zentauro67 » Mon Jan 13, 2014 8:35 am wrote:Yes, it seems a good strategy only for ranging periods. It seems (if I´m not wrong) that the EA uses the slope of the band as a non trade filter (strong trend=no trade). We´ll see if it will work. In manual trading I´ve always had problems with the slope, as it repaints and it is always too late when it appears. My personal trick with TMA bands to know if price is trending or ranging is this: I add an Ichimoku (default settings) to the chart. If the cloud (kumo) is inside the band price is ranging, it the band leaves the kumo price is trending. Kumo is by far more reliable and easier to spot than the slope.
I have not found too much information about TMA+CG. It seems CG refers to Center of Gravity (another big repainter) and could be a mix.
Just to clarify:
This is a good strategy for when price is moving in waves ( i.e, not moving straight up or down ). This more often than not happens during ranging periods, but also applies to trends where price is undulating in one direction.
It's also a good thing (for this EA) that markets range more than they trend, so hopefully we can make something out of that.
The slope that I am using is the LibCSS slope, not the slope of the actual TMA+CG bands. The TMA+CG is purely for defining the outer boundaries (places where we can buy and sell). I'm aware of the repainting; I'm happy to say it doesn't pose a significant threat. ( That is actually partly why a <buffer no. of pips> is subtracted from TP, and why I'm using Ask price to detect upper and Bid price to detect lower ==> stacking odds in my favour ).
As with all EAs, the entry condition (bands) + slope filter are not foolproof, but they don't need to be  I know some may be thinking this is too simple, needs a bit more <extra>, but often-times, simpler is better.
I just need it to be right slightly more times than its wrong, and to make more (on average) than it loses per sample size of trades. Then the rest will be taken care of by quantity of trades + time.
Cheers |
