planned reference; Investor.gov explains why a triggered stop order can fill at a different price. That matters when the retest moves quickly.
For XAUUSD, the gold chart workflow helps identify the correct instrument and units. Then analyze your screenshot and compare the report’s breakout, retest and invalidation with the sequence you recorded.
Your chart. A more structured read.
See how TradeGPT turns a chart screenshot into levels, competing scenarios and an execution plan. Start with the full sample, then bring your own chart.
The sample is open to everyone. Analyzing your own chart requires a subscription.
Sources & further reading
References checked September 8, 2026. Price examples and diagrams in this guide are hypothetical.