Broker/Data: Mirus Futrues, Zen-Fire, IQFeed, Kinetick
Favorite Futures: 6E, CL, GC
Posts: 132 since Apr 2010
Thanks: 22 given,
Great to see some of you guys making headway with some divergence indicators. I think mystiq is correct about some indicators producing better quality signals but I think it's even more important to consider what parameters you are applying to those indicators. Any oscillator, for instance, can be made to show divergence with price but, given different input values, they can be made to indicate the complete opposite.
It's almost impossible to guess at the optimum input values. What would have worked great today usually fizzles quickly when forward tested. That's why any robust divergence indicator must be coded into a backtestable strategy and prove to work sufficiently well in practically any market or timeframe without tweaking (curve fitting). This is a long and tedious process that should ultimately produce many divergence signals with good profit potential. But you still won't be able to trade it profitably because, along with the great signals, you'll get a lot of "trash trades" and spotting the difference is often not easy.
Two big problems with trading divergence are the inevitable trash trades and, by their nature, you miss a lot of great trade opportunities to trade with the trend instead of always against it. I strongly recommend traders include some "with trend" entries in their arsenal of trading weapons. But how do we eliminate those pesky divergence signals that aren't going to make us a dime? The answer is, "We can't eliminate them". But we sure can thin them out a lot...certainly enough to make trading a profitable enterprize for the disciplined trader.
You learn a lot when you study something for 8 hours a day every trading day for over a decade. I'm still discovering new properties of divergence. Fascinating stuff. Why some divergence signals work and others don't is usually pretty obvious even when the signal forms were nearly identical. But the "evidence" is usually scattered over several charts in different timeframes...even different types of charts for that matter. For instance, are we trying to go long in the ES on a 1-minute chart when when all of your indicators on a 3 period Renko shows the ES is ready to tank? Your 1-min chart signal might work, but it could never be considered a high potential trade.
So, if we can gauge the success potential of a well formed divergence signal simply by looking around before we pull the trigger, how can we not be successful on most trades? Because most traders are capable of looking at a few things...a couple of indicators on a couple of charts...but most are not capable of processing all of the necessary data in time to take the trade when it should be taken. Only computers can process boatloads of information fast enough to give you the answers you need in plenty of time. This is why I firmly believe that any good robust trading system should be coded. If it can't be coded then it's obviously loaded with subjectivity and of doubtful value. What info is processed and how it's processed is vitally important. The GI GO rule always prevails.
The attached chart shows most of my signal types...Basic Divergence (BD) and Extended Divergence (ED). Those are my counter-trend signals. The Hidden Divergence (HD) and the Lump Trade (LP) are my with-trend signals. You'll notice a multi colored stochastic for finding divergences and a multi colored trend line. The bottom panel is what I call my Momentum Change indicator. The colored horizontal lines are Murrey and work great for support and resistance levels...and a good number of other uses...but I am not a proponent of Murrey Math as a trading system.
The striped bar indicator in Panel 2 is very useful. It let's me look at what's going on with any number of indicators in an unlimited number of timeframes and chart types from Renko, Line Break, Tick, Range, etc. without having to take my eyes off of my primary trading chart which, in this example, is a 5 Better Renko (available in Big Mike's Elite section) of the CL (oil). I like trading on a Renko chart because I always know what the closing price of my entry bar will be and can easily enter with limit orders and trade with virtually no slippage. The strongest entry indication is when the red & white "candy cane" is present when a short signal fires or blue and white present when a long signal triggers.
These signal entries are highly filtered and were all computer generated. The code is able to look at each indicator separately and it knows exactly what it needs to see for an ok for a signal to appear . I can eliminate all of the indicators that you see here from view and just look at Price and the signal entries that are produced in real time with plenty of warning. Makes the complex very simple but it's taken over 30,000 lines of code to get it to this point.
Anyone wanting to learn more of how I trade using this system can simply contact me to schedule a session 1-on-1. Note that I am a vendor (educator), but understand that I will neither ask nor accept payment in any form for anything I share with my fellow futures.io (formerly BMT) friends. Just glad to help if I can. I'll probably post some of my indicators here on the forum after the holidays. I have more that were not discussed in this post.
Happy New Years!
Last edited by rogerf; December 26th, 2010 at 04:22 PM.
The following 15 users say Thank You to rogerf for this post:
A great post, but curious on what HD (Hidden) or ED (Extended) is? I am guessing Extended divergence is just divergence that continues? Hidden I have no idea...
Your post was very clear and precise. Good work.
Oh, one more comment. ForexFactory and MT4 is lightyears ahead of what guys have on the other platforms it seems to me at least from what I have coded or played with. It is good to see some of your indicators be on par with MT4.