Firebird EA - fixed version - | Page 26 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Firebird EA, please test this one, it is one of the greats 1,228 replies

which version of firebird are you using on live account? 12 replies

Firebird V1.0 an amazing EA ! 33 replies

Going live with Firebird 38 replies

My firebird is a roast cornish game hen 3 replies

Attachments: Firebird EA - fixed version -
Exit Attachments

Firebird EA - fixed version -

  • #501
  • Aug 30, 2006 10:51am Aug 30, 2006 10:51am
I'm running v63G_fixed. I'm using Mono's settings for the s/l, tp. I trading on a mini demo account with IBFX, using .6 lots on each currency. I'm trading USD/CAD, GBP/USD, USD/JPY, EUR/JPY, EUR/USD, USD/CHF. I started with $7000, on 8/22. The largest negative float I've seen in the week I've been running it is eariler today with -$422.

Balance $8775.52
Equity $8522.39
Margin $510
Free Margin $8009.39
Float -$259.66

I'm really pleased with how this is performing, but I think the key to trading FB is to close out all open trades periodically and start anew. It probably makes the most sense to do this a day or so before large news events. And stand aside for a bit while the market settles down and then start FB once again.

I'm testing FB using different settings, but I think I like this on the best because of the small negative float. I'm preparing to go live with this one on Sunday.
  • #502
  • Aug 30, 2006 11:01am Aug 30, 2006 11:01am
Maybe we should do some historical float research.
I have no idea how to code it but I would like the backtester to print to a csv file which can be imported into Excel. The EA should run on D1 because else we get a overload of data.
The .csv file should contain:

- Date.
- Negative float.
- Positive float.

Each pair it's own data .csv
Plus a combined .csv

Then locate the point with a high float and so some historical research.
Things like:
- What day triggered the float?
- Was it related to news?
- Was it related to low volume? (holiday season)
- Are some pairs more resitant to float than others?
- Do the floats of certain pairs even eachother out?
- etc
May the pips be with you.
  • #503
  • Edited 2:14pm Aug 30, 2006 11:52am | Edited 2:14pm
Quoting The Wizard of Oz
Yes it's like living in fantasy land. Just wondering when the storm will hit. When will the bird die?
Well, according to my back testing Firebird may not die, but it certainly may look like a lame duck for extended periods. Took me awhile to get around to it, but I've finally managed to test 3 of better performing pairs lately... EUR/USD, USD/CAD, and USD/JPY back to 12/1/2004. Not that back testing is everything, but I think the larger the time frame (due to more data points), the better idea how a system is likely to behave... at least as long as my test data is good, which I have no idea. (got it from http://www.alpari-idc.com/en/dc/databank.php )

I did my testing with the current popular settings (T/P = 30, S/L = 300 on the 1 minute chart, etc.). The results show a slow bleed from 12/04 - 6/05. It may be longer but my data doesn't go back any further.

I'm not sure what changed about the market conditions in June '05, but it must have been significant because results are definitely different from June 05' - present. And the difference seems to affect all three pairs similarly (probably because they're USD based). I'm thinking that whatever conditions contributed to Firebirds poor performance during that time are likely to come around again. Maybe soon.

Right now, my opinion is that this is a good EA in that it doesn't seem to lose money when left alone (many do), but over the long haul it isn't going to take anyone from "rags to riches". I do think that returns can be improved with smart manual intervention, but those improvements will be incremental not exponential. I hope this doesn't pour cold water on anyone's party... And it is my hope that somebody can provide different/better back test results that will refute my tests.

Edit: Got a little more data back to 6/04. Will add this to my testing...
Edit # 2: After some more testing, I have a pretty good suspicion the test data from Alpari is FUBAR'd prior to 10/05, so forget I said anything. ;-)
  • #504
  • Aug 30, 2006 12:42pm Aug 30, 2006 12:42pm
hi

mid week update of my 5 version which i am testing.

up to now is my favorite the v63g-TS and the 1a, but the 1a has at the moment a lot of float but ~the half of it is only due to GBP/JPY. i have removed today this pair from all 5 versions.

i also have noticed that the v1.0c had today 2 EUR/JPY losses. is anybody else testin v1.0c who can confirm those loosers today?
Attached File(s)
File Type: zip Firebird 1a.zip   9 KB | 467 downloads
File Type: zip Firebird 1b2.zip   12 KB | 489 downloads
File Type: zip Firebird 1c.zip   13 KB | 492 downloads
  • #505
  • Aug 30, 2006 12:43pm Aug 30, 2006 12:43pm
last but not least
Attached File(s)
File Type: zip Firebird 63g-TS.zip   9 KB | 587 downloads
File Type: zip firebird 1.0a MONO settings 10 pairs.zip   15 KB | 550 downloads
  • #506
  • Aug 30, 2006 12:59pm Aug 30, 2006 12:59pm
I've been lurking a little bit.

I have a suggestion on the issue about closing down the EA before news, etc. The EA would still manage stop losses, so, your positions should exit if it goes against you. It would also manage TP as well, if the news is with you.

It seems that the biggest concern about news trades is opening new positions either because of the news, or just before the news.

Instead of closing everything down, why not just have a boolean flag that tells Firebird to not make any new orders. Turn the flag on a day before a big news event. Turn it off a day after the news.

Anyway, I'm not sure if that would really solve the issue or not. I don't know enough about how the EA works to comment intelligently.
  • #507
  • Aug 30, 2006 1:13pm Aug 30, 2006 1:13pm
Quoting stockwet
I've been lurking a little bit.

I have a suggestion on the issue about closing down the EA before news, etc. The EA would still manage stop losses, so, your positions should exit if it goes against you. It would also manage TP as well, if the news is with you.

It seems that the biggest concern about news trades is opening new positions either because of the news, or just before the news.

Instead of closing everything down, why not just have a boolean flag that tells Firebird to not make any new orders. Turn the flag on a day before a big news event. Turn it off a day after the news.

Anyway, I'm not sure if that would really solve the issue or not. I don't know enough about how the EA works to comment intelligently.

i think those open trades which will hit their SL during big news days are the one which are opened for about a week or so...
  • #508
  • Aug 30, 2006 1:39pm Aug 30, 2006 1:39pm
Assuming thats right there is no use closing positions (that are close to SL) just before news.
The news will take the stops out or if you are lucky you catch a spike that hits your TP
May the pips be with you.
  • #509
  • Aug 30, 2006 1:50pm Aug 30, 2006 1:50pm
Over time, the logic of Firebird should prevail - that a price will deviate toward the mean after a spike. FB needs to win 10 for every 1 loss to make a profit. My first demo (now just over a month old) is 308-7. the logic seems to be skewing things well in favour of take-profit. the probability of big news going against is 50%. therefore over time, it shouldn't make any difference. In fact, since FB thrives on spikes-retracements, this EA should work well on news days in the long run.

demo statement attached.

live account currently doing very well also.
Attached File(s)
File Type: zip DetailedStatement.zip   17 KB | 444 downloads
TRADE FOR SWAP
  • #510
  • Aug 30, 2006 2:07pm Aug 30, 2006 2:07pm
Yardie,

The same is here, concerning one of my tests on the 15 minutes time frame the negative float is greater than the profit having the test running for 1.5 weeks.

I think the reason for this is that when hedging Firebird 1.0b versions open two market orders (one buy and one sell) at the same time. So one of the buy or sell orders will always go against us meanwhile the other goes to profit. But one of the two orders will remain there for long and increase the float, particularly in a trending market. According to the best case scenario first the first one then the other takes profit, but in the worst case scenario the first one then the other will stop out - depending on the market movement.

Other strategies, such as news trading straddle EAs, open pending orders both sides of trade (buy and sell). Once the first pending order has been activated the other side is liquidated and the positive is used to go to profit direction. This way only one order will be activated and in a trending market it probably goes to the profitable direction and the float can be retain on a reasonable level.

Quoting Yardie
I am beginning to think version b Hedge has flaws. I am booking a lot of pips, but a lot of it is being eaten up by floats. I didnt see this when I was using the non-hedge versions of the EA. What kind of results have you guys been getting so far this week with the other versions?
  • #511
  • Aug 30, 2006 3:02pm Aug 30, 2006 3:02pm
Quoting A1Cruiser
Yardie,

The same is here, concerning one of my tests on the 15 minutes time frame the negative float is greater than the profit having the test running for 1.5 weeks.

I think the reason for this is that when hedging Firebird 1.0b versions open two market orders (one buy and one sell) at the same time. So one of the buy or sell orders will always go against us meanwhile the other goes to profit. But one of the two orders will remain there for long and increase the float, particularly in a trending market. According to the best case scenario first the first one then the other takes profit, but in the worst case scenario the first one then the other will stop out - depending on the market movement.

Other strategies, such as news trading straddle EAs, open pending orders both sides of trade (buy and sell). Once the first pending order has been activated the other side is liquidated and the positive is used to go to profit direction. This way only one order will be activated and in a trending market it probably goes to the profitable direction and the float can be retain on a reasonable level.
This is a good point. Maybe we need to make a variation of this version and compare the two. The new version would close out or liquidate the other order once one side was activated. I think that this would help the float also. Any other thoughts on this? Is there anyone willing to program this change?
  • #512
  • Aug 30, 2006 3:27pm Aug 30, 2006 3:27pm
Alpari provides "bad" rates. They use "cheap" traffic. Some experts do not work with it at all.

Quoting sm31
Well, according to my back testing Firebird may not die, but it certainly may look like a lame duck for extended periods. Took me awhile to get around to it, but I've finally managed to test 3 of better performing pairs lately... EUR/USD, USD/CAD, and USD/JPY back to 12/1/2004. Not that back testing is everything, but I think the larger the time frame (due to more data points), the better idea how a system is likely to behave... at least as long as my test data is good, which I have no idea. (got it from http://www.alpari-idc.com/en/dc/databank.php )

I did my testing with the current popular settings (T/P = 30, S/L = 300 on the 1 minute chart, etc.). The results show a slow bleed from 12/04 - 6/05. It may be longer but my data doesn't go back any further.

I'm not sure what changed about the market conditions in June '05, but it must have been significant because results are definitely different from June 05' - present. And the difference seems to affect all three pairs similarly (probably because they're USD based). I'm thinking that whatever conditions contributed to Firebirds poor performance during that time are likely to come around again. Maybe soon.

Right now, my opinion is that this is a good EA in that it doesn't seem to lose money when left alone (many do), but over the long haul it isn't going to take anyone from "rags to riches". I do think that returns can be improved with smart manual intervention, but those improvements will be incremental not exponential. I hope this doesn't pour cold water on anyone's party... And it is my hope that somebody can provide different/better back test results that will refute my tests.

Edit: Got a little more data back to 6/04. Will add this to my testing...
Edit # 2: After some more testing, I have a pretty good suspicion the test data from Alpari is FUBAR'd prior to 10/05, so forget I said anything. ;-)
  • #513
  • Aug 30, 2006 3:41pm Aug 30, 2006 3:41pm
ART,

SO far the B1 looks the best in terms of how much your balance would be if all open positions were closed. The return since August 23rd would be over 10%.


Quoting ART_1_
hi

mid week update of my 5 version which i am testing.

up to now is my favorite the v63g-TS and the 1a, but the 1a has at the moment a lot of float but ~the half of it is only due to GBP/JPY. i have removed today this pair from all 5 versions.

i also have noticed that the v1.0c had today 2 EUR/JPY losses. is anybody else testin v1.0c who can confirm those loosers today?
  • #514
  • Aug 30, 2006 7:09pm Aug 30, 2006 7:09pm
You are using on the 1 minute time frame though correct?



Quoting wackena
No, for me with FBv63g-TS and all v.10 versions, these are the changes from the set above.

MA_timeframe=30
Percent=1.0
TakeProfit=30
Stoploss=300
Use_V63D_Divergence=1

Wackena
  • #515
  • Aug 30, 2006 7:14pm Aug 30, 2006 7:14pm
There are some debate on whether FB can handle breaking news, if you look at the nature of FB is to take small profit during normal spike along the trend or ranging market. But breaking news may wipe out the gain profit in a short time of trading. Take for an example of setting 30-300-30 method with one currency.

If there are 5 trades where 2 are closed trade with 30 pips profit, 1 trade with 15 pips in progressiive profit and the other 2 trades in 10 pips loss as float. When the breaking news come out, either up or down heavily with surprise to the market. That is the market may swing more than 300 pips.

If the existing trade is in favor to the news, then the 3 trades will conclude with another 90 pips in total which give a total of 150 pips for 5 trades. But for the other way round, the existing 3 trades will give 900 pips loss and the nett loss is 840 pips.

Hence my point is it may be good to close all the trades before the breaking news.
  • #516
  • Aug 30, 2006 7:24pm Aug 30, 2006 7:24pm
As I mentioned in my earlier post, FB is performing bound scalping for the small profit during the normal spike. It doesn't matter for the market in ranging but not for trend market.

In trend market, the price can start from small pips until the top trend of say 400 pips. Take up trend as an example, if FB trade short when there is a spike for take profit environment in the initial trend. Then that trade will not close when the trend is heading upward until SL is hit. This is why some member do encounter a SL of 300 pips 20 days later.

Hence my suggestion is 1:3 ratio or not more than 1:5 ratio should be used instead of 1:10 ratio.
  • #517
  • Aug 30, 2006 7:48pm Aug 30, 2006 7:48pm
Hi Wizard, thx for the insights. I think in a live account i might go ahead and at least turn off the ea during the actual announcement, and maybe even close out all trades if it is a big day such as NFP. This Friday? Well i'll probably let it just go since it is only a demo and i'm trying to get the feel of it all.

Art1 hope you figured out those losses. I'm not trading that pair, but so far my 1c is the heads-above winner. The only losses i've incurred have been with the 63G-TS in my testing (but to soon to tell anything of real meaning). . .after NY close:

EA demo as of this post at 6PM Pacific Wed IBFX demo all at default settings, opened this last Sunday:

v1-0b2 closed +859 float -797 net = +62
v1-0c1 closed +1165 float -439 net = +726
63G-TS closed +636 float -707 net = -71

======On another note/thought. What if a trend factor of some sort was thrown in the loop? Like take X-kind-of-trade (long or short) if Xtrend (4 hr, daily, weekly??) was X? For instance we know from typical TA that a spike off support in an uptrend will usualy be indicative of a positive-pip-trade? And vise-versa w/downtrend bounce (spike) off resistance. This might be as simple as looking at the relationship of two MAs such as the 13 & 55 (insert your favorite pair/cross here) cross on XtimeFrame? If SlowMA is above FastMA then take spikes off support in direction of SMA/FMA trend?

Cheers,
Thom
  • #518
  • Aug 30, 2006 9:17pm Aug 30, 2006 9:17pm
Yes accrete, 1c has it heads up..Its a pip machine. Time to go live after the NFP..........................



Quoting accrete
Hi Wizard, thx for the insights. I think in a live account i might go ahead and at least turn off the ea during the actual announcement, and maybe even close out all trades if it is a big day such as NFP. This Friday? Well i'll probably let it just go since it is only a demo and i'm trying to get the feel of it all.

Art1 hope you figured out those losses. I'm not trading that pair, but so far my 1c is the heads-above winner. The only losses i've incurred have been with the 63G-TS in my testing (but to soon to tell anything of real meaning). . .after NY close:

EA demo as of this post at 6PM Pacific Wed IBFX demo all at default settings, opened this last Sunday:

v1-0b2 closed +859 float -797 net = +62
v1-0c1 closed +1165 float -439 net = +726
63G-TS closed +636 float -707 net = -71

======On another note/thought. What if a trend factor of some sort was thrown in the loop? Like take X-kind-of-trade (long or short) if Xtrend (4 hr, daily, weekly??) was X? For instance we know from typical TA that a spike off support in an uptrend will usualy be indicative of a positive-pip-trade? And vise-versa w/downtrend bounce (spike) off resistance. This might be as simple as looking at the relationship of two MAs such as the 13 & 55 (insert your favorite pair/cross here) cross on XtimeFrame? If SlowMA is above FastMA then take spikes off support in direction of SMA/FMA trend?

Cheers,
Thom
  • #519
  • Aug 30, 2006 9:48pm Aug 30, 2006 9:48pm
Quoting Yardie
I am beginning to think version b Hedge has flaws. I am booking a lot of pips, but a lot of it is being eaten up by floats. I didnt see this when I was using the non-hedge versions of the EA. What kind of results have you guys been getting so far this week with the other versions?
Yardie,
I am running one live and one demo with Version b.2. I had no winning trades on the demo today while the float crept to 90% of the profit because of additional positions put on. The live account had one trade in profit but also increased float considerably. Perhaps this is just the market today. I have two trades that are six days old, the rest just a few days and newer.
Any thoughts on my comment presented in my earlier post and the one prior to it in regards to closing out a pending order once order in opposite direction is activated, similar to straddle EA's?

I'm also continuing to have problems with "Excessive Loss" Alerts even when line is commented out (I get warnings when I comment it out in V 1b also). Will this alert prevent orders from being placed. I get this alert in FB 63HO2, as well as 1b. Any thoughts on getting rid of this annoyance?

Thanks Yardie. I appreciate your hard work on this EA.
  • #520
  • Aug 30, 2006 9:58pm Aug 30, 2006 9:58pm
I am not having the excessive loss error with Firebird at the moment so I cannot comment much on that. Maybe Wackena or one of the other coders can provide some more assistance.

So far this week I have made over 1200 pips, yet I am in deficit for the week. I suspect that 1.0b doesnt like tight ranging markets. It likes when the market moves up and down. The past day or so the range in the eurusd has been a mere 50 pips..the same with the usdcad. I think things should be shaken up tomorrow with news and Friday with NFP.

Quoting 1trader2
Yardie,
I am running one live and one demo with Version b.2. I had no winning trades on the demo today while the float crept to 90% of the profit because of additional positions put on. The live account had one trade in profit but also increased float considerably. Perhaps this is just the market today. I have two trades that are six days old, the rest just a few days and newer.
Any thoughts on my comment presented in my earlier post and the one prior to it in regards to closing out a pending order once order in opposite direction is activated, similar to straddle EA's?

I'm also continuing to have problems with "Excessive Loss" Alerts even when line is commented out (I get warnings when I comment it out in V 1b also). Will this alert prevent orders from being placed. I get this alert in FB 63HO2, as well as 1b. Any thoughts on getting rid of this annoyance?

Thanks Yardie. I appreciate your hard work on this EA.
0 traders viewing now