I do not personally use MACD to signal a trend. I was only responding to a question about how to use it in conjunction with %R. However, MACD is no worse than any other moving average to define a trend. I prefer price action and trend line breaks to define trends. There are a lot of different ways to use MACD besides signaling a trend. That is only the most basic use. I like it principally for divergence signals on a lower time frame in the direction of the trend on a higher time frame. I am curious what method of trading you find to be effective in chop? I am also very interested in the use of volatility in trading and would be interested in how you use it.
I'll answer the chop question for now.... Firstly I'm talking about stocks only, not indices-Forex etc, as I only trade stocks. Chop can happen at any time, ie unpredictable and most often I've found it will happen when the market as a whole takes a breather. As trading is a probabilities game you need to look at this issue in terms of probabilities I believe. What I do is scan for stocks which historically trend well. This can be done a number of ways, eg, how often do your buy and sell signals indicate on each particular stock. ( you need a program which to do explorations on - like Amibroker) Another method is look for stocks which have historically low bar volatility as well as run up and down steeply in terms of % gains and falls. I've found it's best to use weekly bars, not daily for this type of searching. Yet another method is to use RSI and measure ONLY the average RSI over a longish lookback period. eg If the average RSI over a lookback of 12 months is 60, then you know this has been a trending stock. Now just because its been a trending stock in the past doesn't guarantee it will in the future. So what do you do now? Well by using indicators a little differently in approach you can come up with some novel ideas. Another approach is when the market as a whole is downtrending or chopping, what are some particular stocks doing? When you've found the best performers, then look at the last annual report. Look at who the directors are. If they have experience in their field of endeavour, keep an eye on the stock. If the directors for example of a Resource company are predominantly Lawyers, accountants, politicians, bankers, throw that stock into the junk directory and don't go near it again. As I said, it's a probabilities game, so screen out the junk stocks first, then trade what's left over. Quality companies are what will give best returns, junk companies can also trend well but what happens often is it will often plunge suddenly. That's OK so long as you are very fast and continually sharp with the trigger finger.
It's currently being used by the guy in first place ( or third, depending on what day it is) in an automated trading contest in currencies: http://championship.mql5.com/2011/en/news/93 Larry recommended using it with some kind of moving average. Like if your %R is set to 10, your moving average would be 50 or so. Just a way to buy the dips and sell the rallies. The above trader has given away some code to this effect. In the contest he is experimenting with money management.
As I said 18 months ago... Today's chart, 5 mins, giant sell signal at 10:30. Exit signal is just before 11 am...
That is a failed signal, obviously.... It happens... Otherwise this would be the Holy grail. But good catch...
Thought i would at least report back with the final results of this robot after three months. He started out great but it was all downhill from there. Not sure what happened, i'll have to look closer at the code. http://championship.mql5.com/2011/en/users/gery18
Ive never been a big fan of indicators for a couple obvious reasons , been a price action trader for many years now but have just recently thought of building a mechanical auto system to reflect my price action methods . Due to knowing what i actually want to reflect structurely on a chart ive begun to find indicators are the easiest way to reflect some key levels of interest to me . WilliamsR is a great thing for some of these purposes . on the right period the -50 level will reflect 50% retracements , you can also use it to define when swing lows and highs are challenged or db or dt . I use new X period lows and highs for defining trends As long as you know what the indicator is reflecting it can be a huge advantage trying to mathematize these key areas . i personally think williamsR will be a huge part of a bot to trade price action , you just have to think outside the box