How to choose an EA?

One the main problems I face as a newbie is deciding which EA(s) to use.

I guess there's a few parameters we can look for :

1. Risk/Reward (partly determined by the recommeded default TP/SL settings for each EAs) but also the %WIN/PROFIT for week/month etc.

2. EA type (Trend-based, volitility-based etc.)

3. Average Pips/week.

It would be great to have a feature on this website where a table lists all the critical indicators of each EA to allow us to compare/contrast each.

The problem is often with hundreds of pages to read, and limited knowledge, however enthusastic we might be to learn there is considerable inertia given the range of choices and the amount to digest.

Can anyone recommend which EAs tend towards giving a pretty constant return with modest risk? I guess this is all relative. There's just so many to choose from.

Where do we start?

How about none of them? It is impossible for an EA to work effectively in all market conditions. Impossible! There is no EA money making machine. An EA is a good way to knock out your balance. But I guess no matter how many times you tell a newbie, they still won't listen.

The ones that do appear to work, usually have too much activity, that some brokers wouldn't allow it.

FWIW, my EA works well in trading markets. For much of the last 12 months, it would have made a killing in backtest on EURUSD ($1000->$5000 or $6000). Except for a period in March/April and part of August. As for all of 2005.... it would have been knocked out. Ditto for 2004.

The problem is that an EA that works well in a trending market gets whipsawed. An EA that does well in a trading market gets squeezed. Bye bye equity!

I wouldn't run an EA against a live account unless it met all the following conditions:
(1) Works only on H4 or D1.
(2) Works well on all major pairs. If you think that's possible, look at the charts... all the pairs have their own personality and I don't think you can write an EA that can adapt to each of them.
(3) No more than 5% relative drawdown. I don't care about any of the other statistics but it should be able to make 20-30% a year return.
(4) Can adapt well to changes to trend or trading.
(5) All the MT4 bugs are worked out. I still don't have faith in MT4 yet.

R299, there is no HG EA out there, but if one knows each of the EAs personality's then there is hope. Forward testing several of the more popular EAs here at FF and over at fxtsd have shown me that there is hope in this area of melding man and machine. There are now MQ4 programers that are developing managing EAs that can, with relative accuracy (after all an EA simply looks at the same lagging indicators we look at!) if market mood is range bound our in a trend, and user will have the ability to tell this managing EA to switch between a portfolio of appropriate EAs.

Here is what i have found in my short few months of forward testing some of the popular EAs...a consistent 10% to 15% positive growth in the accounts during the months i've been demoing them. This may not be much to some, but if managed properly a proven portfolio of EAs can be a good tool in one's trading methodology.

: ) Thom

If you're still interested in an EA, I'd be more than happy to sell you mine The results below are a test from October 19, 2005 to October 18, 2006, using H4 on EURUSD, 90% modelling quality. Would you pay $400 US for this EA?

Actually, in all seriousness, I wouldn't sell an EA because I wouldn't give someone false hope. I don't think it would work long term. You see that 10% relative drawdown? Too high! Far too high.

Please don't ask me to run it on other pairs either In other words, that just goes to show you that bottom line: an EA programmer can make their EA show whatever they want given the right data. My EA didn't do too well in the last month (-$6000), and it doesn't do well on other pairs. But does that show in the below stats?

Does a money making EA exist? Probably, but you have to ask yourself, will the EA ever fail due to changing market conditions? Think about it. There are situations in my EA that I know it won't adapt to.

Sorry to sound harsh and all, but I think EAs should be used as tools, not as automated machines.


Bars in test 9278
Ticks modelled 1486320
Modelling quality 90.00%
Initial deposit 100000.00
Total net profit 90546.64
Gross profit 201813.88
Gross loss -111267.24
Profit factor 1.81
Expected payoff 576.73
Absolute drawdown 246.00
Maximal drawdown 20703.28 (11.10%)
Relative drawdown 11.10% (20703.28)
Total trades 157
Short positions (won %) 142 (59.15%)
Long positions (won %) 15 (66.67%)
Profit trades (% of total) 94 (59.87%)
Loss trades (% of total) 63 (40.13%)
Largest
profit trade 5092.50
loss trade -3194.10
Average
profit trade 2146.96
loss trade -1766.15
Maximum
consecutive wins (profit in money) 13 (18713.70)
consecutive losses (loss in money) 11 (-12552.30)
Maximal
consecutive profit (count of wins) 31798.00 (9)
consecutive loss (count of losses) -13502.02 (7)
Average
consecutive wins 5
consecutive losses 4