It's like many other indicators: nice to look at after the facts; impossible to wring money out of it by applying kiddy rules.
If a trader is going to follow the trend, it makes sense to first find the trend. The eff ratio works a heckuva lot better than arbitrary MA crossovers.
I know it's better than nothing, but it's technically inefficient. You'll always be late with that kind of indicator...
Isn't that the point of trend trading? Late in late out, catch the big part of the move? So getting in late should never worry a trend follower in my opinion.
You're full of dreams my dear ! This theory would be great if the market was 'trending' most of the time, but it actually doesn't ! For one big 'trend' you'll catch, you will get burned 10 times and loose in the average. Well tried.
Well that's a problem with trend following techniques, and not my idea. It is however very possible to trade trends provided the market you trade trends well. I think this is the problem with most trend followers, they don't pick their markets well. Anyhow all markets are trending on some timeframe. Take a look at the S&P weekly chart, can you see any trends?!
Identifying a naturally 'trendy' market and identifying a 'trending now' market is a different thing. The first is easier to do : computing autocorrelation and persitence is very easy. However, these autocorrelations/persitences are not large enough to counter slippage, spread and transaction costs. But this is of course something nobody selling TA service will tell you.