Order block trading: finding blocks that are worth trading

What an order block is

An order block is the last candle that moved against the trend before a strong move began. A bullish order block is the last down candle before a rally; a bearish order block is the last up candle before a sell-off. The zone is the range of that candle, high to low (some traders use the body only, which gives a tighter zone and a lower fill rate).

The story told about it, in the smart money concepts (SMC) and ICT communities, is that large participants built their position in that candle, and that when price returns to it they will defend the area or fill the rest of their orders. It is fair to say plainly that nobody watching a retail chart can see this. Order blocks are a framework for describing where moves started, not evidence of institutional activity. Judge them the way you would any other zone: by testing whether trading them, with fixed rules, gives an acceptable result.

Seen without the story, an order block is a precise version of an old idea. Price often comes back to the base of a move before continuing, which is the same observation behind supply and demand zones. The order block simply narrows the zone to one candle, which tightens the stop and makes the setup easier to define and test.

How to tell a valid order block from any candle

Because every move has a last opposing candle, the rules for which ones count are the whole strategy. A block worth trading passes four tests.

1. Displacement after it. The move away from the candle is fast: large candles, small wicks, covering much more distance than recent bars. A slow drift away from a candle does not make it a block.
2. A break of structure. The move closes beyond the last swing high (bullish block) or swing low (bearish block). This is the test that removes most candidates. See break of structure for how to mark swings consistently.
3. An imbalance left behind. The best blocks have a fair value gap just above (or below) them, a sign the move was one-sided. Not compulsory, but a strong tiebreaker.
4. Unmitigated. Price has not yet returned to the zone. Once it has traded back into the block, the block is "mitigated" and should be crossed off.

In the gold chart above, the last down candle before the rally runs from $2,387.50 to $2,395.00. The rally breaks the $2,408 high and leaves a gap between $2,395 and $2,401, so the block passes all four tests. Price runs to $2,421, then falls back to touch the top of the block. A buy limit at $2,395 fills, the stop sits at $2,385 below the block, and the target is the $2,421 high, $26 away against $10 of risk.

Entry, stop and target

Entry. The simplest entry is a limit order at the near edge of the block (the top of a bullish block), which ICT traders call the proximal line. It fills on any touch. A limit at the block's midpoint, the "mean threshold", gives a better price and a tighter stop but misses trades that only tag the edge. A third option is to wait for price to reach the block and then enter on a lower timeframe break of structure, which filters out blocks that price slices through at the cost of a worse price.

Stop. Beyond the far edge of the block (the distal line), plus a buffer for the spread and for wicks. On gold that buffer matters: a dollar or two below the block is common, because round numbers and stop orders sit just past obvious lows. If price closes through the whole block, the zone did not hold, and there is no reason to stay in.

Size. The stop distance is set by the block, not by you. A thick 4-hour block on gold can mean a $15 stop; on EUR/USD a 1-hour block is often 10 to 25 pips. Size the trade from the stop so the loss is a fixed share of the account. The position size calculator does it in seconds, and for gold the pip calculator shows what each dollar move is worth at your size.

Target. The high (or low) the move reached before retracing is the obvious first target, because it is the nearest place where stop orders and fresh breakout orders collect. Require at least twice the risk to it. If the move away from the block was too short to offer that, the block is not worth trading however clean it looks.

When order blocks fail, and what a failed block becomes

Order blocks fail most often in three situations: when they point against the higher timeframe trend, when price has already visited them, and when the move away from them never broke structure. In each case the zone had less behind it than it seemed.

In this EUR/USD example a bullish block forms between 1.0819 and 1.0828, and the rally from it is real but tops out at 1.0858 without breaking anything on the 4-hour chart, which is falling. When price returns, a long at the top of the block would be stopped as a large candle closes below 1.0819. Price then rallies back to the underside of the old block, stalls at 1.0826 and falls again.

That last part is worth knowing. A block that is closed through often acts as resistance (for a failed bullish block) on the next retest. SMC traders call it a breaker block, and it is the same idea as old support becoming resistance. It is a separate setup with its own rules, covered in the academy lesson on breaker blocks; the point here is simply that a close through your block is information, not bad luck.

Two more honest weaknesses. First, marking blocks is less objective than it looks: two traders will disagree about which candle is "the last down candle" when there are several small ones in a row. Write a rule (for example, the last candle with a bearish body before the first displacement candle) and stick to it. Second, blocks on very low timeframes are cheap and plentiful, and the spread eats most of the edge. Stay on the 1-hour chart and above until you have tested something smaller.

How to backtest order block trading

Order blocks are only testable if the definition is fixed before you start. Write down exactly which candle counts, what qualifies as displacement and what counts as a break of structure, and do not adjust them mid-test.

1. Pick one market and timeframe, for example gold on the 1-hour chart, and go back at least six months.
2. Move forward bar by bar. Mark a block only once the break of structure has closed, never earlier.
3. Place the hypothetical limit, stop and target, then advance. Record unfilled orders as well as filled ones.
4. Log each result in units of risk, after the spread, and collect at least fifty trades.

Then test the filters one at a time on the same sample: blocks with and without a fair value gap, first touch only against any touch, with and without the daily trend filter. Look at the average result per trade and the worst losing run, not the share of winners. The risk of ruin calculator shows whether that losing run is survivable at your risk per trade.