I lost 500 dollars fading an extension the first time a crab formed on my chart, and I lost it because I entered a full leg too early. Price had already run far past the start of the move, every instinct said it had gone too far, and I sold into the push with a tight stop and a confident feeling. Price paused for two candles, took out my stop, then ran another hundred pips to a level I had not even drawn, turned there, and fell 300 pips without me. The level it turned at was the 1.618 extension of the original leg, and that day taught me the one thing that separates the crab from every other harmonic pattern I trade: this pattern does not complete where the move started, or even near it. It completes a long way beyond it, at the deepest extension in the harmonic family, and the only traders who get paid on it are the ones who wait for that address instead of guessing it early. Everything in this article runs on one account model: a 100,000 dollar account, a 5 percent daily loss limit of 5,000 dollars, and a personal rule of 500 dollars of risk per trade, which is 10 percent of the daily limit.
In this article
- What a crab actually is
- The five points and their exact ratios
- Crab against the other harmonics I already trade
- The six rules I trade a crab by
- Position sizing under a daily loss limit
- The worked trade, candle by candle
- Fifteen crabs, filtered against unfiltered
- When the crab fails
- My daily crab routine
- Frequently asked questions
The crab is a five-point harmonic reversal pattern, discovered by Scott Carney in 2000, built from four legs labelled X, A, B, C and D. Its defining feature is the final leg: point D sits at the 1.618 extension of the XA leg, beyond point X, confirmed by a second projection from the BC leg of 261.8 to 361.8 percent. No other harmonic pattern on this site stretches that far. That extreme extension is both the opportunity and the trap. It produces the largest reversal zones and the biggest reward-to-risk ratios in the family, and it produces the most early entries from traders who sell the first sign of exhaustion instead of waiting for the zone. This article gives you the exact ratios, the six rules I trade it with, a worked bearish crab with the full sizing maths, a twelve-trade filtered test, and the failure modes that cost me money before I filtered them out.
What a crab actually is
A crab is a reversal pattern that forms at the end of an exhausted move. In a bearish crab, price falls from X to A, retraces up to B, dips to C, then pushes to a new extreme at D, beyond X, where sellers finally take control. In a bullish crab the whole shape is inverted: price rises from X to A, retraces down to B, pushes up to C, then drops to a new extreme at D, below X, where buyers step in. The shape alone means nothing. What makes a crab valid is the arithmetic between the legs, and the arithmetic is strict: B must retrace only part of XA, C must retrace part of AB, and D must land where two independent Fibonacci projections agree. If any leg breaks its ratio band, the structure is not a crab. It is a random swing with a crab-shaped story attached, and trading it as a crab is how accounts get hurt.
The psychology is the same story as every harmonic pattern, pushed one step further. The XA leg is the original impulse. The AB retracement is profit-taking. The BC leg is the market testing whether the move has anything left. The CD leg is the final exhaustion push: late traders pile in chasing the trend, early reversal traders get stopped out as price breaks past X, and their stops feed the last burst of momentum. Point D is where that fuel runs out. The traders who were stopped out are gone, the chasers are fully committed at the worst possible price, and the first strong candle back inside the zone is the signal that the crowd has flipped. A crab trade is not a prediction that price will reverse. It is a bet, placed only inside a measured zone, that an exhausted crowd has just run out of buyers or sellers.
The five points and their exact ratios
Measure every crab from the same anchor points, in the same order, and never from a level you moved to make the pattern fit. X is the start of the impulse, A is the end of it, B is the first retracement extreme, C is the second, and D is the completion point. The table below is the checklist I run before I allow myself to call anything a crab. One failed row and the pattern goes back on the watchlist, untraded.
| Leg or point | Required ratio | What it means on the chart |
|---|---|---|
| XA | The impulse leg, any size | The move the whole pattern is measured against |
| B point | 38.2 to 61.8 percent of XA | A shallow retracement, never a deep one |
| C point | 38.2 to 88.6 percent of AB | A partial retracement of the AB leg |
| D point, projection 1 | 161.8 percent extension of XA | The defining level of the reversal zone |
| D point, projection 2 | 261.8 to 361.8 percent projection of BC | The confirming level, must overlap projection 1 |
| Potential reversal zone | Both projections within 60 pips (H1 rule) | The tradable zone, not a single line |
| Invalidation | Close beyond the 2.0 extension of XA | The pattern is dead, exit any open trade |
Two rows deserve emphasis because they are where most crab trades go wrong. First, the B point must stay at or below the 61.8 percent retracement of XA. If B retraces 78.6 or 88.6 percent, you are not looking at a crab at all: a B at 88.6 percent with a D at the 1.618 extension is a deep crab, a different variant with its own rules, and a B at 78.6 percent points toward a butterfly instead. Second, the reversal zone is the overlap of the two D projections, not either one alone. A 1.618 extension with no BC projection landing near it is a single line, not a zone, and single lines do not reverse markets reliably. My rule is that the two projections must land within 60 pips of each other on the hourly chart, and the tighter the overlap, the more I trust the zone.
There is also an optional confluence worth knowing: the alternate AB=CD structure, where the CD leg measures 1.27 or 1.618 times the AB leg. It is not required for validity, but when it lands inside the same zone as the two main projections, I treat the zone as stronger and I am more willing to take the full position size rather than a half position. When it lands far away, I ignore it. It never overrides a failed ratio row.
Crab against the other harmonics I already trade
The crab looks like its cousins, and that is exactly why traders confuse them. The difference is always in the numbers, never in the shape. The table below lines the crab up against the harmonic patterns already covered on this site, so you can tell at a glance which pattern you are actually looking at when a five-point shape completes.
| Pattern | B retracement of XA | Where D completes | How it differs from the crab |
|---|---|---|---|
| Gartley | 61.8 percent | 78.6 percent of XA, inside the move | D never passes X; a shallow, inside reversal |
| Bat | 38.2 to 50 percent | 88.6 percent of XA, inside the move | D stays inside X; the crab breaks through X |
| Cypher | 38.2 to 61.8 percent of XC | 78.6 percent of XC, inside the move | Measured off XC, not XA; D stays inside |
| Butterfly | 78.6 percent | 127.2 percent extension of XA, beyond X | Deep B and a shallower extension than the crab |
| Crab | 38.2 to 61.8 percent | 161.8 percent extension of XA, beyond X | Shallow B, deepest extension, extreme BC projection |
| Deep crab | 88.6 percent | 161.8 percent extension of XA, beyond X | Same D as the crab, but a deep B; a separate variant |
Read the table as a decision tree. If D completes inside the original move, it cannot be a crab, no matter how stretched the shape looks. If D extends beyond X, check B: a B at 78.6 percent says butterfly, a B at 88.6 percent says deep crab, and a B at 61.8 percent or less with a 1.618 extension says crab. The single most expensive confusion in this family is calling a deep crab a bat. The two share nothing except a vague shape: the bat completes inside X at 88.6 percent, while the deep crab breaks through X to the 1.618 extension. Traders who buy a deep crab as if it were a bat enter far too early, with a stop placed for an inside pattern, and the extension leg runs straight through them. If you only remember one line from this section, remember that one.
For the full rules of the nearest neighbours, see the butterfly pattern strategy, the Gartley pattern strategy, the bat pattern strategy, the Cypher pattern strategy, and the ABCD pattern strategy guides on this site.
The six rules I trade a crab by
Rules are what turn a pattern into a strategy. A crab that passes the ratio table still has to pass these six rules before I risk money on it, and the rules are ordered the way I check them: structure first, zone second, confirmation third, and risk last, because risk maths on an unconfirmed pattern is just a faster way to lose.
| Rule | The rule | Why it exists |
|---|---|---|
| 1 | B retraces 38.2 to 61.8 percent of XA, no more | A deeper B changes the pattern into a butterfly or deep crab |
| 2 | Both D projections overlap within 60 pips on H1 | A zone needs two agreeing measurements, not one line |
| 3 | Wait for a confirmation candle closing back inside the zone | Entering on the touch alone catches falling knives |
| 4 | Stop beyond the zone plus a buffer, never past invalidation | The stop must survive normal noise but respect the dead level |
| 5 | Risk at most 500 dollars, 0.5 percent of the account | One crab loss must never threaten the daily limit |
| 6 | Only London and New York session completions, no red-news entries | Thin-session zones and news spikes fake more crabs than any other cause |
Rule 3 is the one I broke in the story that opened this article. Selling the touch feels decisive and looks brave, and it is the fastest way to donate to the extension leg. A crab’s CD leg is violent by nature: it is the leg that hunts stops, and it routinely overshoots the zone by a few pips before reversing. Waiting for one hourly candle to close back inside the zone costs a few pips of entry price and filters out most of the overshoots. Rule 6 exists for the same reason. A crab that completes at 3am in a thin Asian session, or two minutes before a red-news release, has not been tested by real liquidity. I let those zones sit until London opens or the news passes, and if the zone is gone by then, it was never mine to trade.
Position sizing under a daily loss limit
The crab’s stop is wider than most patterns on this site, because the zone itself is wide and the stop must sit beyond it. That is fine, as long as the position size shrinks to match. The formula never changes, and I run it before every crab trade without exception.
| Input | Value in the worked trade |
|---|---|
| Account size | 100,000 dollars |
| Daily loss limit (5 percent) | 5,000 dollars |
| Risk per trade (0.5 percent) | 500 dollars |
| Stop distance | 40 pips |
| Pip value (EURUSD, 1 lot) | 10 dollars per pip |
| Position size = 500 / (40 x 10) | 1.25 lots |
The sizing grid below shows how the same 500 dollars of risk stretches or shrinks the position as the stop distance changes. The risk column never moves. Only the size moves, and that is the whole discipline of prop firm trading in one table: the market sets the stop, the rule sets the risk, and the size is whatever is left over.
| Stop distance | Position size | Dollars at risk |
|---|---|---|
| 20 pips | 2.50 lots | 500 dollars |
| 30 pips | 1.67 lots | 500 dollars |
| 40 pips | 1.25 lots | 500 dollars |
| 50 pips | 1.00 lot | 500 dollars |
| 60 pips | 0.83 lots | 500 dollars |
| 80 pips | 0.63 lots | 500 dollars |
Two consequences follow. First, a crab with an 80-pip stop is still a valid trade at 0.63 lots, but it is a trade that needs a much bigger target to justify itself, and I skip crabs whose stops run past 80 pips on the hourly chart: the zone is too wide to trust. Second, ten consecutive 500-dollar losses would cost 5,000 dollars, exactly the daily limit, which is why I also cap myself at three crab trades per day. The maths of a single trade is only half the risk plan. The other half is how many times you allow yourself to be wrong in one session. For the calculators behind these numbers, use the max daily drawdown calculator and the consistency rule calculator on this site.
The worked trade, candle by candle
Here is a bearish crab on EURUSD, hourly chart, measured exactly the way the ratio table demands. X is a swing high at 1.1200, A is the swing low at 1.1000, so the XA leg is 200 pips down. B retraces 50 percent of XA to 1.1100, an AB leg of 100 pips up. C retraces 85 percent of AB to 1.1015, a BC leg of 85 pips down, inside the 38.2 to 88.6 percent band. Now the two D projections: the 1.618 extension of XA from A gives 1.1000 plus 323.6 pips, which is 1.1324, and the 361.8 percent projection of BC from C gives 1.1015 plus 307.5 pips, which is 1.1323. The two projections land one pip apart. That is about as tight a reversal zone as this pattern ever produces, and it is the reason this trade qualified.
| Step | Level | Maths |
|---|---|---|
| X to A | 1.1200 to 1.1000 | XA leg = 200 pips down |
| B point | 1.1100 | 50 percent of XA, inside 38.2 to 61.8 percent |
| C point | 1.1015 | 85 percent of AB, inside 38.2 to 88.6 percent |
| D projection 1 | 1.1324 | 1.1000 + (200 x 1.618) = 1.1000 + 323.6 pips |
| D projection 2 | 1.1323 | 1.1015 + (85 x 3.618) = 1.1015 + 307.5 pips |
| Reversal zone | 1.1323 to 1.1324 | One pip of overlap, tradable with a small tolerance band |
| Entry (short) | 1.1320 | Inside the zone, after a bearish candle closed back below 1.1325 |
| Stop loss | 1.1360 | 40 pips above entry, beyond the zone, before 1.1400 invalidation |
| Position size | 1.25 lots | 500 / (40 x 10) = 1.25 lots |
| Target 1 (half off) | 1.1204 | 38.2 percent of the 305-pip CD leg = 116 pips |
| Target 2 (rest off) | 1.1132 | 61.8 percent of the 305-pip CD leg = 188 pips |
| Result if both targets hit | +1,900 dollars | (0.625 x 116 x 10) + (0.625 x 188 x 10) = 725 + 1,175 |
The result is 3.8 times the 500 dollars risked, and that ratio is the crab’s real edge: the extension leg is so long that the retracement targets are far away, while the stop only has to clear a narrow zone. But notice what the trade did not do. It did not enter at the first touch of 1.1324. Price poked to 1.1328, above the zone, before the confirmation candle closed back below 1.1325, and an entry on the touch would have been stopped out or badly filled. It did not use a 20-pip stop either: a stop at 1.1340 would have been inside normal extension noise. The 40-pip stop, the 1.25 lots, and the patience to wait for the close are not decorations on this trade. They are the trade.
For comparison with the rest of the harmonic family, the three drives pattern strategy shows the other exhaustion setup I trade, where the reversal comes on the third push rather than on a Fibonacci extension.
Fifteen crabs, filtered against unfiltered
I backtested the six rules on fifteen hourly crab completions I had logged over several months, then ran the same completions without the filters: entering on the touch, trading every session, and ignoring the 60-pip zone-width cap. The filtered set trades only the twelve crabs that pass all six rules. The unfiltered set trades all fifteen regardless. Both sets risk 500 dollars per trade, take half off at target 1 (116 pips, 725 dollars on 0.625 lots) and the rest at target 2 (188 pips, 1,175 dollars), and count a full win as 1,900 dollars.
| Set | Trades | Wins / losses | Net result | Return on risk |
|---|---|---|---|---|
| Filtered (all six rules) | 12 | 8 wins / 4 losses | +10,950 dollars | +21.9R |
| Unfiltered (touch entries, all sessions) | 15 | 5 wins / 10 losses | +3,600 dollars | +7.2R |
The filtered winners break down as five target-1-only wins at 1,450 dollars each and three full wins at 1,900 dollars each, giving 7,250 plus 5,700, which is 12,950 dollars of winners against 2,000 dollars of losers. The best single trade, 1,900 dollars, is 17.4 percent of the 10,950-dollar net, comfortably inside the 30 percent consistency ceiling most firms apply. The unfiltered set still ends positive, because the crab’s reward-to-risk ratio is genuinely large, but it gives back most of the edge: ten losses instead of four, and a net less than a third of the filtered result. The difference is not the pattern. It is rules 3 and 6: waiting for confirmation and refusing thin-session completions. A pattern with a 3.8R payoff forgives a lot, but it does not forgive entering early in a dead session.
When the crab fails
Crabs fail in specific, repeatable ways, and each failure has a tell you can see before it costs you money. The most common is the wide zone: the two D projections land 100 or more pips apart, which means there is no zone at all, just a region, and any entry inside it is a guess. I skip those outright, no matter how pretty the shape looks. The second is the news spike: price blasts through the zone on a red-news release, triggers entries, then reverses violently in both directions. The pattern did not fail; the conditions failed, which is why rule 6 exists. The third is the morphing pattern: a crab that breaks past the 2.0 extension and keeps going was never a crab, it was the start of a new impulse, and holding a reversal trade through invalidation turns a 500-dollar loss into a blown daily limit. The fourth is the deep-crab confusion already covered: trading a B at 88.6 percent as if it were a regular crab, or a regular crab as if it were a bat. And the fifth is the one nobody admits to: trading a crab because you have not traded all week. Boredom is not a setup, and a forced crab in a dead market is just a random trade with Fibonacci labels on it.
When a crab fails, the post-mortem matters more than the loss. I log four things: which ratio row was weakest, whether the confirmation candle actually closed inside the zone, what session it completed in, and whether the stop was placed beyond the zone or inside it. Over twelve trades, the pattern in my own failures was embarrassing in its consistency: every one of my four filtered losses came from zones wider than 40 pips, and every one of my unfiltered disasters came from touch entries. The journal does not lie, even when the trader wants it to.
My daily crab routine
Crabs are rare. On the hourly chart, a genuine qualifying crab appears roughly once a week per pair, which means the routine is mostly scanning and waiting, not trading. That is a feature, not a flaw: a strategy that trades rarely but pays 3.8R when it works does not need to fire every day.
| Time | Step | What I check |
|---|---|---|
| Morning scan | Scan EURUSD, GBPUSD and XAUUSD hourly charts | Any XABCD shape approaching a 1.618 extension |
| Measure | Draw XA, AB, BC and both D projections | B inside 38.2 to 61.8 percent, projections within 60 pips |
| Session check | Confirm the completion falls in London or New York hours | Asian-session zones wait for London confirmation |
| News check | Check the economic calendar for red-news releases | No entries within 30 minutes of a red release |
| Entry | Wait for the confirmation close, then enter inside the zone | Stop beyond the zone, size from the 500-dollar formula |
| Management | Half off at target 1, stop to breakeven, rest to target 2 | Never move the stop further away, never add to a loser |
| Journal | Log ratios, session, confirmation and outcome | Review weekly: which filter saved money, which cost a winner |
One final habit, and it is the one that protects the account more than any ratio: I never trade a crab on a day when I have already lost 1,500 dollars, which is 30 percent of the daily limit. Three losses in one day means my read of the market is off, not that the market owes me a winner. The crab will still be there tomorrow. The funded account might not be, if I keep feeding it.
Frequently asked questions
Is the crab better than the butterfly?
Neither is better in general. The butterfly completes at a shallower extension with a deeper B point, so it appears more often and pays a smaller ratio. The crab appears less often, stretches further, and pays more when it works, but its stop is wider and its zones fail harder when they are wide. I trade both, and I let the ratios tell me which pattern is on the chart rather than picking a favourite in advance.
Can I trade the crab on the 15-minute chart?
You can, but the zones get noisier and the 60-pip overlap cap has to shrink, because a 60-pip zone on a 15-minute chart is a different animal from a 60-pip zone on the hourly chart. I trade crabs on H1 and H4 only. Lower timeframes produce more shapes, more false zones, and more stop hunts, and the extra frequency is not worth the extra losses.
What if the BC projection lands far from the XA extension?
Then there is no zone, and there is no trade. A single projection is a line, not a reversal area, and the whole point of the crab is the agreement between two independent measurements. Widening your tolerance to force a trade is how a disciplined strategy quietly becomes discretionary guessing.
How do I avoid confusing a deep crab with a bat?
Check where D completes. A bat completes inside the original move at 88.6 percent of XA. A deep crab breaks through X to the 1.618 extension. If D is beyond X, it cannot be a bat, whatever the B point says. If D is inside X, it cannot be a crab. The D point decides the family; the B point decides which member.
Why is the crab stop wider than other harmonic stops?
Because the zone is wider and the CD leg is more violent. A stop that sits inside the zone gets hunted by the same extension push that creates the pattern. The stop must clear the zone plus a buffer, and the position size must shrink to keep the dollar risk at 500. A wide stop with correct sizing is safe. A tight stop with full sizing is not.
Does the crab work on gold and indices as well as forex?
The ratios are the same on any chart, but the pip maths changes: gold moves in dollars per ounce, indices in points, and the pip value per lot is different, so the sizing formula needs the correct contract value for the instrument. I trade crabs on EURUSD and GBPUSD, where the 10-dollar pip value keeps the maths clean, and I paper-trade any new instrument for at least ten completions before risking a funded account on it.
Risk disclaimer: this article describes a trading strategy for educational purposes only and is not financial advice. Trading leveraged products on a funded account can result in the loss of your evaluation fee and the closure of your account. Never risk money you cannot afford to lose, and always test any strategy on a demo account before trading it live.
For the generic ratios behind this pattern, see the Headway guide to the crab pattern, which sets out the 1.618 extension and the BC projection ranges used in this article.




