When not to trade?

Hi,

I'm in the process of putting a few strategies together in a single EA and have got stuck on one important issue. My system is quite vulnerable to tight trading ranges, that is it produces quite a few false signals in such conditions. Do not confuse this with a ranging market, it may range as long as it ranges in a slightly more interesting pattern than up, down, up, down, up, down...
Instead of tweaking the logic of the system which is already quite complicated, I thought I'd come up with an external guard that would monitor the price behaviour and prevent the system from trading when conditions are unfavourable. I have tried two approaches so far, none of which seems to please me. First off, I tried a simple threshold - if the highest high minus the lowest low over the last N periods falls below it - no trading. This obviously doesn't work if the price goes up or down slowly but continuously - you may miss a decent trend.
The I moved on to trying - as I called it - a relative trading range. This is the ratio of the trading ranges over the last N periods and M periods where M > N.
There's no trading when this ratio goes below - say - 0.2 (20%). This method has one drawback - it's got a poor performance when a relatively tight range gets even tighter. In this case you may have a range of 10 pips which gives you 35% relatively.
As I'm getting a feeling that I'm trying to reinvent the wheel, can you share your thoughts on this please? Note - I'm not interested in any ATR-like indis based on moving averages. I need something that quickly reacts to sudden events like breaking out of a range (just like the approaches I've just described). It doesn't have to work in general. I'm targeting GBPUSD on M5 so I'm thinking about exploiting some of its characteristics. Any ideas/suggestions more than welcome.
Cheers

Efficiency ratio. Look into "adaptive moving averages."

Thanks! I'm already using AMAs for other things in the system but didn't think they would be useful here.
Cheers.

What is a "false" signal? Are you trying not to ever have a losing trade?

It's a signal that triggers a loser.

Yep, and I'm quite close to it. Seriously, the vast majority of all my losers comes from these tight ranges. When I trade the system manually I know when to refrain from taking a position. Now I'm looking for a way to quantify it.

Giving a Winning Signal and Not Giving a False Signal are basically the same.
ie:
TRUE = NOT FALSE

Good that someone focused on "NOT LOSING" instead of just winning.

Cheers! and good luck with your EA

You could have it switch to a breakout mode after the first false signal. So once it has a losing trade, the code then sets up a box based on the last X number of candles and then just waits for a breakout of that box. That would keep the EA out of trades until there is a significant move.

I think this is a brilliant idea. I've just given it a quick visual inspection and it looks like many false signals are filtered out especially those I was concerned about. It's a shame I've never taken breakouts seriously

Price is always breaking out of some level.

What is your risk per trade? What is your TP and SL? What is your Risk/Reward?

Risk - anything between 0.5 and 1.5 % depending on market behaviour
TP - 3 levels depending on market behaviour, they may change during a trade
SL - none, I use my own exit signals
R/R - don't know and don't care as I find it useless

What do these have to do with my orignal question?

Good point.

________________________________
REAL ea:http://realea.awardspace.com
Automation is key

Try avoiding large bars,this is a sign of volatility.

________________________________
REAL ea:http://realea.awardspace.com
Automation is key

In order to recommend changes, the more one knows about the system, the better the recommendations.

Price volatility as such is there all the time and the higher it is the better it actually is.
Would you care to elaborate what "avoiding bars" and what a "large bar" is? This is a programming forum.
Cheers

So what do you recommend then?
Cheers

Dynamic Price Channel. Price is always breaking out of a channel. The channel is 2 dimensional - price and time. Identify what channel price is/is not breaking out of and you may solve your dilemma.