LazyBear's WaveTrend port has been praised for highlighting trend reversals with precision and punctuality (minimal lag). But strong "3rd Wave" trends can "embed" or saturate any oscillator flashing several premature crosses while stuck overbought/oversold. This happens when the trend stretches over a longer timescale than the oscillator's averaging window or filter time constant. Our solution: simultaneously monitor many oscillator timescales. Watch for fresh crossovers in "dominant" timescales alternating most smoothly between the overbought (red shade) and oversold (green shade) range.
Fractal Resonance Component facilitates simultaneous viewing of eight timescales that are power of 2 multiples of the chart timescale. Each timescale shows lead line, lag line, lead-lag difference, and crossover marks. Add 4 to 8 copies to your chart for a good multi-fractal read. Format * the "Timescale Multiplier" attribute of each row to be twice that of the row above for a sequence like 1, 2, 4, 8, 16, 32, 64, 128...
Fractal Resonance Component shifts its timescales along with your choice of main chart timescale:
Daily chart: 1 day through 128 day (~4 month) oscillators.
Crossovers in different oscillator ranges tend to have different meanings:
Minor (< 75%) crossovers: small green/red dot [**] usually noise
Overbought/Sold crossovers (shaded 75 to 100%): black outlined dot (o) [**] reliable reversal indicators (when they appear alone)
Extreme Overbought (> 100%) crossovers: black outlined plus (+). [**] Can be a major reversal in fast markets, but usually portend the end of Elliot 3rd waves with just a small corrective (4th wave) retrace before the larger impulsive (5-wave) sequence resumes in original direction. [**] The final 5th-wave terminus should appear later as a lone non-extreme (black outlined circle) crossover on a slower timescale coincident with weaker (non-extreme) dot crosses on this timescale.
Careful examination of historical charts leads to many useful observations such as:
Dominant crossovers punctuating true reversals are usually in the green/red shaded ranges with black outlined dots (o) rather than minor or Extreme (+) ranges.
Due to market's fractal nature, two well-separated timescales like 1 minute and 1 hour can show dominant crosses simultaneously in opposite directions, e.g. the 1 minute showing a very short term high and the 1 hour a medium term low nearby.
Staying Nimble Watch out for embedding on your supposedly dominant timescale -- a second cross while stuck in the overbought/oversold region suggests a stronger, longer trend than expected. Drop your eyes to a slower timescale below for the real dominant whose crossover will validate main trend reversal.
Embedding can often be predicted even at the first cross mark by checking whether the green lead line of the next slower timescale (one row below) has already hit the Overbought or especially the Extreme Overbought range but isn't close to rolling over. Fractal Resonance Bar (to be published) uses this principle to mark embedded timescales with white stripes, warning of a powerful trend wave on longer timescales you shouldn't fight until the white stripes subside.
Overnight gaps surge all timescales in ways that obscure the dominant timescale, so for shorter than daily charts, these methods work best on Futures contracts that only suffer weekend gaps.
EXCELLENT JOB I PERSONALLY THINK YOU ARE NOT ONLY SUPER INTELLIGENT BUT YOU HAVE AMAZING PROGRAMMING SKILLS 2 PEOPLE HAVE NOT UNDERSTOOD THE VALUE AND IMPORTANCE OF THIS AMAZING INDICATOR ..IF YOU JUST CHANGE FEW SETTINGS THIS IS ONE OF THE BEST INDICATORS I HAVE EVER SEEN THANKS A TON MAY TO ALWAYS BE BLESSED AND PROSPER REGARDS FROM INDIA
pythagoras
⋅
@Bornleader, I'm curious what settings you wanted to change. Have to be careful changing settings just to 'overfit' the historical data you're looking at. Most of the settings I've exposed just affect the speed of the oscillators which will necessarily make them fit patterns with particular speeds a bit better or worse.
BarsStallone
⋅
I think another way to solve the issue with the indicator may be to do it the same way as a rolling linear regression which also has a tendency to have oscillation issues. I think this could be solved by implementing a Kalman Filter.
pythagoras
⋅
@justin, Do you mean the saturation issue? Checkout the new Fractal Composite Ribbon. By compositing several timescales it is less susceptible to saturation on a single timescale, yet it can still occur eventually. That's an inherent tradeoff in using oscillators that have a bounded range for you to trade within to represent price that has potentially unbounded range.
See the Fractal Resonance Composite page for explanation of math behind these oscillators and why we consider them superior to many of the standard indicator implementations: