My EA, put to the test

Greetings andersenws, stopped by to show support for a fellow EAfxer. Always enjoy hearing of the hopes and visions of those attempting to put some heart into the machine. Best wishes on your trading journey.

: ) Thom

grin: PS...and if you happen ever want some objective-non-biased testing of your EA, just PM me. Happy to assist. Mums the word.

Congratulations on starting a journal my friend, and best of luck to you! I'm going to follow with interest.

Thanks for the support guys.

Itme, the Risk:Reward is variable, but only on the reward side. From what I learned in the "Best Way to Choose Exits" thread, I decided to implement an exit strategy that will take profit at 10 (most of the time), or if the system warrants, it will let profits run. Just looking at the first 5 trades of my backtest, 3 were for +10 pips, 1 was for +21 pips, and the big one was for +66 pips. However, the stoploss is a constant 50. The report with the 4% risk you were looking at has an average loss of -50,107.68 with a maximum loss of -54,049.05. The average win was 17,490.42, but the maximum was 196,020.00! I think this shows how the profits are allowed to run but the losses are kept to a minimum, which should be the goal of any trader or system.

David, I am jealous that you have over 75% wins in real time! I hope I can live up to that. I, too, wonder about the effect of news on the backtest. Maybe it could have pumped up my % in some way? I don't know. We will see in the coming months. Thank you for your support.

I will be extremely happy if this EA can hold +85% win rate in real time over the coming months. I do not expect +95%, even though that is what the backtest says. I'll take 85% though.

BTW, the EA took its first trade today, profit for +11 pips.

Trade
http://img508.imageshack.us/img508/6709/tradezd0.th.png

I also have another trade open, BUY 0.20 lots at 1.2529. It is currently barely in the money, but I am confident it will reach the 10 pip goal.

Thanks guys,
andersenws

AndersenWS, If I understand your performance report correctly, then the optimum risk factor should be about 25%.

By increasing your risk factor from 4% to 25%, you should avoid the problems of an excessively high risk factor - namely insufficient capital in the account to cover the 50 pips stop, and excessive drawdown caused by sequential or clustered losses, adequate to margin out the account.

By increasing the risk factor eightfold, the weekly rate of return on capital will increase from 8% to about 100%. The result will be a return on capital in 6 months of about 16 million times the initial capital. Is this correct?

My second trade just closed, so I thought I'd post that. The swap ate a little of the profits, but it was another winning trade.

Trade
http://img218.imageshack.us/img218/7696/tradezp7.th.png

So far, so good.
2/2 = 100%
21 pips
0.8% profit

I do not believe there will be any more trades tonight, but i will keep you guys posted.

Thanks,
andersenws

P.S. itme, I read your post and I am tinkering with higher risk factors. I will post a reply later tonight. Thanks

Ok, I have ran some more tests with varying risk levels, and it appears that itme is correct (which I suspected all along). A higher risk level would suit this system better, for higher profits over the long run. However, I tried to make this experiment as realistic as possible, so that was the reason for my initial 2% (which I now think is extremely small). I think I could comfortably implement a 7-10% risk, which would result in larger profits, but keep drawdowns to a level that my emotions could handle. I just finished a test over all of 2006 thus far, which I will post tomorrow, and it turned 10K into 177K. Again, this is from 1/1/2006 to 10/15/2006. The drawdown was larger (3200), but the win% (88%) was pretty close to the overall average.

I will post the report and graph tomorrow.

Thanks to all,
andersenws

Just thought I'd keep you guys updated. There were no trades taken yet today, although I am thinking there will be one this evening or during the night. I will post the 10% risk backtest later tonight.

Thanks,
andersenws

Balance is still $5040.27

Here is the 10% risk backtest from 1/1/2006 to 10/15/2006

Graph
http://img475.imageshack.us/img475/8...0610cj6.th.gif

Report
http://img384.imageshack.us/img384/615/10ki0.th.png

Thanks,
andersenws

Just so you guys know, I have been having some serious problems with MT4 this week. Not being able to connect, not being able to backtest...etc. So I am trying to figure that stuff out. As soon as I do, I will post any trades if they were taken.

Thanks,
andersenws

Ok I got a new account, reinstalled MT4, and re-imported my Alpari Data. I took 3 more trades since my MT4 account has been down. I have decided to just post the week's results with all the trades of this week as well as some analysis.

See you tomorrow,
andersenws

I have been studying the various backtests I'v erun today and discovered that I need to put a filter of some kind on these trades. Most of the losing trades are longer than 12 or even 24 hours. The winning trades on the other hand, are almost never longer than 15 or 30 minutes. So I am thinking of putting a time filter on the trades. In addition, I am thinking of modifying the trade at the open of each new 4H candle, moving the stoploss 10 pips closer.

For example,
I buy at 1.2587 SL at 1.2537
4 hours go by and I have still not closed the trade.
Modify Buy 1.2587 SL at 1.2547
4 Hours go by and I have still not closed the trade.
Modify 1.2587 Sl at 1.2557
Price hits SL, but I only lose 30 pips instead of 50.

Just from visually backtesting this strategy, it seems to cut losses without affecting the winning trades too much.

I'm working on posting the results from this week right now. They will be up later.

Thanks,
andersenws

Sorry I haven't kept this up guys, I've been busy this last week. Here are the results from last week. My demo was down for the middle of the week, so that is why the statement reflects that. I've also attached a backtest of the forward test from last week (sounds confusing, but think about it). Last week went 5/5 for 52 pips. Awesome!

Here is the statement:
http://img60.imageshack.us/img60/448...mentjo8.th.png

And the backtest with the middle of the week's trades that aren't reflected in the statement:
http://img230.imageshack.us/img230/7...testzd5.th.png

My upcoming goals for this EA are:
1. Develop some sort of trailing stop (tighten stop after certain time period)
2. Study bad trades and why they failed (may post some later this week)
3. Research optimal risk levels for the EA

So far,
5/5
52 pips
2.1% gain in a week

Thanks,
andersenws

Looks like a pretty reliable system, cant wait to see more results

Thanks pkeigs.

I was wondering if anyone knows if you can actually trade lot sizes like 1.04 or 1.52 or something like that. I know that tenths of lots can be traded, but I didn't know if hundredths could be traded or not.

Thanks,
andersenws

Yes interbank FX will allow mirco lots. Just e-mail them requesting micro lots in your account.

Thanks David.

I was about to code my stop modifier (don't really know what to call it) and I didn't know when to place it. My idea was to tighten the stop by 10 pips at the beginning of each 4H candle after the trade was taken. Would it be better to tighten the stop after 4 hours elapsed instead? Or do you think it would make that much of a difference?

Also, I have an open BUY trade right now, +2 or +3 pips. We'll see how that goes.

Thanks,
andersenws

goodluck man, keep us posted :-)

Thanks pooshkin. I'm going to do some homework but I thought I'd keep posted on my BUY position. It was at 1.2565, and my quote now is 1.2573, so hopefully this trade will close successfully in the next 20 minutes or so.

Thanks,
andersenws

The results are quite impressive. There is something that I don not quite understand.

On the first report you mentioned that you risked 1% and it reads MaximumRisk=0.1. On the second report that you risked 4% it reads MaximumRisk=0.04. Could you explain that?

You've also changed the LotSize from 2 to 1.2. Could you explain what does that mean?

Thanks and good luck,
Al

Interesting results. Seems like you've got a nicely developing system. Wish you the best when you go live. I agree with a previous poster, you should have rushed to get this into the competition.