Indicator library · Trend
Fibonacci Bollinger Bands
A Bollinger envelope with the multiplier taken from the Fibonacci ratios. The arithmetic is unchanged, the bands are narrower, and the mechanism that makes Fibonacci levels behave elsewhere is not present here at all.
What changes and what does not
The construction is the one set out on the Bollinger Bands page: a moving average of the close, the standard deviation of the same window, and bands at average ± k × standard deviation. The only difference is k, taken here as 1.618, and often 2.618 and 4.236 for further bands, instead of 2.
Everything else is identical, which is worth stating because the name implies a synthesis. The centre line is the same average. The width still breathes with recent volatility, expanding and contracting on exactly the same sessions. The shape of the envelope is unchanged; it has only been scaled.
The borrowed name
The Fibonacci calculator page states plainly why those ratios do anything at all: they are on every trading platform by default, computed identically from the same two visible points, so orders cluster near the resulting levels. The mechanism is coordination between participants, and it is the only mechanism on offer.
None of that transfers to a multiplier inside a volatility envelope. Nobody else can see the constant you chose; two analysts using 1.618 and 2 on the same instrument are not converging on a level, because the level depends on each one’s own average and deviation. Whatever the ratio contributes here, it is not the thing that makes Fibonacci levels behave on a chart.
That leaves a plain question with a plain answer: this is a Bollinger envelope drawn about 19 per cent narrower. If narrower bands suit how you read a chart, that is a legitimate preference, and it deserves to be described as one.
What Bollinger’s own readings survive
Usefully, the two readings that carry any evidence do not depend on the multiplier at all.
The squeeze, bandwidth reaching an unusually low level for that instrument, still identifies an unusually quiet period, because scaling every value by a constant leaves the percentiles unchanged. A squeeze at 1.618 deviations is the same session as a squeeze at 2.
The walk, price pressed against a band through a sustained advance, still says the same thing, and arrives more often with a narrower envelope. In both cases the reading that matters comes at the end of the walk, when a new push fails to reach the band.
That both survive unchanged is itself informative about how much the multiplier was ever doing.
How to test the claim yourself
The advantage of a hybrid whose difference is a single constant is that its claim is unusually easy to check. Plot both envelopes on the same instrument, then count: how many touches did each produce over a year, and what happened in the following five bars after each.
Two outcomes are possible and both are useful. If the outcomes after a 1.618 touch and a 2.0 touch are indistinguishable once you allow for the difference in frequency, the multiplier is doing what the arithmetic says it does, scaling the envelope, and nothing more. If they differ, the next question is whether the difference survives on a second instrument, because a constant that helps on one sample and not on another has found the sample rather than the market. Either way the test costs an afternoon and settles what the name cannot.
Where it misleads
| Situation | What goes wrong |
|---|---|
| The name read as a mechanism | Fibonacci levels behave through coordination between participants. A private multiplier produces none. |
| Touch read as a signal | It was not one at two deviations and is not one at 1.618, there are simply more touches. |
| Normal-distribution reasoning | Closing prices have fat tails and autocorrelation. No multiplier turns the envelope into a confidence interval. |
| Several bands plotted | Enough lines and price touches one. Hindsight then finds the level that "worked". |
| Multiplier unstated | Two charts labelled "Bollinger" with different constants are not comparable, and neither are levels quoted from them. |
| After a gap | Inherited from the original: the bands widen at once and narrow abruptly when the gap leaves the window. |
Why the page is worth having
Hybrids of this kind are common, and they are usually presented as combining the strengths of two methods. The test is always the same and it is easy to apply: identify the mechanism each parent relies on, then ask whether the hybrid preserves it.
Here one is preserved and one is not. The volatility envelope still measures dispersion, exactly as before. The Fibonacci ratios have been separated from the coordination that gives them their behaviour, and what remains is a number. Saying so is not a dismissal of the tool. It is what makes the tool usable for the thing it actually does.
Frequently asked questions
What are Fibonacci Bollinger bands?
An envelope built exactly like Bollinger Bands — a moving average with bands set a multiple of the standard deviation either side — with the multiplier taken from the Fibonacci ratios rather than set at two. Common versions plot several bands at 1.618, 2.618 and 4.236 standard deviations. The average and the deviation are unchanged; only the numbers multiplying the deviation are different.
Does using a Fibonacci ratio make the bands "Fibonacci"?
It makes the multiplier a Fibonacci number, and nothing more than that. Fibonacci retracements work (to the extent they do), because a great many participants draw the same levels from the same swing and orders cluster there. A multiplier inside a volatility envelope is not visible to anyone else and produces no coordination, so the mechanism that gives the ratios their behaviour elsewhere is entirely absent here.
What does changing the multiplier actually do?
It scales the envelope, and only that. Going from 2 to 1.618 makes the bands about 19 per cent narrower at every point; the shape, the timing of the expansions and contractions, and the position of the centre line are identical. Nothing about the market changes, only how often price touches the lines, which is a decision about how many touches you want to see.
Is a narrower band better?
It is more sensitive, which is not the same thing. More touches means more of them are ordinary movement; fewer touches means the ones that occur are more extreme by construction. The touch was never a signal at two standard deviations and it is not one at 1.618. What the multiplier changes is the frequency of an observation whose meaning has not improved.
Do the standard-deviation levels have statistical meaning?
Less than the language suggests, at any multiplier. The familiar "95 per cent of observations lie within two standard deviations" holds for a normal distribution, and closing prices are not normally distributed; they have fat tails and are autocorrelated. Both the conventional 2 and the Fibonacci 1.618 are conventions dressed in statistical clothing, and the honest way to read either is as a line drawn at a chosen distance.
Are multiple bands useful?
They are informative in the same way contour lines are: several envelopes at increasing multiples show how far price has moved relative to its own recent dispersion, without any of them being a level. The risk is the same as with pivot points and retracement grids, plot enough lines and price will touch several, and hindsight will find the one that worked. Two envelopes is a picture; six is a decoration.
What survives from Bollinger’s own guidance?
Everything that mattered, because none of it depended on the multiplier being two. The squeeze, bandwidth reaching an unusually low level for that instrument, is still the reading with a mechanism behind it, and a walk along the upper band is still evidence of strength rather than exhaustion. Those hold at any multiple, which is itself a hint about how much the multiple was ever doing.
Should I use it?
If narrower bands suit the way you read a chart, use them and say what the multiplier is. What is worth avoiding is the implied claim in the name: this is a Bollinger envelope with a different constant, not a synthesis of two methods. Calling it that keeps the reasoning clear and stops a coordination effect being attributed to a number nobody else can see.