How to define a good EA?

Hi all,

I am in the process of creating an EA. I would like to know from the EA developer's and most importantly users, what is a good EA? what is a good return from an EA? How much draw down is acceptable? Like if you start with 10k, how much your EA would make with with max draw down? Does the EA has to produce somewhat similar returns on all pairs? Does your EA work the same on all TF? Which EA is better, with fewer trades/year (with high pips target) or the one with more (and lower pips target)?

As always, thanks for your time!

Well, the most ideal EA:

 

1. 100% win rate.

If the EA wins all the time, there's no need to worry about drawdown, profit factor, or ROI

Well, I was talking about EA's designed by lesser mortals like me . Here is my first EA tested with a SL of 100 pips and exits at 75 EMA with a constant lot size of 2.

Small Mistake: The lot size was 10 for 50k. For 10k, I use 2 lots.

Well, again it depends on what sort of characteristics you want the EA to have.

There are a series of trade-offs you need to take into consideration:

 

1. Accuracy vs Drawdown
2. Straight vs stepped equity curve
3. Risk vs. Reward

General rule of thumb:

Trending EAs:

Shoots for the astronomical gains. However, there are usually many false entries for those gains, leading to extended periods of losses.

Scalping EAs:

Aims for small consistent gains. Most of these scalping systems have stoplosses larger than their takeprofits which means it takes more wins to offset one loss.

Breakout EAs

Look for that burst of movement outside of a predefined range of some sort. This can sometimes lead to several false breakouts if not done properly.

It largely depends on the investment characteristics you want from that single EA.

My most recent independent EA developments have been trying to determine the difference between the four stages of a market cycle.

 

1. UpTrend
2. Peak
3. DownTrend
4. Trough

Once I can get an EA to determine where it is in a market cycle, I can then program the EA to decide which system it uses: (Trending, Ranging, Breakout, Scalping, Counter Ranging, News). I suggest that's where you direct your research.

My EA is based on the current trend in the pair. Going long or short in the direction of the trend. As you said, my EA has more losing trades (be it small ones comparatively) then the winnings, I am already filtering out quite few trades (it open 29 trades on 4H frame for a period of more then two years). I am looking at ways if i can further filter them out. I am testing it with all the majors, but EURUSD seems to be the most trendiest.

Filtering out trades is certainly one option. How do you plan to profit from that gap where you filtered out the trades?

Well, even if my EA opens less trades, and also if it can have a better success rate (my current SL is 100 pips, tested it with around 80 pips SL, that seems to be OK with almost the same results), I'll then be trading more lots per order (not sure if that was what you asked )