Community technique; refined by numerous practitioners
Fibonacci retracements: a structure for pullback entries, and an honest look at the evidence
No technique in this library is more widely used or more thinly supported. The claim that markets respect ratios derived from the Fibonacci sequence does not survive careful testing, and saying otherwise would be repeating folklore. What does survive is more modest and more useful: a pullback trade needs a short list of candidate entry prices and a level beyond which the idea is dead, and this method supplies both — which is most of why it works at all.
- Style
- Mean reversion
- Approach
- Discretionary
- Difficulty
- Intermediate
- Horizon
- Swing (days to weeks)
- Holding period
- Days to weeks
- Time needed
- 20-30 minutes a day
- Markets
- FX · Single stocks · Commodities
- Source
- Standard technical analysis practice; a rigorous treatment appears in Constance Brown's Technical Analysis for the Trading Professional — Community technique; refined by numerous practitioners
The rule set
- Identify a clean impulse leg and mark its start and end — this anchoring decision determines every level that follows
- Draw the 38.2%, 50% and 61.8% retracement levels between those two points
- Wait for price to reach the zone and then confirm a turn — a reversal candle, a volume shift, or a shorter-timeframe structure break
- Place the stop beyond the 78.6% retracement, where the premise of the setup no longer holds
- Target the prior swing extreme, or an equal-measure extension of the impulse leg
What makes it distinctive
- It reduces 'where do I get in on a pullback' from an unbounded question to three or four candidate prices
- The invalidation point is naturally clear — beyond the 78.6% retracement the setup is dead, with no interpretation required
- It demands reversal confirmation, because a level on its own is not a signal and the method is explicit about that
When it works
Trending markets with orderly pullbacks, and best where a retracement level coincides with something else — a prior high, a moving average, a round number.
When it fails
Choosing the impulse leg is subjective, and different anchors produce completely different levels. Strong trends retrace only shallowly before continuing, and weak ones slice through every level without pausing.
How a decision moves through it
Input
Swing-timeframe price data
Daily or 4-hour bars for most applications. The method needs identifiable swings, which very short timeframes do not reliably produce.
Measure
The impulse leg, anchored by hand
The single most consequential step and the only genuinely subjective one. Every level downstream is a fraction of this choice.
Measure
Retracement levels at 38.2%, 50% and 61.8%
Note that 50% is not a Fibonacci ratio at all — it is included because half a move is a natural reference, which is a hint about what these levels actually are.
Decide
Price at a level, with confluence and a confirmed turn
Three requirements. The level alone is the weakest of the three and the only one most people use.
Act
Enter on confirmation, stop beyond 78.6%
The stop level is the method's most defensible component: a retracement of more than about three quarters means the move being retraced was probably not an impulse.
What the evidence actually supports
The usual justification runs through the golden ratio: 0.618 and 0.382 appear in the Fibonacci sequence, the sequence appears in nature, therefore markets respect these proportions. Each step of that argument is weaker than the last, and the conclusion does not follow from any of them.
Studies looking for a special tendency of price to reverse at Fibonacci ratios have generally found no effect distinguishable from what you would expect at any arbitrary set of levels covering the same range.
It is worth noticing that 50% — the most commonly watched level of the three — is not a Fibonacci ratio at all. It is included because half a move is an obvious reference point, and its presence rather gives away what the whole set is doing: providing a small number of sensible places to look inside a pullback.
What the method actually provides
Strip away the ratios and something useful remains. 'Buy the pullback' is an instruction with no entry price and no invalidation point. This method supplies both, and that is a real contribution regardless of where the numbers came from.
Buy the dip, unstructured
- No specific price to act at
- No level that says the idea is dead
- Entry decided by feel, in the moment
- Position size has nothing to anchor to
With retracement levels
- Three candidate prices, decided in advance
- 78.6% is a clear invalidation
- Orders can be placed before the emotion arrives
- Stop distance is known, so size is computable
The right-hand column is worth having even if the ratios are meaningless. Deciding your entry and your exit before price gets there is most of what separates a plan from a reaction.
Confluence is where any real edge lives
The version of this method with the best claim to working is not 'price reaches 61.8%, buy'. It is 'price reaches 61.8%, which also happens to be where a previous high sits and where the 50-day average is, and it turns there'.
In that description the Fibonacci level is doing very little work. The prior high is a real level with a real mechanism — traders who bought there, traders waiting to buy a retest. The moving average is watched by many participants. The retracement level's contribution is mostly to have narrowed your attention to that price in advance.
Five ways into this system
- Anchoring the leg, drawing the levels, and requiring a turn before actingFive steps, of which the first is subjective and determines everything else, and the fourth is the one most people skip.7 min read
- The level you enter at decides your stop distance, and therefore your sizeThe three entry levels are not interchangeable: each has a different distance to the same stop, so each implies a different position size.6 min read
- Where clean swings exist, and why FX traders use this more than anyoneThe method needs identifiable swings. Markets and timeframes that do not produce them make the anchoring step arbitrary, and everything downstream inherits that.5 min read
- The anchor you chose, the level that always fits, and the trend that never pulls backThe failures here are mostly epistemic rather than mechanical. The method's greatest risk is that it is very hard to be wrong with it.7 min read
- Fibonacci retracements for beginners: useful lines, dubious theoryHow to draw them, what they are for, and why you should be sceptical of every explanation of why they work — while still finding them useful.6 min read
The ideas behind it
This system assumes you already know these. Each one is explained from scratch in Investing 101.
Compare with
- Holy Grail Setup (Raschke)When ADX confirms a strong trend, buy the first pullback to the 20-period moving average and target the prior swing high.
- Wyckoff Accumulation/DistributionDecide whether a trading range is large buyers accumulating or distributing, then join the move once the structure confirms it rather than guessing the low.
These are documented methods described for study. Nothing here is investment advice, a recommendation, or a claim about future returns — every system on this page has losing periods, and the pages say where.
Reading about a system is not having one.
Plutux is where you write your own rules down, test them against real data, and keep the record your memory would otherwise rewrite. Join the waitlist for early access.