Art Phantom EA | Page 10 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Phantom 6 System 1,540 replies

Phantom 6 - Manual System 8 replies

Phantom Entry on ODL Japan. 2 replies

Tradeview CAD/JPY "Phantom" spike 0 replies

Phantom high, phantom low 0 replies

Attachments: Art Phantom EA
Exit Attachments

Art Phantom EA

  • #181
  • Nov 23, 2010 7:54pm Nov 23, 2010 7:54pm
Quoting zznbrm
Caillou,

I wrote a custom script to backtest.

The attached file contains summary statistics of the 3 month test. Detailed stats are available but I don't feel like accumulating all that data.

By the way, before I made several tweaks to the strategy, I compared the results of my backtest with the demo results you had posted on that external website - the results were surprisingly quite consistent.
if the results are true and accurate, then i have a solution for this huge DD issue.

lets say in a month you get DD above 10% 3 times. or take 5% DD 4-5 times. dont take the trades as per the original EA. let EA run and get into 5% DD, now buy/sell this basket so that you are ahead of others by 5%. set the PP at 5% (anyway basket will recover the DD). with this approach, you can tame the DD issue. you cant blow out your account, because that would have required 5 times the account to blow out by the original EA

this 5% can be variable. you will get less opportunities with 5% DD as most baskets close below that number. so you are missing those profits. but when you get that 5% DD basket your PP is 5%, so you recover some. overall this may not give same PL as original EA, but surely wont blow out yr account.

to test this concept, have 2 accounts. on account1 start EA. if and when the DD reaches 5%, start EA on account2 with PP 5% and watch this account2.
  • #182
  • Nov 23, 2010 8:04pm Nov 23, 2010 8:04pm
Quoting zznbrm
I made some tweaks to the full strategy and ran through my backtester. I used a $10,000 account balance, traded 2.0 mini lots per trade, and a take profit of $50 per basket. Since I am in the US, I used 50:1 leverage. There are a few other specifications about the strategy that are different from the full EA, but that is not really important at this time.

I started two baskets per day: at 2 AM ET and 8 AM ET. A conservative spread was used, commissions calculated. However swap was not included. The goal of the test was to see if and when we would...
Hi zznbrm,

Good work! Please continue with your testing and provide any and all testing analysis.

Some analysis of your results.

1/ Do your results reflect the combined effect on the account of both EA's combined DD's - since you were running two baskets per day?

2/ Can you run one copy of the EA and have it run 24/5 through the sample period you were testing for - so we can see it enter after the close out of the last equity pop. These usually occur during times of high volatility and so further equity pops can occur in quick succession. The way you have them opening at a stated time each day means the EA is unable to take advantage of further trade flows.

3/ Your risk is MASSIVE, you are currently trading 50 times my position size for an account that is only 8 times bigger than my little live account ($1200). For five days in the market I have taken out 9 equity pops or approx 2.25% profit in a week of live trading. Since my risk is much lower than yours I have only had -0.75% max DD to date. (Sure it's only 5 days...)

4/ I can live with 8- 10% per month profit (compound this for over 100% p.a.). I appreciate that I would not have a risk 1/50th of your risk - but it will be WAY smaller than what you were seeing. So a risk of, lets say 20% of your worst DD ~50%, so I might see a ~10% DD (just a guesstimate), based on your results. (Note my profit result is disproportionate to yours since my EA is running 24/5 and not just twice per day. So, like yesterday, I get multiple equity pops (four for the day). At 0.25% each.

5/ Assuming realistic risk management is applied then the primary issue as Frazerd/Caillou mentioned is that the trade sequences that run for many days will cause us to miss out on profit. So it would be sensible - if you are trading with good risk management (and therefore conservative DD), to open a second EA on the same pairs, that will likely keep spitting out the equity pops (simply because it will be started at a different time and day from the first) - maybe in a sub account. This could be initiated if we are stuck in a sequence for more than ~3 days. Once the original EA spits out a profit - go back to just one EA running...???

I would be keen for you to test the EA over the same period with more concervative risk and if possible maintain it running on the basis of 24/5 as we are. And, if possible, add a second EA to trade whenever the original goes for more than ~3 or 4 days.

Please do not take my comments as criticism. I am really keen to see more of your analysis. Especially why/how the EA ended up being stuck in a sequence for so many days. What was the specific dates and what was the market doing that created this situation....

Thanks
Tim
  • #183
  • Nov 23, 2010 8:11pm Nov 23, 2010 8:11pm
Version = Art_Phantom_Extreme
MMTU = 1.0
PPCT = 0.5
Max Float Down = -6.7%
Total Trades = 25
Duration = almost 24 hours

Action taken: put in MPTM's Basket SL to close the whole basket if it breaches -7%. If this works, I will set the SL to a lower value
  • #184
  • Nov 23, 2010 8:28pm Nov 23, 2010 8:28pm
I guess i'll share my results to see if anyone can benefit from it:

Before market open this week, I setup 6 demos, same broker
1. extreme, default pairs
2. full default pairs
3. light default pairs
4. extreme, gj, eg, ej
5. full, gl, eg, ej
6. light gj, eg, ej

All started with 1000, 300:1 leverage, .4 mmtu, 1 pp

Stats as of right now:
1. 1 closure, current balance 1011 (up 1.1%), current odd -$60 (6%), max dd 9.3%)
2. 4 closures, current balance 1038 (up 3.8%), current odd -$70 (7%), max dd 8%)
3. 0 closures, current balance 1000 (up 0%), current odd -$67 (6.7%), max dd 8.8%)
4. 0 closures, current balance 1000 (up 0%), current odd -$504 (50.4%), max dd 63%)
5. 1 closure, current balance 1010 (up 1%), current odd -$180 (18%), max dd 22.6%)
6. 1 closure, current balance 1010 (up 1%), current odd -$61 (6.1%), max dd 8%)
  • #185
  • Nov 23, 2010 8:32pm Nov 23, 2010 8:32pm
but has the basket closed in profot (considering the losses taken from losers) ?

Hi Roundrock,

Yes, it has closed multiple times. Please keep in mind that when you close a loser (or winner for that matter) your equity does not change, only your balance. We dont care about balance only that equity reaches our target.

I dont want to hijack the thread with a phantom 6 ea so this will be the last post over here. This is the EA Im running. Its the basic Phantom6_01x only change is I use different pairs than tthe original.

Attached File(s)
File Type: mq4 Phantom6_01x_A.mq4   34 KB | 389 downloads
  • #186
  • Nov 23, 2010 8:54pm Nov 23, 2010 8:54pm
[quote=Caillou;4193946]
Quoting pipchaser888

Hi Pipchaser888,

Could you attach a screenshot to see your information ? Your broker has increments 0.01 and what is minimum lot size?

Hi Caillou,

Here is the screen shot. I started the EA with default setting, and balance of 10K. Still no position open till now.

Thanks.
Attached Image (click to enlarge)
Click to Enlarge Name: art_phantom.gif Size: 66 KB
  • #187
  • Nov 23, 2010 8:58pm Nov 23, 2010 8:58pm
Hi Thorn, Roundrock
This may work well as it won't disrupt the trading system for the EA. All you are doing is extending the profit requirement based on the equity target being the same at all stages.

Roundrock can you recode a version of the Art full with this worst trade close out function in it? (Full being the best version so far I think). Also can you make it a variable that uses a multiple of the $ target value. So 2 x $target would mean any trade that exceeds this $ value would close - and it has to be the worst looser also.

I think it is worth testing further.

Tim
  • #188
  • Nov 23, 2010 8:59pm Nov 23, 2010 8:59pm
Quoting thorn
but has the basket closed in profot (considering the losses taken from losers) ?

Hi Roundrock,

Yes, it has closed multiple times. Please keep in mind that when you close a loser (or winner for that matter) your equity does not change, only your balance. We dont care about balance only that equity reaches our target.

I dont want to hijack the thread with a phantom 6 ea so this will be the last post over here. This is the EA Im running. Its the basic Phantom6_01x only change is I use different pairs than tthe original.

Attachment ...
here you go.... i added a external parameter CutLoserMultiplier, you can change this from default 2 to whatever number. this is a multipler of Dollar Profit. losing trade should have a loss greater than this to cut short...
Attached File(s)
File Type: mq4 Phantom6_01x_A.mq4   36 KB | 201 downloads
  • #189
  • Nov 23, 2010 9:26pm Nov 23, 2010 9:26pm
Thanks Roundrock,

I will set up a couple of demos and run them at same time. One with the trim and one without. It will probably take a few weeks of trading to see if it makes a big difference.
  • #190
  • Nov 23, 2010 9:36pm Nov 23, 2010 9:36pm
Timmy,

You can copy the code off that phantom6_01x_A EA that roundrock coded up for me. I think it does what you are looking for.
  • #191
  • Nov 23, 2010 9:37pm Nov 23, 2010 9:37pm
Quoting the-game
Mate, don't get defensive, I obviously had no idea you can backtest a multipair strategy like you have done.
Well, I guess it is better to be defensive than offensive.
  • #192
  • Nov 23, 2010 9:43pm Nov 23, 2010 9:43pm
Quoting timmy
Hi Thorn, Roundrock
This may work well as it won't disrupt the trading system for the EA. All you are doing is extending the profit requirement based on the equity target being the same at all stages.

Roundrock can you recode a version of the Art full with this worst trade close out function in it? (Full being the best version so far I think). Also can you make it a variable that uses a multiple of the $ target value. So 2 x $target would mean any trade that exceeds this $ value would close - and it has to be the worst looser also.

I think it...
here you go ...
Attached File(s)
File Type: mq4 Art_Phantom_Full_A.mq4   50 KB | 254 downloads
  • #193
  • Nov 23, 2010 10:08pm Nov 23, 2010 10:08pm
Quoting charvo
ZZ's programming is way beyond the pathetic strategy tester in MT4. however, multipair backtesting is for sure not impossible, and it is realised in other threads if you read into them (like t101's method). I personally believe that ZZ's backtesting is trustworthy.
Thanks charvo!

Quoting timmy
1/ Do your results reflect the combined effect on the account of both EA's combined DD's - since you were running two baskets per day?
Each basket was a separate test. Each basket started with $10,000.

Quoting timmy
2/ Can you run one copy of the EA and have it run 24/5 through the sample period you were testing for - so we can see it enter after the close out of the last equity pop. These usually occur during times of high volatility and so further equity pops can occur in quick succession. The way you have them opening at a stated time each day means the EA is unable to take advantage of further trade flows.
I did not run the EA for my tests. I created a customized script to perform the tests.

But yes, the idea is to eventually run my test script in the same way the EA runs. I just have not done so yet. Again, the results I posted were for a particular test case - to determine the kinds of drawdowns to be expected over a 3-month period for particular set of inputs.

As I stated, I really just wanted to point out that the last few days were pretty good for this type of strategy. For the one set of tests I posted, you can see that Nov was very good, but Oct and Sept were not as promising.

Quoting timmy
3/ Your risk is MASSIVE, you are currently trading 50 times my position size for an account that is only 8 times bigger than my little live account ($1200). For five days in the market I have taken out 9 equity pops or approx 2.25% profit in a week of live trading. Since my risk is much lower than yours I have only had -0.75% max DD to date. (Sure it's only 5 days...)
I used the same risk Caillou was using for his tests. He posted results on an external site - looked like a $10,000 account trading 0.2 lots. I assumed it was a standard account. So, for this test I though it would make sense for me to use something comparable - 2.0 lots on a $10,000 account (MBTrading account...similar to a mini account). I have messed around with many variations (lot size, tp size, minimum margin levels, leverage) - you only saw the output of one of those variations.

I hope your live trading continues in a positive direction.

Quoting timmy
4/ I can live with 8- 10% per month profit (compound this for over 100% p.a.). I appreciate that I would not have a risk 1/50th of your risk - but it will be WAY smaller than what you were seeing. So a risk of, lets say 20% of your worst DD ~50%, so I might see a ~10% DD (just a guesstimate), based on your results. (Note my profit result is disproportionate to yours since my EA is running 24/5 and not just twice per day. So, like yesterday, I get multiple equity pops (four for the day). At 0.25% each.
I'm not thinking about profit per month or anything like that (not yet anyhow). I'm trying to determine the effectiveness of the strategy during all (most) market conditions.

Quoting timmy
5/ Assuming realistic risk management is applied then the primary issue as Frazerd/Caillou mentioned is that the trade sequences that run for many days will cause us to miss out on profit. So it would be sensible - if you are trading with good risk management (and therefore conservative DD), to open a second EA on the same pairs, that will likely keep spitting out the equity pops (simply because it will be started at a different time and day from the first) - maybe in a sub account. This could be initiated if we are stuck in a sequence for more...
That is all something to think about as the strategy progresses.

Quoting timmy
I would be keen for you to test the EA over the same period with more concervative risk and if possible maintain it running on the basis of 24/5 as we are. And, if possible, add a second EA to trade whenever the original goes for more than ~3 or 4 days.

Please do not take my comments as criticism. I am really keen to see more of your analysis. Especially why/how the EA ended up being stuck in a sequence for so many days. What was the specific dates and what was the market doing that created this situation....
I appreciate your observations. Yes, it is obvious to me that the strategy will have many successful trades regardless of risk/tp levels. However, as is the case with all strategies, I am more interested in the bad days and how I can minimize the impact of these days on the profitability of the strategy.
  • #194
  • Nov 23, 2010 10:34pm Nov 23, 2010 10:34pm
wow, you quick, RR.

i just finished my mod per frazerd's suggestion:

when margin level drops to an alarming level, the EA will close the biggest losing position, until the margin level rise over the alarming level again.

though it might do same thing as yours, i think that i just upload incase it is useful to anyone.

mine is based on art_phantom_light.mq4, and my coding may be easily moved to the other versions. if you open the code, search for "charvo added", those are the lines that I added on Caillou's original coding.

there are two new external var:
1. LowMarginKillLoser, this is the switch, if true, the "kill biggest loser" is in effect.
if false, the EA is exactly same as Caillou's original.
2. AlarmingMarginPer, this is the % of margin level that will trigger the "kill biggest loser" function, of coz, when 'LowMarginKillLoser" is set true.
default is 500, means 500% is the lowest % margin the EA will stand.

Also, on screen comments, under margin info, there is one more line stating the alarming margin level.

there is very slight chance for a bug, if there is, let me know, please.

Quoting roundrock
here you go ...
Attached File(s)
File Type: ex4 Art_Phantom_Light_charvo_kill_loser.ex4   37 KB | 326 downloads
File Type: mq4 Art_Phantom_Light_charvo_kill_loser.mq4   51 KB | 314 downloads
  • #195
  • Nov 23, 2010 10:52pm Nov 23, 2010 10:52pm
Quoting charvo
wow, you quick, RR.

i just finished my mod per frazerd's suggestion:

when margin level drops to an alarming level, the EA will close the biggest losing position, until the margin level rise over the alarming level again.

though it might do same thing as yours, i think that i just upload incase it is useful to anyone.

mine is based on art_phantom_light.mq4, and my coding may be easily moved to the other versions....
Hi Charvo,

Better to code your ExitBiggestLoser() function in the following way, since MC is already tracking the Biggest Loser Real Trade.

Inserted Code
void ExitBiggestLoser()
{
   if (BiggestLoserTicket > 0)
   { 
     OrderSelect(BiggestLoserTicket , SELECT_BY_TICKET, MODE_TRADES);
     if ( OrderType() == OP_BUY )  OrderClose( OrderTicket(), OrderLots(), MarketInfo(OrderSymbol(), MODE_BID), 5, Red ); 
     if ( OrderType() == OP_SELL ) OrderClose( OrderTicket(), OrderLots(), MarketInfo(OrderSymbol(), MODE_ASK), 5, Red ); 
   }
}
  • #196
  • Nov 23, 2010 10:57pm Nov 23, 2010 10:57pm
Quoting charvo
wow, you quick, RR.

i just finished my mod per frazerd's suggestion:

when margin level drops to an alarming level, the EA will close the biggest losing position, until the margin level rise over the alarming level again.

though it might do same thing as yours, i think that i just upload incase it is useful to anyone.

mine is based on art_phantom_light.mq4, and my coding may be easily moved to the other versions. if you open the code, search for "charvo added", those are the lines that I added on Caillou's original coding.

there are two...
Charvo,

If the margin is less than 500 %, it cuts down biggest looser, ok. In that case, the overall basket will be loser considering the fact that it had already accepted loosing trade?

Thanks.
Intraday only.
  • #197
  • Nov 23, 2010 11:06pm Nov 23, 2010 11:06pm
dang, thanks a lot, man. i didn't pay my attention there.

Quoting the-game
Hi Charvo,

Better to code your ExitBiggestLoser() function in the following way, since MC is already tracking the Biggest Loser Real Trade.
  • #198
  • Nov 23, 2010 11:12pm Nov 23, 2010 11:12pm
the fact is the EA will start accept losing trades, and stoploss happened.

however, the overall basket might recover faster, this is the point frazerd elaborated at the other phantom EA thread.

take an example, you have 3 pos of eurusd, -100 pip, -200 pip, and -300 pips. these -600 pips make your account lower than 500% margin. then EA starts killing the -300 pips pos. after the pos is stopped, the account is only having two pos, -100 and -200. and the EA now has better chance to recover and close back at 'profitpct'.

or in the above example, the -100 and -200 pos may be +10, and +50, so after -300 is stopped, the EA's basket is now in positive side.

Quoting navk
Charvo,

If the margin is less than 500 %, it cuts down biggest looser, ok. In that case, the overall basket will be loser considering the fact that it had already accepted loosing trade?

Thanks.
  • #199
  • Nov 23, 2010 11:17pm Nov 23, 2010 11:17pm
Quoting charvo
dang, thanks a lot, man. i didn't pay my attention there.
That's okay This way it will guarantee that you close "THE" Biggest Loser.
  • #200
  • Nov 23, 2010 11:41pm Nov 23, 2010 11:41pm
Quoting charvo
the fact is the EA will start accept losing trades, and stoploss happened.

however, the overall basket might recover faster, this is the point frazerd elaborated at the other phantom EA thread.

take an example, you have 3 pos of eurusd, -100 pip, -200 pip, and -300 pips. these -600 pips make your account lower than 500% margin. then EA starts killing the -300 pips pos. after the pos is stopped, the account is only having two pos, -100 and -200. and the EA now has better chance to recover and close back at 'profitpct'.

or in the above example,...
So the overall is -300 +10+50 ? This my question.
Intraday only.
5 traders viewing now