Help Test my new EA, FLYER! | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

My trade explorer Flyer > 1000% return for May 3 replies

Flyer! EA - New Thread - Part 2 147 replies

Flyer! EA - New Thread! 456 replies

Fozzy, Pip Boxer, Flyer 4 replies

Flyer! Version 2.00 is released. 0 replies

Attachments: Help Test my new EA, FLYER!
Exit Attachments

Help Test my new EA, FLYER!

  • #1
  • First Post: Edited 9:31pm Sep 30, 2006 8:34pm | Edited 9:31pm
Hello Everyone:

<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o></o>

I would like to introduce an EA I have developed called “Flyer!”. It is a moving average EA based upon the principles used in the “Sidus” method listed here in FF. Thanks goes to Sidus for the inspiration to write this. However, it does have some differences. I have finished backtesting Flyer! and am now ready for the forward testing. I would appreciate any help from anybody for assistance with the forward testing and further back testing on other currency pairs besides the ones I have done.

<o></o>

So why should you test this one? Well, for one thing the back testing results are excellent. 80-85% win ratios with maximum relative draw downs between 10-20% depending on the currency pair. The EA also adjusts lot sizes as the account grows. A 1 year back test on the GBP/USD resulted in $10000 turned into over $7 million. This high rate of return seems too good to be true, but other pairs also resulted in returns over $1 million. Note, in ranging markets, the EA will not gain a lot in value and may draw down some too, but once it starts trending, look out… the pips will come!! So, now with this kind of preliminary rate of return, I encourage everyone to try it out and see how it works and offer improvements where available. See the html files in the zip for the reports on each currency pair tested.

<o></o>

After you have tried it out, please post your results here on FF
<o></o>

Many thanks!

<o></o>

GirlFlyer
Attached File(s)
File Type: zip FLYER!.zip   153 KB | 2,278 downloads
  • #2
  • Sep 30, 2006 9:29pm Sep 30, 2006 9:29pm
Thanks, GirlFlyer, I'll start testing next week.

Does your preliminary work suggest what account size is needed to trade this on the optimal pairs you indicate in the EA's text? I see you've set the default to 8%. I'm just wondering, for example, with a $10k account, how many pairs you would trade. I'd be curious to see what type of drawdown it experiences.
Quoting GirlFlyer
Hello Everyone:


<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o></o>

I would like to introduce an EA I have developed called “Flyer!”. It is a moving average EA based upon the principles used in the “Sidus” method listed here in FF. Thanks goes to Sidus for the inspiration to write this. However, it does have some differences. I have finished backtesting Flyer! and am now ready for the forward testing. I would appreciate any help from anybody for assistance with the forward testing and further back testing on other currency pairs besides the ones I have done.

<o></o>

So why should you test this one? Well, for one thing the back testing results are excellent. 80-85% win ratios with maximum relative draw downs between 10-20% depending on the currency pair. The EA also adjusts lot sizes as the account grows. A 1 year back test on the GBP/USD resulted in $10000 turned into over $7 million. This high rate of return seems too good to be true, but other pairs also resulted in returns over $1 million. Note, in ranging markets, the EA will not gain a lot in value and may draw down some too, but once it starts trending, look out… the pips will come!! So, now with this kind of preliminary rate of return, I encourage everyone to try it out and see how it works and offer improvements where available. See the html files in the zip for the reports on each currency pair tested.

<o></o>

After you have tried it out, please post your results here on FF
<o></o>

Many thanks!

<o></o>

GirlFlyer
  • #3
  • Sep 30, 2006 9:49pm Sep 30, 2006 9:49pm
Thanks, GirlFlyer, I use the Sidus method and will try this out and get back with the results.............
  • #4
  • Oct 1, 2006 5:31am Oct 1, 2006 5:31am
Quoting BenQ
Thanks, GirlFlyer, I'll start testing next week.

Does your preliminary work suggest what account size is needed to trade this on the optimal pairs you indicate in the EA's text? I see you've set the default to 8%. I'm just wondering, for example, with a $10k account, how many pairs you would trade. I'd be curious to see what type of drawdown it experiences.
BenQ,

In the tests, an 8% usage of equity for 1 pair on a $10,000 account worked fine. Maximum relative drawdown averaged about 15%. None of my tests brought the account balance below $8000. If you change the percent used to 3-4%, you should be able to handle 2-3 pairs per $10000 in your account and grow rapidly. Of course, you can have larger or smaller accounts. If you want to run many pairs in the same account, set the equity to 1 or 2% or even lower. Half percents would be fine, too. Remember, the program will adjust your lot size according to the equity in your account and the percentage you choose. It is not a fixed lot size EA. If the drawdown on the account lowers the equity, the next trade will have a smaller lot size. If a profit is taken, the next trade will have a larger lot size, etc.. I used a relatively small equity size because money management tells us not to over leverage. Find an usage of equity per lot that suits your taste.

GirlFlyer
  • #5
  • Oct 1, 2006 9:58am Oct 1, 2006 9:58am
What's the suggested PointAdjust parameter setting for gbpusd? It's not listed in your suggested optimal settings.
  • #6
  • Oct 1, 2006 10:19am Oct 1, 2006 10:19am
Hello Girl Flyer or BenQ
Could you please let me know how to attach the EA to
the MT4 platform?
I also trade Sidus method on my Demo
Senbanda
  • #7
  • Oct 1, 2006 1:45pm Oct 1, 2006 1:45pm
you might want to import some historical data from

http://www.alpari-idc.com/en/dc/databank.php

the modeling quality is only 25%, you should be able to get it up to 90%

with out proper data mt4 "fakes" what the interday data would be by splitting the daily bar into 12 pieces, so interday systems always looks like the holy grail.

the first ea i made netted 65billion in 2 years. after i got the modeling quality to 90% it was only $1
  • #8
  • Oct 1, 2006 5:29pm Oct 1, 2006 5:29pm
Quoting BenQ
What's the suggested PointAdjust parameter setting for gbpusd? It's not listed in your suggested optimal settings.
Ben, GBP/USD uses .0001
  • #9
  • Oct 1, 2006 5:35pm Oct 1, 2006 5:35pm
Quoting witchazel
you might want to import some historical data from

http://www.alpari-idc.com/en/dc/databank.php

the modeling quality is only 25%, you should be able to get it up to 90%

with out proper data mt4 "fakes" what the interday data would be by splitting the daily bar into 12 pieces, so interday systems always looks like the holy grail.

the first ea i made netted 65billion in 2 years. after i got the modeling quality to 90% it was only $1
Witchhazel...

I've already had the Alpari Data. The EA uses the M1 dataset from Alpari along with the M30 data loaded from the server. I've seen the modelling quality issue and am working on that problem. The only data that I have don't have in my system is the M5 and the M15 so far. But working on that, too. Once I have them, I will see if it changes the modelling quality.
  • #10
  • Oct 1, 2006 6:26pm Oct 1, 2006 6:26pm
Quoting witchazel
you might want to import some historical data from

http://www.alpari-idc.com/en/dc/databank.php

the modeling quality is only 25%, you should be able to get it up to 90%

with out proper data mt4 "fakes" what the interday data would be by splitting the daily bar into 12 pieces, so interday systems always looks like the holy grail.

the first ea i made netted 65billion in 2 years. after i got the modeling quality to 90% it was only $1
Witchhazel,

I've did another backtest using a timeframe in which I have all the historical data in my system on every timeframe. The modeling quality was still only 25%. But still had great success rate. Any further suggestions?

Girlflyer
  • #11
  • Oct 2, 2006 2:20am Oct 2, 2006 2:20am
Hi Everyone,

I'm attaching an updated version of my Flyer!. This version will now allow you to switch chart timeframes while the EA is running without resetting the EA and causing multiple entries.

It also allows for recovery in case of power failure and most importantly, during the weekend, you can shut down the program and then come back to it later.

GirlFlyer
Attached File(s)
File Type: mq4 Flyer!.mq4   16 KB | 1,367 downloads
  • #12
  • Oct 2, 2006 4:34am Oct 2, 2006 4:34am
Hi Girlflyer

Just started testing your EA - have you got a pointadjust setting for Cable? Couldn't find it in the preamble!

Andy
Andy
  • #13
  • Oct 2, 2006 4:39am Oct 2, 2006 4:39am
Quoting andydoc
Hi Girlflyer

Just started testing your EA - have you got a pointadjust setting for Cable? Couldn't find it in the preamble!

Andy
Sorry, its missing the point adjust for GBP/USD. Use .0001

Thanks for testing.

GirlFlyer
  • #14
  • Oct 2, 2006 8:28am Oct 2, 2006 8:28am
Girl Flyer,

EA looks like it has great potential! I have been testing it since market opened last evening. Thus far it has closed five trades as of 9:30 EST. Four of the five have been with GBP/USD, one with USD/CHF. All of the cable pairs have been for losses. The Swiss pair was only for $2.90. The USD/GBP has had the point adjust set at .0001 from the beginning of trading. Most of the entries/exits have not been at the EMA crossovers and have missed the trend. In most instances, when a pair should be exiting and the trend reverses, it should also buy and/or sell with the reversal. In other pairs (majors) there have been quite a few opportunities to enter and exit but the EA has failed to do so. I hope this helps in tweaking the entry/exit. You may also want to allow a TP with TS switch for the EA for those who want to ensure that a certain amount of PIPS is obtained rather than allowing the second crossover for the exit. Just a thought. I'm looking forward to working with you on this. Alas, I wish I was a programmer (I'm studying) but at this point can only offer my observations. I hope this helps.

Mike Johnson
1Trader2
  • #15
  • Oct 2, 2006 8:35am Oct 2, 2006 8:35am
Ok, I am getting quite a few responses about the 25% modelling quality of my tests which make people doubt the veracity of the EA which I can understand. However, further research shows that on the M1 chart which I am running the strategy tester it is impossible to get better than 25% according to the method strategy tester uses to calculate that percentage. Quoted from the Metatrader MQ4 website (see link at: http://articles.mql4.com/93 )

Quoted from above link:

"Reasoning from the same formula, the one-minute data modelled cannot be of quality exceeding 25 per cent since data of a smaller period are not used for modelling of these data. For client terminal, there are simply no data of a smaller period, but there are tick data. However, tick data are not stored in the client terminal by standard means.

In themselves, one-minute data are the most detailed data about the price movement. And, using them, one can model the price movement inside of bars of larger periods. At that, discrepancy between the real price movement and the ticks modelled will not be large due to presence of one-minute reference points: Open, High, Low, Close. These reference points form the first approximation of the price movement function (the first-order approximation). The higher is the value of the period modelled, the more there are reference points and the more exact is the modelling. And vice-versa, the lower the value of the modelled period is, the lower is the modelling quality. For example, up to 240 one-minute reference points can be used to model a one-hour bar. When modelling a 5-minute bar, the amount of reference points drops to 20.

As we mentioned before, there are just 4 reference points for a one-minute bar to be formed! This is why the rating of modelling quality for one-minute bars does not exceed 25 per cent. However, if one uses signals received from data of a larger timeframe when testing on one-minute timeframe, the data modelled should, of course, be rated much higher because the price movement within the bar of timeframe larger than the on-minute one is considered."

End of Quote

So, if you read this and understand it, it is now shown that better than 25% on the M1 is impossible.

Thus, the EA test results are probably about as accurate as they can get.

I hope this helps for people to understand that my results are not bogus.

GirlFlyer
<!-- / message -->
  • #16
  • Oct 2, 2006 8:39am Oct 2, 2006 8:39am
Quoting 1trader2
Girl Flyer,

EA looks like it has great potential! I have been testing it since market opened last evening. Thus far it has closed five trades as of 9:30 EST. Four of the five have been with GBP/USD, one with USD/CHF. All of the cable pairs have been for losses. The Swiss pair was only for $2.90. The USD/GBP has had the point adjust set at .0001 from the beginning of trading. Most of the entries/exits have not been at the EMA crossovers and have missed the trend. In most instances, when a pair should be exiting and the trend reverses, it should also buy and/or sell with the reversal. In other pairs (majors) there have been quite a few opportunities to enter and exit but the EA has failed to do so. I hope this helps in tweaking the entry/exit. You may also want to allow a TP with TS switch for the EA for those who want to ensure that a certain amount of PIPS is obtained rather than allowing the second crossover for the exit. Just a thought. I'm looking forward to working with you on this. Alas, I wish I was a programmer (I'm studying) but at this point can only offer my observations. I hope this helps.

Mike Johnson
1Trader2
1Trader2,

Hi, thanks for the testing. I'm running it as well. Currently at 9:40est I'm in 5 trades, USD/CHF,GBP/USD,EUR/USD,USD/JPY, and AUD/USD.

All 5 positions are positive in value.

Was stopped out on to USD/GBP positions earlier today when it bounced up and down twice for +50 pips which exceeded the stop loss. That can be expected to happen occasionally.

As test results come in further, I will continue to post updates and tweaks as needed.

GirlFlyer
  • #17
  • Oct 2, 2006 8:42am Oct 2, 2006 8:42am
Girl Flyer,

After having been involved in testing Firebird extensively as well as a couple of others, it has been stated many times by members of that thread, that backtesting especially with Strategy Tester is unreliable. It is better to do forward testing and use back testing with the best modeling quality you can get to get a sense of what the EA is capable of. I wouldn't fret over those who are requesting a better modeling performance. Lets go forward...see my previous post.

1Trader2
  • #18
  • Oct 2, 2006 10:35am Oct 2, 2006 10:35am
Hi GirlFlyer,
just to let you know I'm testing this EA as well.

So far only the one trade opened on GBP/USD (I'm running a bunch of different pairs on demo account), its currently up after this mornings run on cable and the SL has been adjusted automatically by the EA.

It will be interesting to see how much of that profit I'll retain by the time the current trend reverses a bit and the cross occurs.

Nice EA so far.. Please keep tweaking it..
  • #19
  • Oct 2, 2006 11:12am Oct 2, 2006 11:12am
Quoting JokerZ
Hi GirlFlyer,
just to let you know I'm testing this EA as well.

So far only the one trade opened on GBP/USD (I'm running a bunch of different pairs on demo account), its currently up after this mornings run on cable and the SL has been adjusted automatically by the EA.

It will be interesting to see how much of that profit I'll retain by the time the current trend reverses a bit and the cross occurs.

Nice EA so far.. Please keep tweaking it..
Thanks Joker for the support.

I agree with you on the profit taking. I'm up significantly on cable, too along with eur/usd and usd/chf. Looking forwards to seeing what the exit will be.

GirlFlyer
  • #20
  • Oct 2, 2006 2:49pm Oct 2, 2006 2:49pm
Hi Everybody,


Update on my EA. I have been running live demo test since the market open Sunday night. The EA has done 8 trades for me so far since I started. Of those 8, 2 have been stopped out (they were on GBP/USD) and the rest remain open still having not hit the exit signal. Below is a synopsis of the current standings.

For reference: My current time is 19:30GMT

EUR/USD - Open Buy at 09:32 GMT Price - 1.2686 - Current Price 1.2739 - Up +53 pips

USD/CHF - Open Sell at 08:30 GMT Price - 1.2500 - Current Price 1.2431 - Up +69 pips

GBP/USD - Open Buy at 08:35 GMT Price - 1.8739 - Current Price 1.8867 - Up +128 pips

USD/JPY - Open Sell at 13:00 GMT Price - 118.06 - Current Price 117.67 - Up +39 pips

AUD/USD - Open Buy at 13:31 GMT Price - 0.7466 - Current Price 0.7481 - Up +15 pips

USD/CAD - Open Sell at 13:47 GMT Price - 1.1160 - Current Price 1.1161 - Down - 1 pip

Two GBP/USD trades earlier

Open buy at 06:49 GMT, stopped out -50 pip loss at 07:40 GMT
Open sell at 07:40 GMT, stopped out - 50 pip loss at 08:31GMT

These two trades came when GBP went up and down and back up over 50 pips early this morning.

Besides those two trades, the EA has performed quite well so far today. Waiting to see how the exits go.

That's my standings so far.

GirlFlyer
0 traders viewing now