Free trader tool
Do Order Blocks Work? 125,936 Measured Against Fake Ones
Every order block on 13 markets since mid-2022, on the 15-minute and 1-hour charts. How often price comes back, whether the block holds, and whether any of it beats a zone that was never an order block.
1 Jul 2022 – 18 Sept 2026 (the DAX from Mar 2025), CRTLAB's own 15-minute bars. An order block is the last opposing candle at the start of the move that broke a swing high or low.
Price came back to the block about as often as to a random zone. And once it did, the block failed just as often: 90.8% were closed through within 5 days, fakes 90.4%.
Reached the 50% (mean threshold) within 24 hours: 80.1% (fake 78.6%).
Limit order at the block's near edge for 24 bars, stop past the far edge, target 2R.
- Win rate
- 31.9%
- Expectancy
- −0.04R
- Fake zone
- −0.06R
Below break-even. −0.04R a trade over 72,430 trades before spread and commission; the same trade on the fake zones made −0.06R.
Opens the 15-minute chart with order blocks drawn as they form, the future hidden.
How the order block is worked out
Gold, 17 Sep 2026, 15-minute, bearish block at 10:00 New York · price came back 30 minutes after the break, then closed through it at 02:00. The calculator runs this for your own numbers, instrument by instrument, as you type.
Calling a revisit proof the block mattered
On 15-minute charts price came back to 86% of order blocks within a day, and to 85% of fake zones the same size and distance away. Once it came back, 91% of blocks were closed through within five days.
The short answer
Price comes back to order blocks constantly. Of 99,428 order blocks on 15-minute charts, 86.4% were revisited within 24 hours, a median 75 minutes after the break of structure. On the 1-hour chart it was 73.9% of 26,508.
Then we drew a fake zone for every block: the same size, the same distance from price, on the same side, at a random time in the same hour of the day, with no order block anywhere near it. Price came back to the fake zones 85.0% of the time. And when it did come back, the real blocks failed just as often: 90.8% were closed through within five days, against 90.4% of the fakes.
The textbook retest agrees. A limit at the block's near edge, stop past the far edge, 2R target, taken 72,430 times, won 31.9% and averaged −0.04R a trade before costs. The same trade on the fake zones averaged −0.06R. An order block marks where a move started. It does not, by itself, make price respect it.
How often price comes back to an order block
A block counts as revisited when price trades back to its near edge (the high of the last down candle before a rally, the low of the last up candle before a drop), as reaching the mean threshold at its 50%, and as failed when a candle closes beyond its far edge, measured from the bar after the break of structure.
| All 13 markets | Within 1 hour | 4 hours | 24 hours | 5 days |
|---|---|---|---|---|
| 15-minute: revisited | 44.2% | 67.4% | 86.4% | 94.0% |
| 15-minute: reached 50% | 25.5% | 53.4% | 80.1% | 91.1% |
| 15-minute: closed through | 9.2% | 33.5% | 68.1% | 85.3% |
| 1-hour: revisited | 23.5% | 46.4% | 73.9% | 87.9% |
| 1-hour: reached 50% | 8.3% | 28.3% | 62.2% | 82.3% |
| 1-hour: closed through | 1.8% | 11.4% | 42.9% | 71.4% |
Order blocks are everywhere: every market printed about 5–8 a day on the 15-minute chart. And most of them do not survive the day. Within 24 hours of the break, 68.1% of 15-minute blocks had a candle close straight through them; within five days, 85.3%. That is the breaker block, and on these numbers it is the usual ending, not the exception.
The test that matters: a zone that was never an order block
"Price respected the order block" sounds like evidence until you ask how often price respects any level that close. So for every block we built a placebo: a zone of exactly the same height, the same distance from the price at the moment of the break, on the same side, placed at a random bar of the same market and timeframe within 30 days, at the same New York hour so the session's volatility matches. No swing, no break, no last opposing candle. Then we walked both forward and traded both the same way.
| 15-minute | Revisited in 24h | Closed through in 24h | Failed after the revisit | Retest, 2R |
|---|---|---|---|---|
| Order blocks | 86.4% | 68.1% | 90.8% | −0.04R ± 0.02 |
| Fake zones | 85.0% | 66.5% | 90.4% | −0.06R ± 0.02 |
| 1-hour | Revisited in 24h | Closed through in 24h | Failed after the revisit | Retest, 2R |
| Order blocks | 73.9% | 42.9% | 81.3% | −0.06R ± 0.02 |
| Fake zones | 72.1% | 41.6% | 80.5% | −0.07R ± 0.02 |
The rows are close to identical. Price came back to real order blocks a point or two more often than to fake zones, held no better once it arrived, and the trade finished in the same place. Across every split below, by direction, size, displacement and session, real blocks and fake zones never differ by more than 3.6 points in how often price came back within a day.
Displacement, size and session
ICT teaches that a real order block leaves displacement behind: the move away from it is fast enough to leave a fair value gap. We split every block on exactly that.
| 15-minute blocks | Blocks | Revisited in 24h | Fake zone | Retest, 2R | Fake zone |
|---|---|---|---|---|---|
| With a fair value gap | 44,751 | 82.8% | 81.2% | −0.03R ± 0.02 | −0.07R |
| Without one | 54,677 | 89.3% | 88.1% | −0.05R ± 0.02 | −0.05R |
| Small (under 0.75 ATR) | 22,161 | 87.4% | 86.1% | −0.05R ± 0.02 | −0.10R |
| Medium (0.75–1.5 ATR) | 52,440 | 86.7% | 85.2% | −0.05R ± 0.02 | −0.05R |
| Large (1.5+ ATR) | 24,827 | 84.9% | 83.6% | −0.02R ± 0.02 | −0.04R |
| Asia (18:00–02:00) | 33,538 | 89.2% | 87.7% | — | — |
| London (02:00–07:00) | 22,952 | 87.2% | 86.6% | — | — |
| New York AM (07:00–12:00) | 24,555 | 82.2% | 80.7% | — | — |
| New York PM (12:00–17:00) | 18,383 | 86.0% | 83.7% | — | — |
Blocks with displacement were revisited less often (83% against 89%), which looks like strength until you read the fake-zone column: it falls by the same amount, because a fast move simply leaves its starting point further behind. On the 15-minute chart the displacement blocks retested for −0.03R against −0.07R for their fake zones; on the 1-hour chart the order flips, −0.07R against −0.05R. A difference that changes sign between timeframes is noise. Large blocks hold more often than small ones, and so do large fake zones: the far edge is simply further away.
The order block retest trade
We fixed the trade before looking at any result, in the form it is usually taught. When the break of structure closes, place a limit order at the block's near edge and leave it working for 24 bars. The stop goes past the far edge, the extreme of the move, because a block that is closed through has failed. The target is 2R, with 1R as a variant, and a second version enters at the 50% mean threshold instead. A trade still open after five days is closed at the market. A bar that touched both stop and target is scored as a loss, and the best-case column shows the result had every one of those been a win.
| All 13 markets | Trades | Win rate | Expectancy | Best case | Same trade, fake zones |
|---|---|---|---|---|---|
| 15-minute, near edge, 2R | 72,430 | 31.9% | −0.04R ± 0.02 | −0.02R | −0.06R |
| 15-minute, near edge, 1R | 72,430 | 46.4% | −0.07R ± 0.00 | −0.03R | −0.11R |
| 15-minute, mean threshold, 2R | 60,565 | 28.1% | −0.16R ± 0.02 | −0.07R | −0.20R |
| 1-hour, near edge, 2R | 19,579 | 31.3% | −0.06R ± 0.02 | −0.01R | −0.07R |
| 1-hour, mean threshold, 2R | 16,492 | 26.5% | −0.21R ± 0.02 | −0.08R | −0.24R |
A 2R target only needs to win one trade in three to break even, and this won 32%. Close, which is exactly what a price with no memory of the block would produce: the fake zones won 31%. Entering deeper, at the mean threshold, made it worse rather than better (−0.16R), because a block that price has already pushed halfway into is usually on its way through. None of this includes spread or commission, which only push every row further below zero.
Every market
| Market | Blocks a day | Typical block | Revisited in 24h | Fake zone | Retest, 2R | Fake zone |
|---|---|---|---|---|---|---|
| Gold (XAU/USD) | 6 | $8.60 | 86.2% | 85.3% | +0.01R ± 0.04 | −0.04R |
| Nasdaq 100 | 7 | 37 pts | 86.9% | 86.3% | −0.03R ± 0.04 | −0.04R |
| Dow Jones (US30) | 6 | 43 pts | 87.4% | 86.6% | −0.07R ± 0.04 | −0.08R |
| S&P 500 | 6 | 6.8 pts | 86.5% | 85.9% | −0.04R ± 0.04 | −0.06R |
| DAX (GER40) | 6 | 31 pts | 85.9% | 85.3% | −0.06R ± 0.06 | −0.09R |
| EUR/USD | 6 | 6.0 pips | 86.3% | 85.0% | −0.06R ± 0.04 | −0.01R |
| GBP/USD | 6 | 7.5 pips | 85.9% | 84.4% | −0.05R ± 0.04 | −0.05R |
| USD/JPY | 5 | 10 pips | 85.2% | 83.2% | −0.04R ± 0.04 | −0.06R |
| AUD/USD | 6 | 4.8 pips | 86.3% | 85.0% | −0.04R ± 0.04 | −0.04R |
| Silver (XAG/USD) | 6 | $0.27 | 86.5% | 84.8% | −0.04R ± 0.04 | −0.04R |
| Crude oil (WTI) | 6 | $0.25 | 86.6% | 85.0% | −0.02R ± 0.04 | −0.02R |
| Bitcoin | 8 | $184 | 86.8% | 84.8% | −0.06R ± 0.02 | −0.09R |
| Ethereum | 8 | $6.89 | 86.2% | 83.9% | −0.05R ± 0.02 | −0.10R |
Thirteen markets, one result. The retest came out between −0.07R and +0.01R on every one, none of them clearly above zero, and the fake zones landed in the same range. The 15-minute figures are shown; the tool above switches to the 1-hour.
What this means if you trade order blocks
- Use the block as a map, not a wall. It tells you where the move that broke structure started. It does not tell you price will turn there, and 91% of revisited 15-minute blocks were closed through within five days.
- "It respected the OB" is not evidence. A random zone the same distance away gets respected just as often. Only a count against a baseline means anything.
- Deeper is not better. The mean-threshold entry was the worst version tested (−0.16R).
- Test what you add, against this baseline. The bare retest is worth about −0.06R more than nothing, which is to say nothing. If your edge is an order block after a liquidity sweep, inside a kill zone, or in the direction of the daily bias, measure that version and compare it to these numbers.
That is what the order block backtester is for: blocks and breakers are drawn on the chart as they form, you replay any day one candle at a time with the future hidden, apply your own rules and log every trade. The order block and breaker block reference entries cover the definitions, and the fair value gap, CRT and previous day high and low statistics run the same kind of test on other ICT setups.
How we measured it
CRTLAB's own 15-minute price history, 1 Jul 2022 – 18 Sept 2026, for every market except the DAX (from 11 Mar 2025), and 1-hour bars built from it on clock hours. A swing high is a bar whose high is above the two bars on either side of it; it exists once those two later bars have printed. The first candle to close above the most recent swing high is the break of structure, and that swing is then used up. The move that broke it starts at its lowest low since the swing. The order block is the last down-close candle at that low or up to two bars before it, and its zone runs from that candle's high down to the low of the move. Bearish blocks are the mirror. Breaks with no opposing candle there (13% of them on the 15-minute chart) have no order block and are left out, as are zones under a tenth of the average true range.
Everything is measured from the bar after the break, in clock time: 1 hour is four 15-minute bars or one 1-hour bar. The fake zones are seeded, so the numbers repeat exactly on a re-run. No spread or commission: this measures the market, not a broker. The definitions, the placebo, the trade and every split were fixed before the results were seen.
Frequently asked questions
What is an order block?
The last opposing candle before the move that broke structure: the final down-close candle at the bottom of a rally that closed above a swing high, or the final up-close candle at the top of a drop that closed below a swing low. Traders mark it as a zone and expect price to react when it comes back. On the 15-minute chart a typical recent order block was $8.60 on gold, 37 pts on the Nasdaq 100 and 6.0 pips on EUR/USD, and every market printed about 5–8 of them a day.
Do order blocks work?
Not on their own. Of 99,428 order blocks on 15-minute charts across 13 markets, price came back to 86.4% within 24 hours, and to 85.0% of fake zones the same size and distance from price. The textbook retest (limit at the near edge, stop past the far edge, 2R) won 31.9% and averaged −0.04R a trade before costs; on the fake zones it averaged −0.06R. The order block itself added nothing measurable.
How often does price return to an order block?
Most of the time, and quickly: 67.4% of 15-minute order blocks were revisited within four hours and 86.4% within a day, a median 75 minutes after the break of structure. On 1-hour charts it was 73.9% within a day. Fake zones drawn the same distance from price were revisited 85.0% and 72.1% of the time, so the return is about distance, not the block.
Do order blocks hold when price comes back?
Rarely for long. Once price got back to a 15-minute order block, a candle closed through its far edge within five days 90.8% of the time, against 90.4% for the fake zones. Within 24 hours of the break of structure, 68.1% of 15-minute blocks had already been closed through.
What is the mean threshold of an order block, and is it a better entry?
The mean threshold is the 50% level of the order block. Price reached it within a day on 80.1% of 15-minute blocks (fake zones 78.6%). Entering there instead of at the near edge, with the same stop and a 2R target, was worse: −0.16R a trade against −0.04R, because the stop is closer and a block that is back to its middle usually goes through.
Are order blocks with displacement (a fair value gap) better?
Barely, and not reliably. On 15-minute charts, blocks whose move left a fair value gap retested for −0.03R against −0.07R for their fake zones; on 1-hour charts it was −0.07R against −0.05R. A difference that flips sign between timeframes is noise, and neither is above zero.
What is the difference between an order block and a breaker block?
A breaker is an order block that failed: price closed through it, and traders then expect it to act from the other side. On these numbers that is the normal outcome, not the exception: 85.3% of 15-minute order blocks were closed through within five days. The breaker block reference entry covers the definition.
Embed this calculator on your site
Free for any trading blog, forum or resource page — no signup, no key, nothing to break. Paste this snippet where you want the calculator to appear. The credit line underneath is the only thing we ask for.
<iframe src="https://crtlab.pro/embed/order-block-statistics" width="100%" height="1080" style="border:0;max-width:720px" title="Order Block Statistics by CRTLAB" loading="lazy"></iframe>
<p style="max-width:720px;font:14px/1.5 system-ui,sans-serif;opacity:.7;margin:.5rem 0 0">
<a href="https://crtlab.pro/tools/order-block-statistics">Order Block Statistics</a> by <a href="https://crtlab.pro">CRTLAB</a>
</p>Find out if your order blocks beat a fake one.
Replay real market history candle by candle and measure your setup across a proper sample, on 38 markets and every timeframe.
Open the chart