PS: here is a sine wave indicator that is made exactly as the original (found the tradestation code in the "Rocket science for traders" book) Now, you will find out that what Ehlers tells about since wave indicator and when you tray to apply it to forex time series are two different things. Also there are some versions that are using Cycle period to adjust the calculation of sine wave, but then all you get is nicer looking and smoother values that are making bigger mistakes
Attaching both the metatrader version as well as the tradestation version
In my opinion sine wave indicator should be disregarded. Hilbert homodyne discriminator on the other hand as a source for phase accumulation and then using that (with a couple of changes and deviations from Ehlers way of calculating it and using it) has proven to be very useful tool in adaptive indicators and we have been using it in quite a few indicators. So, even then it is not used directly, but indirectly as a modifier for some other indicators calculations. I do not see such a possibility for a sine wave indicator (the cycle period part can be used as a modifier but there are much better way to sue something similar to that than to try to use smoothing to make cycles more usable)
To cut it short (after being long )
Try this out, but I think that you will see how wrong it can be. I saw one version from igorad using cycle period but that version is making those even bigger trend estimation errors and that was the reason why I never found interesting the Ehlers sine wave indicator for usage in trading. Anyway, try it out. Who knows. Maybe I am missing something
PS: this version work in new metatrader 4 and does not need any external indicator to work
here is this version too
Adding these two simply as an experiment : one is Ehlers sine wave of rsi and the other is Ehlers sine wave of stochastic. Some toys to see if sine wave can be used on some other "price" sources than price itself and what would be the result of the original John Ehlers' code in such cases (as a kind of estimation of sine wave indicator usefulness)
Interesting...when compare it with SSA...
It is the 48 bar period limit that is going to kill you. If you look at the cycle extraction thread you will find tools to help you. The strongest cycles are a lot longer than 48 bars. Ehlers has this tendency to view longer period lengths as trends. He did the same thing with his Corona charts and everyone wonders why they don't work very well.
I think you are on the right track but this might not be the place to look.
Many thanks for your helpful advice.
This what he is calling 'roofing filter' is just band pass filter which passes the cycles in length
48-10 for example, this idea he started in 90ties to use it for trading, now it's just a new name.
Anyway I evaluated roofing filter and super smoother quite deeply. I have a AI system with 170
inputs so I tried 1st to smooth input series before preprocessing by SS than by roofing filter.
In 1st case profit factor was the same like without smoothing, in 2nd case PF
was ALWAYS LOWER than for raw data.
Than I had a deeper look at behaviour of roofing filter itself and discovered that it has very
strong tendency to overshoot, just plot it together with e.g. RSI and you will see what I mean.
This is very unwanted behavior which introduces extra noise and whispaw trades.
So most likely this and cutting off longer cycles was causing profit factor to fall. This system makes several thousands trades so results are significiant
Many thanks for those findings. Can I ask if you have any preferred Ehler's indicators after that ? Also it is of course noting that an indicator, or indicators, alone do not a trading system make.
Sorry for late answer, just noticed it recently. At the moment no Ehler indis on my preference list
and believe me I studied his work quite deeply.
Almost all Ehler work assumes existing of cycles on high TF, Im trying to build strategies on 1min chart and can not find his filters working on 1min FOREX data but maybe on another data
with cycles it works.
Rocket Science for Traders
For everybody who likes ebooks, here is a download link for Rocket Science for Traders of J.Ehlers .
Damiani RSTD_EnhancedSignal_to_Noise indi
Hi there, here is Mladen's fixed version of the Damiani RSTD_EnhancedSignal_to_Noise indi which is using the Hilbert Transformation ideas including the Homodyne_Discriminator described in Rocket Science for Traders.
Please enable the necessary setting in your browser, otherwise you will not be able to log in.