Disliked{quote} I was also thinking about Trailing Stops, but now I think it does NOT make sense for the way this EA is designed. This EA is designed to run and only close when it hits a SL or TP. In this way it can run for long stretches and overcome small bumps along the way until it reaches its TP. However, if you add a Trailing Stop -- with the intent to allow it to run even longer -- then I think you have to remove the TP setting. This is because the TP will be hit before you can take advantage of the Trailing Stop. I think @Pawel has created something...Ignored
My "Ultron" EA for GBPUSD H1 timeframe
- #188
- Edited 4:41pm Nov 9, 2018 4:31pm | Edited 4:41pm
- Joined Jul 2012 | Status: Trader | 1,880 Posts
- #190
- Edited 5:28pm Nov 9, 2018 5:13pm | Edited 5:28pm
- Joined Dec 2017 | Status: Trader | 4,378 Posts
I know you all understand, that when market is in the narrow range, we will loose money with this technique. Look at 2014. But that is a given when trading via trend technique. Chops always loose money. {image}
And here is 2013 January - 2015 December. 75% in 3 years, 10000usd with 0.5 lot size position.
- #193
- Nov 9, 2018 6:55pm Nov 9, 2018 6:55pm
- Joined Jul 2012 | Status: Trader | 1,880 Posts
Here are the settings...
Here are the settings... {image}
There is a reason why in my EA yours only choice is time zone and position size.
This is the only version I can take responsibilities for, you can talk about your modifications etc. but if you are using some variation not the one that I'm using then it make a sense to tell us that you are using some variation of this EA. I would like to put some rules into this thread. If you are making any changes, then when you share your results please inform that this is some variation of my EA. Since now every EA from me will have OP in its title. Which means it's from me - original poster. If you are doing some modifications, please remove...
Regards.
- #200
- Nov 10, 2018 4:09am Nov 10, 2018 4:09am
- Joined Jan 2015 | Status: Error 404: LuXing not found | 1,586 Posts
Test on 25yo period was positive too, but the slope was not so impressive
Using the very last version and a starting lot size of 0.5 / 10K capital, period January 2008 - December 2010, the result is roughly a 65% loss. Yes, I did quickly modify that last version using adjusted lot sizing to account for capital increase/decrease.
Here's the equity curve for that period.
Results for this period are not acceptable at all for any trader in his right mind.
From 2011 onwards, it performs quite nicely again.
So I was wondering, what is it that could possibly explain these huge differences in performance ?
Looking at your code, the first thing that struck me is that you use "static" distances between current price and MA (xx) as entry criterium. The market is not static at all though !
To illustrate this, see the ADR for GBPUSD (and ball park figure backtest performance in %) measured year over year.
Is it coincidence that during highly volatile years the EA underperforms ? I don't think so.
Now, based on yearly ADR of 2010 vs. 2016 we could of course argue that the average yearly range is not a plausible explanation for the underperformance.
Therefore, I pulled GBPUSD H1 OHLC history into excel and had a closer look and candle ranges.
When drilling down 2010 results and looking month per month at H1 candle sizes in pips, there is a common denominator.
Losing streaks coincide with many "outliers".
To illustrate this (albeit just a bit at random), see the difference below for a 10 period average candle size.
vs.
Not quite the same, right ?
I guess my point is that (IMHO) it could be a good idea to think about ways of flexible scaling of the min/max distance in pips.
Think average H1 candle size in pips measured over xxx periods, maybe throw in a standard deviation for good measure.
Just food for thought, that's all.
Cheers,
LuXing
Wisdom begins in wonder