Phoenix is here! | Page 34 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Phoenix Forex Manager trading journal 25 replies

Phoenix Foundation 1 reply

Phoenix 2007 (new thread) 974 replies

Can someone fix the Phoenix EA? 0 replies

Here is a free Calculator...I don't know if its been posted here before 3 replies

Attachments: Phoenix is here!
Exit Attachments

Phoenix is here!

  • #661
  • Sep 19, 2006 5:37pm Sep 19, 2006 5:37pm
Hendrick,

I have been reading this thread with real interest and I am waiting with bated breath for Phoenix 4. I am new to EAs and I'm glad I came across this thread early on in the piece.

Glad to see some people are still sharing great things on this fourm

May the pips be with!
  • #662
  • Sep 19, 2006 7:47pm Sep 19, 2006 7:47pm
Hendrick,
I was thinking of trying something that sounds similar to what you are talking about.
My plan was to have multiple forward tests in progress feeding into database that filtered out the best current settings. Basicly an ongoing forward optimisation test. I would map the progress of the settings over time to see if the best ones had changed and then switch the live account to them.
There would be a bit of lag between when new settings were performing best and when they are confirmed.
I suspect this is what many people do at the moment manually, but i guess where my approach would be different is it would be using a database to track things over time and get daily reports on current trends for the settings.

I was thinking this would give better results than back testing regularly.

Hendrick, Is this similar to what you are doing with phoenix 4?

Regards
thebts
  • #663
  • Sep 19, 2006 10:58pm Sep 19, 2006 10:58pm
Quoting thebts
Hendrick,
I was thinking of trying something that sounds similar to what you are talking about.
My plan was to have multiple forward tests in progress feeding into database that filtered out the best current settings. Basicly an ongoing forward optimisation test. I would map the progress of the settings over time to see if the best ones had changed and then switch the live account to them.
There would be a bit of lag between when new settings were performing best and when they are confirmed.
I suspect this is what many people do at the moment manually, but i guess where my approach would be different is it would be using a database to track things over time and get daily reports on current trends for the settings.

I was thinking this would give better results than back testing regularly.

Hendrick, Is this similar to what you are doing with phoenix 4?

Regards
thebts
Hi thebts,

Your idea sounds interesting. I'm still using the old fashioned way. If it's possible then please tell me more about your idea.<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o></o>

<o></o>

I do a manually backward test for the last 6 months for every pair. If there's a smooth upward curve from the beginning till the end then it's ok to use the settings. Every 4 weeks I do the same test again. If the curve is going downwards at the end I do a new backward test for the last 3 months to find better settings. Works ok for now.<o></o>

<o></o>

Of course all tests needs a modelling quality of 90%<o></o>

<o> </o>
  • #664
  • Sep 20, 2006 10:27am Sep 20, 2006 10:27am
It's not something i have had time to do, but here is what i was thinking of:
Along side the live account i would have a number of demo accounts collecting forward testing data in a database. Each demo account would run a number of different settings and pairs. To keep the number of demo accounts down i would probably use multiple magic numbers on the same pair. This is really like your forward version of MT4's optimisation test.

Along with recording the actual settings used in the database I would automaticly collect data such as: max draw down of a trade, number of hours open, dates, profit/loss, etc.
To do this i would build it into the ea to log to a database all this information.

With all this feeding into a database i would be collecting a lot of info. There would be some reports that i would run each day looking for patterns and mapping any trends in what settings are working best. I was thinking of some 3D models here, but there would be a number of ways to show the collated information. Last time i looked MT4s analysis capabilities of backtesting data were pretty limited, what i had imagined would be much more comprehensive with the ability to map and filter on all the extra info i have collected.

Hopefully it would provide enough feedback to adjust the settings on an ongoing basis.

Once it is set up the main leg work would be keeping your demos running and maybe adding new settings to test.
  • #665
  • Edited 2:09pm Sep 20, 2006 11:52am | Edited 2:09pm
Quoting Hendrick
What is <st1:city w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:city> 4 all about? In the first place it has nothing to do with Firebird (so please don’t discuss Firebird in this thread) and secondly it’s completely different from <st1:city w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:city> 1, 2 and 3.<o></o>

<o></o>

For <st1:city w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:city> 4 I’ve tried a different approach: whatever strategy you are using, they have all one thing in common: you’re studying the past in order to predict the future. <st1:city w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:city> 4 is not different from that, only with one exception: optimum settings! With optimum settings I mean a special mix between, Number of Trades, Net Profit and Maximal Drawdown. When you have found these settings for lets say the last 6 months and you have a smooth upwards curve (the past) you may expect that for the coming weeks (the future) this upwards curve will continue and the EA will make more winners than losers. And that’s what we want an EA to do! <o></o>

<o></o>

A lot of EA’s I’ve seen had the same problems: fixed settings that worked for the past but if the settings will not be adjusted the EA will fail sometime in the future ( Firebird is an example of this).<o></o>

<o></o>

On all the forums there is a lot of talking about the “holy grail”. In my opinion Metatrader with its ability to make Experts Advisors, is the grail and you only have to make the right EA with the right settings (and adjust them in time!) to make it a “ holy grail”. <o></o>

<o></o>

One thing I’ve discovered is, that by tweaking the settings of an EA you are able to discover strategies you have never heard of before! These are all new strategies and only possible to discover by using Metatrader!

If you are back testing you have the possibility to open a chart after testing. I did it several times for <st1:city w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:city> 4 and it showed me nothing! It was a complete mess! But still it has this straight upwards curve and <st1:city w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:city> 4 made a lot of money in these 6 months. How is this possible? It’s possible because there are a lot more of strategies in Forex than taught in all these Forex-books and you have now the possibility to discover these strategies by using the power of your PC and the Expert Advisors of Metatrader!<o></o>
<o></o>

Ok this is my point, so now you can shoot at it.<o></o>

<o></o>

Let me give some examples: I’m using a TP=42 and a SL=84 for all pairs (I will explain later why I’m doing this) and nowhere in <st1:city w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:city> 4 have I programmed something like resistance and support lines. But if I’m looking at my open trades, I see that SL and TP are functioning as support and resistance lines. Weird, not? So it seems that the EA has made its own strategy. Stay tuned and more tomorrow! I think we can have a lot of fun with <st1:city w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:city> 4!!<o></o>

Hendrick,

I am impressed by your observations and way of thinking! I think you are near to invent something great.
Thanks,

A1Cruiser
  • #666
  • Sep 21, 2006 1:45pm Sep 21, 2006 1:45pm
Just a word about tweaking here. This is a daily chart of the EUR/USD with a 10 period average true range plotted. The higher the ATR = the more price movement, from the high to low of the day. The boxes on the ATR show a low ATR and high ATR time period. I'm working on a range breakout system that uses a predetermined take profit level. If I see that the ATR trend is on the rise then I may increase my TP level. If I see that the ATR is going down, I may lower my TP level. Is this tweaking? Yes. But, how do you know when to adjust your EA parameters? If a backtest on three years of data shows that the max consecutive losses was say, 7. And, you have a current consecutive loss of 5 trades should you be concerned? I don't think so. If you hit 10 in a row then it's time to change your parameters. I think the main question is trying to figure out when to tweak things. Whatever Hendrick is brewing with P4 I'm sure it will be a great piece of work. There may be some that question why should any EA have to be 'tweaked' and the answer is simple. It's because price action alone dictates how an EA will perform. If there is a global economic summit meeting for 5 days and there is no eco news to be released during that same week, no earnings releases, terrorist or oil news I would expect the ATR to be low. I would tweak my take profit zone to the low end because of the probable lack of movement in the markets during that week. It's basically the same as shutting off your EA before a NFP release or FOMC statement. I'm just throwing in my 2 cents here.


Good Trading To All
Attached Image
  • #667
  • Sep 21, 2006 11:13pm Sep 21, 2006 11:13pm
Quoting Palmer
Just a word about tweaking here. This is a daily chart of the EUR/USD with a 10 period average true range plotted. The higher the ATR = the more price movement, from the high to low of the day. The boxes on the ATR show a low ATR and high ATR time period. I'm working on a range breakout system that uses a predetermined take profit level. If I see that the ATR trend is on the rise then I may increase my TP level. If I see that the ATR is going down, I may lower my TP level. Is this tweaking? Yes. But, how do you know when to adjust your EA parameters? If a backtest on three years of data shows that the max consecutive losses was say, 7. And, you have a current consecutive loss of 5 trades should you be concerned? I don't think so. If you hit 10 in a row then it's time to change your parameters. I think the main question is trying to figure out when to tweak things. Whatever Hendrick is brewing with P4 I'm sure it will be a great piece of work. There may be some that question why should any EA have to be 'tweaked' and the answer is simple. It's because price action alone dictates how an EA will perform. If there is a global economic summit meeting for 5 days and there is no eco news to be released during that same week, no earnings releases, terrorist or oil news I would expect the ATR to be low. I would tweak my take profit zone to the low end because of the probable lack of movement in the markets during that week. It's basically the same as shutting off your EA before a NFP release or FOMC statement. I'm just throwing in my 2 cents here.


Good Trading To All
Hey Palmer!

I fully agree with you tweaking-story. Tomorrow I will send you the BETA of Phoenix 4 with a lot of twaeking inside. Hope you can test it for the coming weeks.
  • #668
  • Sep 22, 2006 5:16am Sep 22, 2006 5:16am
Fire away...Did you try the strawberries yet?
  • #669
  • Sep 22, 2006 9:14am Sep 22, 2006 9:14am
Statement first week with <?xml:namespace prefix = st1 ns = "urn:schemas-microsoft-com:office:smarttags" /><st1lace w:st="on"><st1:City w:st="on">Phoenix</st1:City></st1lace> 4.2. First impression: I think <st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City> 4 and the JPY likes each other very much. I'm now testing all pairs for another week to see which pairs are most suitable. Let's see how it goes! For the rest: well, I think it's not too bad for one week. This weekend more work to do: I will tweak the settings for most pairs to see if there's more profit to obtain.<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o></o>

<o></o>

(Palmer: aside testing <st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City> I will definitely test your strawberries this weekend!)<o></o>
Attached File(s)
File Type: zip Phoenix4_B.zip   8 KB | 315 downloads
  • #670
  • Sep 22, 2006 10:53am Sep 22, 2006 10:53am
Hendrick,

These are really impressive results! If this kind of profit level can be stabilized (in average and approximately, naturally) that is really worth the tweaking periodically.
  • #671
  • Sep 22, 2006 9:19pm Sep 22, 2006 9:19pm
The last 5 trades made by Phoenix 4.2 for the GBPJPY:
Attached Image
  • #672
  • Sep 23, 2006 6:06am Sep 23, 2006 6:06am
Hendrick you're such a tease!
  • #673
  • Sep 23, 2006 6:27am Sep 23, 2006 6:27am
Let’s talk about MetaTrader Strategy Tester, the Past, the Future and <?xml:namespace prefix = st1 ns = "urn:schemas-microsoft-com:office:smarttags" /><st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City> <st1:metricconverter w:st="on" ProductID="4. In">4. In</st1:metricconverter> my previous posts I told you that there are more strategies possible in Forex than written in all the Forex-books in the world. I think strategies are not the right word. Laws or rules is a better word. <?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o></o>

<o> </o>

I’m one of these guys that believes that everything in this world is subject to rules and laws. If you hold an apple and drop it will fall to earth. It did the same in the past and it will do so in the future. That’s the law of gravity. Forex in general is also subject to laws and rules. We still don’t know what these laws and rules really are, but we can see the effects. The same is valid for gravity: we know that the apple will fall down, bit the real nature of gravity (and electricity, and magnetism etc etc) is still unknown. <o></o>

<o> </o>

When we study are charts (the past) we are trying to predict the future from that. What we really do is studying the effects of the rules and laws of Forex. Is it really necessary to know the laws and rules of Forex? No! Studying the effect is all we have to know. Back to the apple again, we don’t know the real nature of gravity, but what we do know is the apple will fall down. It did the same in the past and will do the same in the future (I hope).<o></o>

<o> </o>

When you are using the Strategy Tester of MetaTrader you sometimes discover settings that gives a completely different effect than the settings nearby. That’s the moment that you discovered the effect of one off the Laws of Forex (the apple fall down). In the tester you have the ability to open a chart after a test. In this case you probably will see nothing. It all looks like a complete mess. And this is my point: a lot of Forex strategies are hidden, can’t see on the chart, but they do reveal them in the tester.<o></o>

<o> </o>

And that’s the idea behind <st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City> 4: keep on testing and adjusting the settings until you have a smooth up going balance-curve. There’s no doubt about it that this balance-curve will continue in at least the coming four weeks. My forward tests shows me that I’m on the right track. <o></o>
  • #674
  • Sep 23, 2006 6:50am Sep 23, 2006 6:50am
Quoting Palmer
Hendrick you're such a tease!
Palmer, you won't believe this! My green-grocer couldn't deliver strawberries today. I told him about Forex, <?xml:namespace prefix = st1 ns = "urn:schemas-microsoft-com:office:smarttags" /><st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City>, Palmer and your recipe and he didn’t have a clue where I was talking about (well, I can't blame him). He did however promise me to deliver the strawberries next week.<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o></o>

<o></o>

I will mail the BETA-version of <st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City> 4.2 to you within the coming hour. I hope you can test this one on a demo-account in the coming week. <o></o>

<o></o>

Let’s discuss your findings in this thread.<o></o>

<o></o>

When using on a demo-account:<o></o>

PrefSettings=true<o></o>

MM=true<o></o>

<o></o>

For testing:<o></o>

PrefSettings=false<o></o>

MM=false
<o></o>

You're the first one Palmer!<o></o>
  • #675
  • Sep 23, 2006 7:00am Sep 23, 2006 7:00am
Tradeigel,

I just send you a copy of Phoenix 4.2 BETA. Lets find the right settings!!
  • #676
  • Sep 23, 2006 8:04am Sep 23, 2006 8:04am
alright, got it! Thank you very much. I look forward to testing it!
  • #677
  • Sep 23, 2006 9:02am Sep 23, 2006 9:02am
Phoenix 4.2 is almost finished. If you want a copy to test it please PM me your e-mail address.

(Some of you send their e-mail address earlier but I had a disk-crash and lost these addresses. Yes I know.. I should have made a backup, but well ...I didn't. Sorry about that!)

Hendrick
  • #678
  • Sep 23, 2006 9:45am Sep 23, 2006 9:45am
For those of you who received a copy of Phoenix 4.2 I have a request:
for some reason I can't find the right settings for the EURGBP. If you can please let me know:

These are the "rules":

backward test for the last 6 months
modelling quality 90%
at least 60 trades
MD below 10%
balance-curve upwards (of course!)

Who is the first one!
  • #679
  • Sep 23, 2006 10:05am Sep 23, 2006 10:05am
Hendrick:

What timeframe should we use for testting???
  • #680
  • Sep 23, 2006 10:36am Sep 23, 2006 10:36am
Quoting Palmer
Hendrick:

What timeframe should we use for testting???
Hey Palmer! Nice to see that you've finally dropped you "company" and have the time for some serious business! TF=15M.<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o></o>
2 traders viewing now