this is my first post, I am not a coder, but I love this forum and the ideas that are circulated in here.....
Hopefully you can make some sense out of my idea......
I have been reading most of the threads and one that caught my eye was "Joe"..... and even though the "Joe" thread seems to have trickled down to a crawl, the concept intrigued me.....so I started looking at the charts from a different point of view than I usually do, and came up with this idea.... it's incredibly simple, and it has most likely been done before, maybe even by Steve....but I haven't found it...
Now I want to start by saying that I have not actually traded this, so if you don't want to look into coding this I totally understand......
This was done on H4 so that is what I'm gonna explain, I am curious what it will do on different TF's too though...
I'm gonna use buy for the example, sell will be the same, but opposite....
Anyway here it is:
1. We start the week flat.
2. We do nothing on the first candle. (This part could maybe be user defined, how long before we start looking.)
3. At the open of the second candle we look back, if the close of the first candle was higher than the open of the first candle, we buy at market with a Candle Trailing Stop two candles back (which I am actually not sure how that will work, with the second candle back being from previous week, like if there is a gap, that would not work)
.4. At the open of the third candle we look back, if the close of the second candle was higher than the open of the second candle, we buy at market, if the close was lower than the open we do nothing but wait to either get stopped out if the trade goes far enough against us, or see what next candle will do..
5. We use a candle trailing stop two candles back until we get stopped out, and then the count starts over at the open of next candle.
I have manually back tested this for just 4 months starting January 2011 on the EURUSD H4 with a win ratio of roughly 35%......am I getting you excited yet? (It's low, but that is roughly the win percentage the actual trader "Joe" had trading the Flying Buddha's)
I know 4 months is less than nothing as far as an adequate sample size, but I think it might have some legs.....
There are extended runs of losses, so some intestinal fortitude might be needed, with a few really big short winning runs.....in 4 months the backtest racked up roughly 3000 pips.( I know it's backtesting and manual on top of that, so bias, mistakes reading the numbers, and mental mistakes likely makes the numbers different.) But it got me excited enough to throw it out here for scrutiny...
Like I said previously, I'm not a coder, but it seems to me, one of you coding geniuses here should not have too much trouble to make this EA, if my explanation was good enough... If nothing else, maybe it can be a starting point, where ideas from the group might make it better.........
Thanks for letting me be part of this great forum.
xfr8dog