Phoenix is here! | Page 16 | 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!

  • #301
  • Jul 15, 2006 12:12pm Jul 15, 2006 12:12pm
Quoting mono
Hi Zarni

Well, I agree with you to a certain degree, it does sound scary. I am waiting for that day and until it comes, hopefully i'll have enough funds to cover my losses.

What do you define as short term? I'm currently testing Firebird ( into its 5th week) and with its record of 171/1. I had my doubts, so I started a another test and then another, all with the same/similar results. You cannot ignore the results. Its doing something right.

Hopefully Phoenix will do the same or similar job.

P.S.
If u can make any system work at near to no loss, I would like u to show it. Maybe I'll use yours to go live.

Mono
Hey mono,
I can literally give you 10s of systems that would book similar results with a SL of 300. That is not saying that Phoenix is not good system. I am not knocking the system. Just knocking the method that you are using
All that needs to happen is when you go live for the market to go in strong trend and your account can be wiped out pretty soon with that kind of SL.
You could do 300 SL but you need to trade microlots on a 5-10k account otherwise the drawdown and risk can kill you.
I do have some systems that are a lot less drawdown however they are very complex and probably impossible to be coded in EA.
  • #302
  • Jul 15, 2006 12:27pm Jul 15, 2006 12:27pm
Quoting Yardie
Mono but your results are a bit deceiving because of the big floating loss. Furthermore one losing trade can would wipe out the equivalent of 10 winning trades.
The more volatile pairs with bigger ranges (GBP and CHF) seem to do the best with the 150 or 300 S/L and the wins are consistently more than 10 to 1 on all the timeframes I've tried. Just my observation on backtesting v1.1 so far.
  • #303
  • Jul 15, 2006 1:34pm Jul 15, 2006 1:34pm
This looks interesting, Hendrick. Are you proposing this in addition to RVI or in place of? I've placed it on a chart with the old firebird indicator, and it looks like the best trades would be initiated during the yellow areas, which by definition seem to be the ranging areas.

Quoting Hendrick
I need some help. The BUY and SELL signals of <?xml:namespace prefix = st1 ns = "urn:schemas-microsoft-com:office:smarttags" /><st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City> are generated by the following piece of code:<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o></o>


<o></o>

MA =iMA(NULL,MA_Timeframe,MA_Length,0,MODE_SMA,PRICE_OPEN,0);<o></o>

RVI=iRVI(NULL,0,10,MODE_MAIN,0)-iRVI(NULL,0,10,MODE_MAIN,1);<o></o>

<o></o>

The MA is the heart of <st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City> and the RVI (Relative Vigor Index) is an additional signal. For a BUY, RVI needs to be > 0 and for a sell <0. I’m not happy with this because it’s far to simple. What I want is an additional signal that takes account of the trend. So no SELL if the trend is up and no BUY if the trend is down. I think this can be a major improvement.<o></o>

<o></o>

I think a simple fast and slow EMA will do. I have made an indicator (PhoenixAdditionalIndicator) that makes it visible what this signal will do. If it’s red, only SELL’s are valid, green for BUY’s and yellow for BUY’s and SELL’s.<o></o>

<o></o>

<o></o>
<o></o>
  • #304
  • Jul 15, 2006 1:52pm Jul 15, 2006 1:52pm
Quoting BenQ
This looks interesting, Hendrick. Are you proposing this in addition to RVI or in place of? I've placed it on a chart with the old firebird indicator, and it looks like the best trades would be initiated during the yellow areas, which by definition seem to be the ranging areas.
Hi Ben,

In version 2 this new additional signal will replace RVI.
  • #305
  • Jul 15, 2006 5:03pm Jul 15, 2006 5:03pm
Quoting Hendrick
I need some help. The BUY and SELL signals of <?xml:namespace prefix = st1 ns = "urn:schemas-microsoft-com:office:smarttags" /><st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City> are generated by the following piece of code:<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o></o>

<o></o>

MA =iMA(NULL,MA_Timeframe,MA_Length,0,MODE_SMA,PRICE_OPEN,0);<o></o>

RVI=iRVI(NULL,0,10,MODE_MAIN,0)-iRVI(NULL,0,10,MODE_MAIN,1);<o></o>

<o></o>



<o></o>

Hendrick,

Firebird v63g had a trend code in it, but wasn't used. I use it in a modiifed Firebird and it seems to work in stopping trades during trends.

extern int DVLimit = 10; // included by Renato

int TrendUp=0, Trenddown=0;
double iFXAnalyser(int FXA_Period, int mode, int shift)// Made local function by MrPip

if ( (iFXAnalyser(0,MODE_DIV,0)>DVLimit*Point && iFXAnalyser(0,MODE_SLOPE,0)>0 )) Trendup=1;
if ( (iFXAnalyser(0,MODE_DIV,0)<-DVLimit*Point && iFXAnalyser(0,MODE_SLOPE,0)<0 )) TrendDown=1;

Then pur "&& TrendUp !=1" following DealTime==1 in SELL trigger. And "&& TrendDown !=1" following DealTime==1 in BUY trigger.

Another method is to put a controlling factor for RV! to regulate trading activity.

extern double RVI_Factor =50; // 0-200 factor added to RVI value to regulate trading activity

Then in SELL and BUY triggers change RVI codes to "RVI<(0-(RVI_Factor*Point))" and "RVI>(0+(RVI_Factor*Point))" respectifully.

Another method is to put a second larger iMA envelope to tighten the active trading area to encourage trading during non-trending periods.

extern double PercentLimit = 0.15; // 2nd higher envelope (band) that stops trading.

int Safe2=0;
if ( myMA*(1+(PercentLimit/100))<=Bid || myMA*(1-(PercentLimit/100))>=Ask) Safe2=1;

Put "&& Safe2==1" in SELL and BUY triggers. You can attached a 2nd Firebird type Indicator set to PercentLimit to see active trading area.

I use one Firebird with all three filters and it averages ~50% less trades with ~90% win rate. The problem I have is the effect news has on the market. I've reported to this thread, that the last 4 Fridays had a big negative effect because of carry over open trades from Thursday. I am testing eliminating trading days and/or times to see if I can better manage the historic group loses.

Hendrick, this would be a further development of your trading time results chart and it would be a lot of work. But, I would like to have an EA that is 6/24 trading time programable and can survey its history, determine bad trading times & stop trading during these times, while maintaining a ghost trade history for these times. When historic ghost trades are again profitable, these times are re-established as good trading times. Wishful thinking, but I'm going to try to work on this or see if a coder can make this work. Right now, I use yor type of chart and try to do this manually.

Way too much typing. Well, this is what happens when you drink wine while on a thread. If all of this doesn't make sense, it's the wine's fault.

Wackena
  • #306
  • Jul 15, 2006 6:01pm Jul 15, 2006 6:01pm
I like the idea of the larger envelope. I mentioned that possibility early in the firebird thread, but didn't have the ability to code it. I gotta learn to code MT4. Since finding MT4, I'm no longer using esignal at all -- which is a shame, because I was getting pretty good at coding that. Oh, well. Great work, Wackena.

Quoting wackena

Another method is to put a second larger iMA envelope to tighten the active trading area to encourage trading during non-trending periods.

extern double PercentLimit = 0.15; // 2nd higher envelope (band) that stops trading.

int Safe2=0;
if ( myMA*(1+(PercentLimit/100))<=Bid || myMA*(1-(PercentLimit/100))>=Ask) Safe2=1;

Put "&& Safe2==1" in SELL and BUY triggers. You can attached a 2nd Firebird type Indicator set to PercentLimit to see active trading area.



Wackena
  • #307
  • Jul 15, 2006 6:57pm Jul 15, 2006 6:57pm
Quoting wackena
Hendrick,

Firebird v63g had a trend code in it, but wasn't used. I use it in a modiifed Firebird and it seems to work in stopping trades during trends.

extern int DVLimit = 10; // included by Renato

int TrendUp=0, Trenddown=0;
double iFXAnalyser(int FXA_Period, int mode, int shift)// Made local function by MrPip

if ( (iFXAnalyser(0,MODE_DIV,0)>DVLimit*Point && iFXAnalyser(0,MODE_SLOPE,0)>0 )) Trendup=1;
if ( (iFXAnalyser(0,MODE_DIV,0)<-DVLimit*Point && iFXAnalyser(0,MODE_SLOPE,0)<0 )) TrendDown=1;

Then pur "&& TrendUp !=1" following DealTime==1 in SELL trigger. And "&& TrendDown !=1" following DealTime==1 in BUY trigger.

Another method is to put a controlling factor for RV! to regulate trading activity.

extern double RVI_Factor =50; // 0-200 factor added to RVI value to regulate trading activity

Then in SELL and BUY triggers change RVI codes to "RVI<(0-(RVI_Factor*Point))" and "RVI>(0+(RVI_Factor*Point))" respectifully.

Another method is to put a second larger iMA envelope to tighten the active trading area to encourage trading during non-trending periods.

extern double PercentLimit = 0.15; // 2nd higher envelope (band) that stops trading.

int Safe2=0;
if ( myMA*(1+(PercentLimit/100))<=Bid || myMA*(1-(PercentLimit/100))>=Ask) Safe2=1;

Put "&& Safe2==1" in SELL and BUY triggers. You can attached a 2nd Firebird type Indicator set to PercentLimit to see active trading area.

I use one Firebird with all three filters and it averages ~50% less trades with ~90% win rate. The problem I have is the effect news has on the market. I've reported to this thread, that the last 4 Fridays had a big negative effect because of carry over open trades from Thursday. I am testing eliminating trading days and/or times to see if I can better manage the historic group loses.

Hendrick, this would be a further development of your trading time results chart and it would be a lot of work. But, I would like to have an EA that is 6/24 trading time programable and can survey its history, determine bad trading times & stop trading during these times, while maintaining a ghost trade history for these times. When historic ghost trades are again profitable, these times are re-established as good trading times. Wishful thinking, but I'm going to try to work on this or see if a coder can make this work. Right now, I use yor type of chart and try to do this manually.

Way too much typing. Well, this is what happens when you drink wine while on a thread. If all of this doesn't make sense, it's the wine's fault.

Wackena
That's a great idea. You will probably want to write statistics to a file and read from it. Don't know how to do that in MQL. Hope that is possible.
  • #308
  • Jul 16, 2006 12:39am Jul 16, 2006 12:39am
Quoting wackena
Hendrick,

If I read the code correctly in Phoenix v1.1, if using Micro account and MM that calculates, using 5% Risk, the lot size to 0.025, then the code below will return LotMM=0.

lotMM = MathRound(lotMM*10)/10; MathRound(0.025*10) will return 0.

You might consdier using the NormalizeDouble() function to round the lots. You can set number of decimal points needed.

if(Lots>0.1) Lots=NormalizeDouble(Lots,1); else Lots=NormalizeDouble(Lots,2);

Is my assumption above correct and could this code be useful?

Wackena
Hi Wackena,

MM is not using Lots from the input parameters. The MM code is calculating the lot size for a normal account and this piece of code is adjusting the lot size for a micro-account:
if(MM)
{
Lots = subLotSize();
if(AccountIsMicro)
{
Lots=Lots/10;
}
}

I've tested it and it worked as it should. Thanks voor your contribution!
  • #309
  • Jul 16, 2006 5:32am Jul 16, 2006 5:32am
Quoting Hendrick
I've tested it and it worked as it should. Thanks voor your contribution!
Was that Dutch?

PS: I will look into you new MA-Filter. As I am capable of coding MQL4 I decided to go more into development of Phoenix. I hope I will have some ideas for you then. It's worth it - Phoenix is really good!
  • #310
  • Jul 16, 2006 7:12am Jul 16, 2006 7:12am
Quoting Hendrick
Hi Wackena,

MM is not using Lots from the input parameters. The MM code is calculating the lot size for a normal account and this piece of code is adjusting the lot size for a micro-account:
if(MM)
{
Lots = subLotSize();
if(AccountIsMicro)
{
Lots=Lots/10;
}
}

I've tested it and it worked as it should. Thanks voor your contribution!
Hi Hendrick,

If you backtest with $500 deposit, such as a Micro (Mini) accout minimum deposit, with AccountIsMicro=true, MM=true and Risk=5, there are no orders placed. You need Risk=9 to get a 0.01 lot.

double lotMM = MathCeil(AccountFreeMargin() * Risk / 1000) / 100;
double lotMM = MathCeil(500 * 5 / 1000) /100 = 0.03

Then with RoundupLots==true

lotMM = MathRound(lotMM*10)/10;
0.03 = MathRound(0.03*10)/10 = 0


Wackena
  • #311
  • Jul 16, 2006 7:30am Jul 16, 2006 7:30am
Quoting zarni
That's a great idea. You will probably want to write statistics to a file and read from it. Don't know how to do that in MQL. Hope that is possible.
We have the existing History file as the source file. We need code to scan this file, locate loses, sort loss time frequencies and adjust trading time scheduler accordingly in real time. Then place ghost trades in these time zones, keep in a separate file. Then scan, sort and re-adjust trade scheduler accordingly.

Wackena
  • #312
  • Jul 16, 2006 8:49am Jul 16, 2006 8:49am
Of course, all of that is possible in MQL4. However, these kind of self-adjusting, forward-optimization EAs don't bring as much improvment as you could think. I coded a few...
  • #313
  • Jul 16, 2006 9:00am Jul 16, 2006 9:00am
Quoting wackena
Hi Hendrick,

If you backtest with $500 deposit, such as a Micro (Mini) accout minimum deposit, with AccountIsMicro=true, MM=true and Risk=5, there are no orders placed. You need Risk=9 to get a 0.01 lot.

double lotMM = MathCeil(AccountFreeMargin() * Risk / 1000) / 100;
double lotMM = MathCeil(500 * 5 / 1000) /100 = 0.03

Then with RoundupLots==true

lotMM = MathRound(lotMM*10)/10;
0.03 = MathRound(0.03*10)/10 = 0


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

<o></o>

I believe you. I'm just too tired at the moment to check. Have been testing <?xml:namespace prefix = st1 ns = "urn:schemas-microsoft-com:office:smarttags" /><st1lace w:st="on"><st1:City w:st="on">Phoenix</st1:City></st1lace> all day. I have so many ideas for improving <st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City> that I don't know where to start. I'm working hard on version 2 and version 3 at the same time. Well, next week first my holydays and after that that back to <st1:City w:st="on"><st1lace w:st="on">Phoenix</st1lace></st1:City>. BTW I really liked one of your previous posts where you told us that you were enjoying a glass of wine while writing your message! I have to make a confession: while I'm writing this I'm doing just the same (I love Cabernet Sauvignon). Is it a habit of the visitors of this thread or just a coincidence? I really appreciate all your contributions to this thread and I will drink my next glass of wine in honour of you. Cheers mate!<o></o>

<o> </o>

!
  • #314
  • Jul 16, 2006 9:01am Jul 16, 2006 9:01am
Quoting tradeigel
Was that Dutch?

PS: I will look into you new MA-Filter. As I am capable of coding MQL4 I decided to go more into development of Phoenix. I hope I will have some ideas for you then. It's worth it - Phoenix is really good!
You got me there Tradeigel!! Yes I'm Dutch!
  • #315
  • Jul 16, 2006 10:28am Jul 16, 2006 10:28am
ha, that's cool, so you can write on in Dutch now, works for me...Hehe, you can post some of your ideas if you want, maybe I can help you. Right now, I am trying to improve Phoenix, too.
  • #316
  • Jul 16, 2006 1:01pm Jul 16, 2006 1:01pm
I am trying out Mono's wide stoploss method this week and will be using phoenix with 25 tp 250 SL. Lets see how it goes. Mono's results so far is too good to ignore.
  • #317
  • Jul 16, 2006 1:37pm Jul 16, 2006 1:37pm
Quoting Yardie
I am trying out Mono's wide stoploss method this week and will be using phoenix with 25 tp 250 SL. Lets see how it goes. Mono's results so far is too good to ignore.
I backtested 25/250 on GBP from 1/1/06 to today and its basically break-even (you make $250 on 10k). 30/300 is up about 50% so there's quite a difference. I've found GBP and CHF to be the best candidates from backtests. I'm gonna demo 30/300 on GBP starting this week.
  • #318
  • Jul 16, 2006 1:52pm Jul 16, 2006 1:52pm
Hey Yardie, I agree, seems hard to ignore. But it also seems like we've done these settings before in the early days of firebird and it eventually crashed -- which is what gave rise to Phoenix. I guess the difference now would be no pipstep -- which really! hurt in the strong trends -- and the addition of the new SafePipsMinutes/Distance feature and UseCloseSignal. Hopefuly those along with the SafeArea will make a real difference this time. What concerns me though is his open orders can get quite huge. I'm not sure, but I think I noticed on one of his posted results that his negative open orders were almost as large as all the positive closed trades.

Quoting Yardie
I am trying out Mono's wide stoploss method this week and will be using phoenix with 25 tp 250 SL. Lets see how it goes. Mono's results so far is too good to ignore.
  • #319
  • Jul 16, 2006 2:40pm Jul 16, 2006 2:40pm
In backtesting I get bad results with CloseSignal = TRUE but really good results with CloseSignal = FALSE. Anybody experiencing this on a live account?
  • #320
  • Jul 16, 2006 2:49pm Jul 16, 2006 2:49pm
Quoting jhathawa
I backtested 25/250 on GBP from 1/1/06 to today and its basically break-even (you make $250 on 10k). 30/300 is up about 50% so there's quite a difference. I've found GBP and CHF to be the best candidates from backtests. I'm gonna demo 30/300 on GBP starting this week.
Guys, sorry but you must be joking right...any system with an SL of 300 cannot be viable...THere will be an inevitable time when it hits 2 or 3 bad trades and that would be it...
0 traders viewing now