I loved this thread. What you described is your experience at something that I have also experienced, but from a different view, as we all interpret differently. But, the same thing.
The time you spend in trying to share is immense and you are very good at breaking down your explanation.
I skipped a few pages and wound up at this post. You put so much into that indicator, then stopped using it. I'm just curous if that is also some parallel experience; Simpler is better?
I used to pull displaced-alternates (100%) and DALTXs, and retracements, and extensions, and exhaustion fibs, until I had lines everywhere. And where I found confluence, I felt stronger about. But really, without a turn, and without volume, it meant nothing most of the time. I've had greater success not looking at anything less than 30 minutes, and limiting my fibs to simple and basic.
My view today is that there are no harmonic laws or golden ratios that can be found in financial data. The only reason that Fibonacci retracements and expansion work is self-fulfilling prophecy. It is a fad used by traders, let us call them meeting points, which are drawn somewhere on the chart. If many traders use the same meeting points, they start to work.
Today I am looking for confluence zones, which are made up from
- prior swing highs and lows
- Fibonacci retracements across different timeframes
- Fibonacci expansions across different timeframes
- Fibonacc projections
- measured moves (also called alternates)
and I do not draw the individual lines on my chart (too many), but just the areas, where different concepts of support and resistance reinforce.
The first chart shows the individual lines detected by the FibonacciCluster indicator. These lines do not include projections and measured moves. The second chart shows a selection of different Fibonacci lines. The third chart shows the confluence zones, where the Fibonacci lines add up. Which indicator can be used best for your trading?
The following 5 users say Thank You to Fat Tails for this post:
We look at the same things, and I agree with your take on why they work. While there may not be any statistical advantage of one over another, collectively when all traders are watching the same areas, and when you have confluence on multiple timeframes, those areas have a possibly higher probability of having something happen, particularly when something does happen...
I pull my internals, externals and APPs manually on a daily and a 30 or 60 or 90 minute. Id be curious to see how your indicator with the least number of lines (image on the far right) works. Is it available for download?
I agree, the least number of lines make sense. I am very interested in how do you evaluate the significance of each pivot or zone. I would be very pleased having a look into this analyzer as shown on the third picture.
I am looking on the statistical success of reversal lines to qualify the trading levels to trade off. i.e. a price must have shown a certain number of time in the past monthes that it has impact on price action to qualify as a level.
Your approach would give me a good starting point to combine my rating algo with your statistical approach.
The statistical approach, which I use is semi-scientific only. The indicator detects 54 support and resistance lines in 10 different time frames, which is a total of 540 lines. Each timeframe is represented by a ZigZag, which runs over the daily lookback and the chart period. The 10 zigzags vary by their minimum deviation, which is calculated from the average true range over the entire chart period.
Starting from the swing highs and lows, the indicator calculates from each of the 10 zigzags
- prior swing highs and lows (2 levels up and 2 levels down)
- prior micro swing highs and lows (4 levels up and 4 levels down)
- Fibonacci retracements (5 levels up and 5 levels down)
- Fibonacci expansions (6 levels up and 6 levels down)
- Fibonacci projections (6 levels up and 6 levels down)
- Measured moves (4 levels up and 4 levels down)
which makes up for a total of 54 S/R lines per zigzag. For each type of line I have run distributions over various instruments to make an estimate of the conditional probabilities that price reverses when hitting a zone with a width of x percent around that line. Based on those findings, I have attributed a statistical weight to each of the 54 line types. The weights are then modulated by a timeframe factor. Larger timeframe lines are considered being more important than smaller timeframe lines. The modulation factor is based on coherence of results, that is empirical testing of the outcome.
I am not willing to share all the statistics, as I want to keep a few of my secrets.
The following 6 users say Thank You to Fat Tails for this post: