# Proving that Fibonacci retracements have an edge...

#### Davidee

##### Member
I've met many advocates of using Fibonacci retracement levels in ones trading who claim that these levels 'work' but offer no actual evidence that they do.

Can anyone 'prove', or even offer some evidence that Fibonacci retracement levels actually work?

I've done a couple of little tests of Fibonacci retracement levels and it appears that (on large time frames at least) they are fairly significant and therefore could be said to 'work' -

Has anybody else done any tests that they'd like to share? I'm looking for actual evidence that these levels are statistically significant and that it therefore means something when the price reaches them. So, can anybody help 'prove' that Fibonacci retracements work?

I've met many advocates of using Fibonacci retracement levels in ones trading who claim that these levels 'work' but offer no actual evidence that they do.

Can anyone 'prove', or even offer some evidence that Fibonacci retracement levels actually work?

I've done a couple of little tests of Fibonacci retracement levels and it appears that (on large time frames at least) they are fairly significant and therefore could be said to 'work' -

Has anybody else done any tests that they'd like to share? I'm looking for actual evidence that these levels are statistically significant and that it therefore means something when the price reaches them. So, can anybody help 'prove' that Fibonacci retracements work?

like everything, they do not work just by themselves, they are a guide.
ANything that worked with any statistical significance would soon be traded away and stop working. If you want to define s trate based on fibs then fine but you'll need to look at other things as well like current trend, SR, etc etc

Good research so far but I'd add in a few things. The first thing you may want to look at is doing the same analysis with lines put in at 'random' percentages.

I am sure the results will be enlightening...

like everything, they do not work just by themselves, they are a guide.

Well, if they are indeed a useful guide, then the Fibbo levels must have some special significance that makes them different from other levels, right? Now here's the question, how do you actually know this? What research is there to show that this is actually the case? Any links?

Good research so far but I'd add in a few things. The first thing you may want to look at is doing the same analysis with lines put in at 'random' percentages.

I am sure the results will be enlightening...

That's a good idea actually, thanks And I'll left you know what the results are if/when I do it

Well, if they are indeed a useful guide, then the Fibbo levels must have some special significance that makes them different from other levels, right? Now here's the question, how do you actually know this? What research is there to show that this is actually the case? Any links?

The Fibonacci series is tightly related to the golden ratio. The golden ratio is deeply embedded in nature and I can find no reason to doubt it is a component of human psychology which is what the market is all about.

Well, if they are indeed a useful guide, then the Fibbo levels must have some special significance that makes them different from other levels, right? Now here's the question, how do you actually know this? What research is there to show that this is actually the case? Any links?

You enter a swing trade and see it go up 300 pips in your favour...then it starts to come back. At what point do you crap out and close your trade?

I haven't ever seen convincing evidence that they work. They seem very arbitrary to me in where you draw them. So IF they work, then I would check ways of drawing fibs that are not arbitrary. I don't buy this golden ratio nonsense. The ratio isn't in nature as often as people claim. That is a myth if you look into it.

The only one that I would think MIGHT work is the 50% retracement, for various reasons, but that isn't really Fibonacci.

Then again, if enough people believe in it, it could become self-fulfilling. But that requires enough people to all see it, which is why I suggest a non-arbitrary way of drawing the fibs.

'Proving that Fibonacci retracements have an edge'
or any method or technique for that matter is 'is it profitable ?'
that's the Only criteria required

'Proving that Fibonacci retracements have an edge'
or any method or technique for that matter is 'is it profitable ?'
that's the Only criteria required

So it's best to throw money at something rather than doing a bit of research first ?

The Fibonacci series is tightly related to the golden ratio. The golden ratio is deeply embedded in nature and I can find no reason to doubt it is a component of human psychology which is what the market is all about.

The connection between psychology and Fibonacci can be summed up simply:

If you look hard enough for something, you'll probably see it.

A retracement level represents a change in direction between 2 prior swing points. So - you first mark out a high & low. Then you draw in a number of lines between those points. Then you allow a couple of percentage points either side of your lines and suddenly you see something of significance.

Now - instead of using fib numbers - just try lines at 15%,30%,45%,60%,75% and 90%. You will see just as much significance in those numbers as you do with fibs.

You could write a program that looked for significant retracement levels with no bias whatsoever. Would anyone bet money that this program would come out with the fibs as the most significant ?

Our brains are made to make short-cuts in order for our survival. For example "Tiger, run". It is these shortcuts that give us tendencies to see things that or not there.

So it's best to throw money at something rather than doing a bit of research first ?

In important step in qualifying any strategy is to use it with small money. It's not the first step, but it is a critical step.

The connection between psychology and Fibonacci can be summed up simply:

If you look hard enough for something, you'll probably see it.

There is ample evidence that many do not agree with this. Including a current client of mine that manages significant investment portfolios.

I've always wondered why use Fibonacci? Why not just simple quartiles or something? Personally, knowing my own weaknesses in trading, I can see it's use in helping to provide some structure or consistency to trade planning but as a tool to measure price movement not convinced it's anything but witchcraft.

I've met many advocates of using Fibonacci retracement levels in ones trading who claim that these levels 'work' but offer no actual evidence that they do.

Can anyone 'prove', or even offer some evidence that Fibonacci retracement levels actually work?

I've done a couple of little tests of Fibonacci retracement levels and it appears that (on large time frames at least) they are fairly significant and therefore could be said to 'work' -
Has anybody else done any tests that they'd like to share? I'm looking for actual evidence that these levels are statistically significant and that it therefore means something when the price reaches them. So, can anybody help 'prove' that Fibonacci retracements work?

Fibonacci retracement and extensions are widely used in all kinds of markets...forex being an obvious example...Just this morning in cable the 1hr downtrend off 6181-Friday's Hi that had developed found demand at the 61.8% of the recent 5950-6181 swing up, the recovery from a 2nd test of this 6039 low found supply at the current pullback Hi at the 50% of said 6181-6039 move. Trouble with fibs is that they can be subjective which is why they are potentially really more significant on clear 1hr + swings. Because of the potential subjectivity it is always preferable to look for a confluence of potential support/resistance factors and a price action confirmation at a potential swing point...these may include trend lines and previous price pivots (ie near-term obvious fractal hi/lo's that show where imbalances of supply over demand / demand over supply actually existed.)

G/L

The Fibonacci series is tightly related to the golden ratio. The golden ratio is deeply embedded in nature and I can find no reason to doubt it is a component of human psychology which is what the market is all about.

You can see no reason to doubt it's a component; but do you know it. What research has actually been done to prove it.

I haven't ever seen convincing evidence that they work. They seem very arbitrary to me in where you draw them. So IF they work, then I would check ways of drawing fibs that are not arbitrary. I don't buy this golden ratio nonsense. The ratio isn't in nature as often as people claim. That is a myth if you look into it.

The only one that I would think MIGHT work is the 50% retracement, for various reasons, but that isn't really Fibonacci.

Then again, if enough people believe in it, it could become self-fulfilling. But that requires enough people to all see it, which is why I suggest a non-arbitrary way of drawing the fibs.

When I did the experiments I linked to I didn't expect to find any significance in the levels, but nevertheless the results of the experiment were the results. I have to follow the results of the experiments, not my feelings, but I'm still not convinced so I want more evidence

The only one that I would think MIGHT work is the 50% retracement, for various reasons, but that isn't really Fibonacci.

Some argue that 50% isn't a fib ( as you pointed out ), I'm not into Fibonacci so I have no opinion on the matter.

Having said that your comment about the 50% level caught my eye.

Can I ask you to expand a little bit on why you believe the 50% retracement level MIGHT work?

dd

Last edited:
IMO, this is a lazy investors method of selecting a level and blaming Fibs when it does not work.

BTW, When you find the level do you buy or sell?

Poor Fibonachi. What has he done to deserve this?

BTW, When you find the level do you buy or sell?

depends on which way the price trending when you identify your pivots for drawing fib dunnit

Replies
12
Views
4K
Replies
10
Views
13K
Replies
1
Views
7K
Replies
2
Views
6K
Replies
12
Views
3K