Flyer! EA - New Thread! | Page 8 | 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

Fozzy, Pip Boxer, Flyer 4 replies

Help Test my new EA, FLYER! 243 replies

Flyer! Version 2.00 is released. 0 replies

Attachments: Flyer! EA - New Thread!
Exit Attachments

Flyer! EA - New Thread!

  • #141
  • Oct 20, 2006 5:36am Oct 20, 2006 5:36am
Quoting et_phonehome_2
Its annoying when trying to test EAs with the internet down or loss of connection from broker during the middle of the night...
Tell me about it.... For reliable internet connections, you generally have to have 2 different internet providers. One for primary, one for backup. This is not just for EA's. Its for your own trading. If you do a lot of trading, you need to have backup sources.
  • #142
  • Oct 20, 2006 5:39am Oct 20, 2006 5:39am
Quoting HAGeek
i noticed on another post somewhere that adding a time to trade function on the EA might be good.

i.e. only trade certain currency pairs during when the best times to trade are... to prevent trades during the horrible flat sideways movements.

Hey how about looking at the ATR value continually... if it is > than a certain amount then ok to trade..... then if there is a large move overnight for some reason we can catch it.

thoughts?
I am working currently on an improvement now to reducing trades during flat market times, basically it checks to see if the high/low range for the last several periods is very small. If so, it will ignore the entry signals. But for now, I'm still working on it.
  • #143
  • Oct 20, 2006 5:43am Oct 20, 2006 5:43am
Quoting athalon7
GirlFlyer-

Thanks so much for your work and willingness to share. I have read all of the posts and have seen the changes. I am currently testing another EA and can only test two more pairs with my current computer/connection.

I have seen all of the changes from A to B and the different settings. My request is this: If you had to pick the two currency pairs to demo test for the next 2 months, what would they be? What time period? What version of Flyer!? and what optimal settings?

I hope this is not too much to ask as I know you are busy. Best regards.
Right now, GBP/USD and USD/CHF are my two favorites as they are performing well on Version A with the settings I have given.

Only further forward testing will allow me to give any further opinion.
  • #144
  • Oct 20, 2006 9:41am Oct 20, 2006 9:41am
  •  jdl
  • | Joined Oct 2006 | Status: going to master this or die trying | 101 Posts
GirlFlyer,

I'm running Version A and Version B with suggested presets and no manual intervention for a week, will then post the results. In short, yesterday was a big down day, today is a big up day.

Is it correct to say that the trailing stop will take effect when the price has moved up or down by the TrailingStopValue, but that it only rachets the TS a fixed 10 points? If the TS is set to 60, and the currency goes up 110 pips (GBPJPY today), the trailing stop will be racheted up 10 pips one time and will be 100 pips below the current price, right? Does that risk a big drain of profits before the TS is hit?

Many thanks for all your hard work, both with the EA and with answering everyone's questions. We really appreciate it.
jdl
  • #145
  • Oct 20, 2006 10:41am Oct 20, 2006 10:41am
hi GirlFlyer,

could you post the most recent forward testing results?
  • #146
  • Oct 20, 2006 12:16pm Oct 20, 2006 12:16pm
first of all thanks GirlFlyer for the EA.

i've made some backtests for periods before 02.15.06 and results are very poor. i've got some questions.
1.what parameters did You optimized besides extern variables?
2. have You made optimization (of extern and other variables if there are others to optimize) let say for a 6 months period, than run a backtest for a month?
3. are those results in .doc achieved with parameters optimized for the same period which was used to make a backtest?
  • #147
  • Oct 20, 2006 2:40pm Oct 20, 2006 2:40pm
Please read the old and new thread of GirlFlyer as to what she has used to test. I have done back tests using the posts in her threads, and the results was very impressive. I believe too many posts on this has been beaten to death already, in backtesting with her EAs, you must use the 1 minute timeframe, other timeframes will not produce desirable modelling quality....
  • #148
  • Oct 20, 2006 3:43pm Oct 20, 2006 3:43pm
Quoting et_phonehome_2
Please read the old and new thread of GirlFlyer as to what she has used to test. I have done back tests using the posts in her threads, and the results was very impressive. I believe too many posts on this has been beaten to death already, in backtesting with her EAs, you must use the 1 minute timeframe, other timeframes will not produce desirable modelling quality....
et_phonehome_2, thank You for Your kind reply.
i've already read all previous posts (old and new thread) and i know i should use M1 tf in backtesting. my results for period mentioned in the .doc file are very impressive too. for previous dates i can not optimize parameters (extern variables) to get decent profit (win/loss ratio neither). that is why i'm asking my questions. maybe i didn't notice answers. if they are answered somewhere than very sorry.

and sorry for my english, it's not my native language.
  • #149
  • Oct 20, 2006 5:15pm Oct 20, 2006 5:15pm
Wait a minute... I think you got that mixed up in you head . If you use EVERY TICK MODELLING - method in MT4 then it uses ALL timeframes below the one you're testing on. So don't test on M1, just test in any TFrame but make sure you have set it to EVERY TICK.
BTW, you have to get good data, too (e.g. from Alpari-DataBase). Then use this data for backtesting to get a modelling quality of 90% (Anything below you can throw away!!).

I think there's material about reliable MT-backtesting (which is really not THAT easy) in the beginner's / MT section.

Quoting et_phonehome_2
Please read the old and new thread of GirlFlyer as to what she has used to test. I have done back tests using the posts in her threads, and the results was very impressive. I believe too many posts on this has been beaten to death already, in backtesting with her EAs, you must use the 1 minute timeframe, other timeframes will not produce desirable modelling quality....
  • #150
  • Oct 20, 2006 5:27pm Oct 20, 2006 5:27pm
I had to reload Flyer 1.04. I downloaded the 2 files attached to your post #146 connected to the old thread. I forgot which settings are the reccomended settings. The ones that appear in the Flyer!.doc Word document or the settings that appear as comments in the Flyer!.mg4 file.

Thanks,

dbcmkx
  • #151
  • Oct 20, 2006 5:32pm Oct 20, 2006 5:32pm
Quoting tradeigel
So don't test on M1, just test in any TFrame but make sure you have set it to EVERY TICK.
flyer uses own period for mas (ie 15 for version A) so there is no need to test it on M15. one can test it on lower tfs. M1 gives best results.

Quoting et_phonehome_2
other timeframes will not produce desirable modelling quality....
this is sentence is not valid. it should rather be: other timeframes will not produce desirable profit.... tradeigel is right modeling quality is sth different.
  • #152
  • Oct 20, 2006 7:38pm Oct 20, 2006 7:38pm
Quoting dbcmkx
I had to reload Flyer 1.04. I downloaded the 2 files attached to your post #146 connected to the old thread. I forgot which settings are the reccomended settings. The ones that appear in the Flyer!.doc Word document or the settings that appear as comments in the Flyer!.mg4 file.

Thanks,

dbcmkx
Best settings are those that appear in the .DOC file. I think I forgot to update them in the code file.

Girlflyer
  • #153
  • Oct 20, 2006 7:41pm Oct 20, 2006 7:41pm
Quoting jdl
GirlFlyer,

I'm running Version A and Version B with suggested presets and no manual intervention for a week, will then post the results. In short, yesterday was a big down day, today is a big up day.

Is it correct to say that the trailing stop will take effect when the price has moved up or down by the TrailingStopValue, but that it only rachets the TS a fixed 10 points? If the TS is set to 60, and the currency goes up 110 pips (GBPJPY today), the trailing stop will be racheted up 10 pips one time and will be 100 pips below the current price, right? Does that risk a big drain of profits before the TS is hit?

Many thanks for all your hard work, both with the EA and with answering everyone's questions. We really appreciate it.
jdl
JDL,
After the first TS level is hit, the S/L is set to + 10 pips, if it goes another level of TS above that, then the S/L is set to 10+TS pips, and so on. That way it keeps jumping the S/L as the profit increases.

Hope that helps.

Marci
  • #154
  • Oct 21, 2006 3:52am Oct 21, 2006 3:52am
Quoting GirlFlyer
Best settings are those that appear in the .DOC file. I think I forgot to update them in the code file.
Hi GirlFlyer,

May I make a suggestion, sometimes it is hard for people reading this thread (myself included) to track down where are the latest settings and where is the latest version of the EA. Maybe it would be best to keep the latest settings in the actual EA?

I, for one, just grabbed the EA and ran in on 4 demo accounts over the last week and have noticed that the results were less then good. 3 accounts n the red one in the black, but only just. And now reading the above post I have just realised that I must have been using the wrong settings. :surprised
Just my 2 pips worth
  • #155
  • Oct 21, 2006 6:54am Oct 21, 2006 6:54am
Quoting kalamari
flyer uses own period for mas (ie 15 for version A) so there is no need to test it on M15. one can test it on lower tfs. M1 gives best results.


this is sentence is not valid. it should rather be: other timeframes will not produce desirable profit.... tradeigel is right modeling quality is sth different.
Thanks for correcting that statement....
  • #156
  • Oct 22, 2006 11:46am Oct 22, 2006 11:46am
Hello Everyone:

For those of you following the development of Flyer!, I wanted to let you know I discovered a gross error in the backtests that created a situation where the profitability of the EA was grossly exaggerated. I finally discovered this problem when after doing a test run on an improvement I did to the EA caused nearly a 100% success rate. At that point I began to question everything because I could not believe that kind of success rate could be achieved. If it could, I would not be talking about it now and go into hiding making millions of dollars.

Anyway, long story short is that the EA uses the 15M timeframe internally for its calculations and runs on Strategy Tester in the M1 setting.... we all know that by now. The catch was, the data I had in my computer for each of the 2 time frames had a 2 hour differential. My 1minute data came from Alpari and the 15 minute came from IBFX. IBFX indexed their data 2 hours ahead of Alpari. Consequently, Strategy Tester was "seeing" into the future and making perfect decisions.

So, what does that mean? I'm still reanalyzing everything, but after resolving the problem, the accuracy of the EA appears to have dropped down to about 65-70% with the drawdowns being much larger. This creates serious questions as to the long term profitability of the EA at this time.

What now? I still believe in the EA and its system because I use this method in my personal trading (much more accurate btw) on a manual basis, but filter out some of the entries according to the larger picture which is much more difficult to program. Accordingly I will continue to work on improvements to the EA and release them as I see fit and show my results for those that wish to continue to follow this method. However, it may be sometime before we see results that come close to what was achieved in the early back tests.

I apologize for the error in the results as I should have been a bit more careful in examining the data and the results.

That's all for now and I'll write some more later regarding new tests and results.

Happy Trading to All!

GirlFlyer
  • #157
  • Oct 22, 2006 2:10pm Oct 22, 2006 2:10pm
It was still a major effort and a lot of blood,sweat and tears to date..in spite of the "setback"


Thank you for your willingness to share!!!

Do not give up...stick to what you believe in and what works for you.

Best of luck!!
" Never let the past hold the future to ransom "
  • #158
  • Oct 22, 2006 3:32pm Oct 22, 2006 3:32pm
I have been reading your postings and you do have it all together so can the rest of us show our appreciation for all of Marci's efforts... I am new here to this EA and i can see all the repetitions.. maybe the forum format doesn't properly lend itself to finding the summary of information posted by someone as helpful as Marci but we should read it all first instead of seeking to be spoonfed..

Appreciate it Marci.. from a newbee...

Rich
  • #159
  • Oct 22, 2006 7:57pm Oct 22, 2006 7:57pm
Marci,

I have setup the B version and noticed it stating a bid and ask but '0' for the trade. Am i missing something. I believe i've read everything in the document and looked for more at the forum. Sorry as i'm a newbee here...

Appreciate your help and all you've done

Rich
  • #160
  • Oct 22, 2006 8:17pm Oct 22, 2006 8:17pm
Quoting GirlFlyer
I apologize for the error in the results as I should have been a bit more careful in examining the data and the results.
GirlFlyer
Keep working at it. Your determnation to succeed reflects your integrity.
Thanks for your efforts, and I'm grateful for your honesty.
0 traders viewing now