Four basket closures in five days this week. See my settings in previous posts. You guys who are blowing accounts are trading too big of lots. I'm trading with 50:1 leverage and haven't blown up since I started this EA in early november. Of course my next basket could blow.
Its a shame that nobody is taking up running with this EA. It has a crystal clear concept that simply needs tweeking to address the DD problem.
A coder is urgently needed to (a) make 2 synthetically hedged sub-baskets out of the six pairs to get to PP quicker and (b) code a recovery system that comes into play when ONE particular pair contained in a sub-basket goes sufficiently against position so that it needs to be neutralized, ie, simply a 1:1.1 hedge and let the rest of the pairs within the sub-baskets continue their work to get us back to PP (or somewhere close to our last flat balance where we can manually intervene accept a small loss and live to fight another day)
Is there not a coder out there who can take on the challenge?
Honestly I'm not seeing much need to "improve" on this system. Perhaps so code cleanup and efficiency, but otherwise the system seems pretty solid, and my own tests so far have been positive.
I'm looking at the posts of people who are having problems with it and I have to wonder if greed and lack of patience is the real reason the system is failing for them.
Try running AddSix_02a with default settings and LOTS set to 0.01.
Hi, Tooth, you used 0.4 for MMTU, right? so it is 0.004% of $40000, that is $160 for each position. and that's approx. 0.06 lot (micro lot brokers).
en, that's pretty conservative. good to know for me.
Ignored
Yes, .4 MMTU which gave me .08 standard lots at around $40000 and automatically bumped the lot size to .09 when balance reached around $44000. I'm using only the AddSix 02a EA. I even ran it hands off clear through the holiday chop. With these setting even at 50:1 leverage, I can withstand well over 100 trades in a basket. Have seen that twice. There's always the risk of that death basket though. Nicastro's ideas would require a lot more margin and being able to hedge. US citizens forget it.
Its a shame that nobody is taking up running with this EA. It has a crystal clear concept that simply needs tweeking to address the DD problem.
...
Is there not a coder out there who can take on the challenge?
Ignored
There are plenty of coders up to the challenge. However, programmers like detailed requirements. Supply details and you will probably get some interest.
I am posting this for all to comment upon, but have quoted Frazerds post as being most relevant to what I have to say
Firstly it would be a real pity if this thread and the development of these ea's stopped now that MetaCoder has moved on
I long ago came to the conclusion trying to pick any trade direction based on applying analysis to the past is a futile exercise
[color=black][font=Verdana]The market is driven, pulled...
Ignored
JCS,
Good post.. I have been testing different phantoms since around september. If I may, just a few suggestions for you I have found from my testing. You can take em or leave em its up to you.
MMTU- yours is fine. I use .5.
CLM- depends on leverage. For 50:1 I use 1.2., 100:1 use 2.0.
PP- 0.25- 0.5% (less profit on close yes, but quicker close, less DD)
Pip Step--Im up in air on this one.
"X" factor- I would never use these EA's without it. I see most people having trouble with these phantoms are NOT using it.
Pairs- I know this is not something that is discussed too much in these threads. I dont use the original pairs. I put together a 6 pair basket that is syntheticaly hedged but is also more correlated than the original 6 pairs.
Yes, .4 MMTU which gave me .08 standard lots at around $40000 and automatically bumped the lot size to .09 when balance reached around $44000. I'm using only the AddSix 02a EA. I even ran it hands off clear through the holiday chop. With these setting even at 50:1 leverage, I can withstand well over 100 trades in a basket. Have seen that twice. There's always the risk of that death basket though. Nicastro's ideas would require a lot more margin and being able to hedge. US citizens forget it.
Ignored
I wouldn't say us citizens forget it. You just move your money oversees. A lot of places do micro of micro accounts. that's a whole political discussion in itself.
Good post.. I have been testing different phantoms since around september. If I may, just a few suggestions for you I have found from my testing. You can take em or leave em its up to you.
MMTU- yours is fine. I use .5.
CLM- depends on leverage. For 50:1 I use 1.2., 100:1 use 2.0.
PP- 0.25- 0.5% (less profit on close yes, but quicker close, less DD)
Pip Step--Im up in air on this one.
"X" factor- I would never use these EA's without it. I see most people having trouble with these phantoms are NOT using it.
Four basket closures in five days this week. See my settings in previous posts. You guys who are blowing accounts are trading too big of lots. I'm trading with 50:1 leverage and haven't blown up since I started this EA in early november. Of course my next basket could blow.
Ignored
Hello Toothman,
I have re-read your previous post and made some calculations too. Correct me if I made a mistake but this demo account gained 71 pips/day as average. It means more than 25,000 pips/year: is it right?
Thorn,
Can I ask what pairs you're trading? Why would you want more correlation? Are you talking about negative correlation? thanks
Ignored
Toothman,
I trade the following 6 pairs (Phantom 6_01x)
Longs: GBP/USD EUR/CAD EUR/GBP
shorts: EUR/USD EUR/AUD AUD/CAD
What I meant about correlation is the are + correlated, (ex: if EUR/Cad is moving up EUR/AUD probably is too) so if I have a EUR/AUD short and price starts to go up then I will start to get longs in EUR/Cad to balance the basket. Of course we now its not perfect and correlations can and do change at times. I didnt like the origanal pairs as there was some negative correlations which I felt just didnt work as well. I used mataf.com and looked mostly at the daily correlation chart to make my decision on pairs. EU/GU and EC/EA are very possitive, EG/AC not as much but still possitive and enough to use to make the synthetic hedge.
My longest running phantom using these pairs is up 13.5% since 11/24/10.
Just about 2 months now. max DD is at 5.65%.
I also just started testing a 4 pair phantom using the same settings as the 6 pair. Using highy correlated pairs eur/usd eur/cad aud/usd aud/cad. So far in 2 weeks its doing well (+ 4.5%) with max DD at 0.66%.
I trade the following 6 pairs (Phantom 6_01x)
Longs: GBP/USD EUR/CAD EUR/GBP
shorts: EUR/USD EUR/AUD AUD/CAD
What I meant about correlation is the are + correlated, (ex: if EUR/Cad is moving up EUR/AUD probably is too) so if I have a EUR/AUD short and price starts to go up then I will start to get longs in EUR/Cad to balance the basket. Of course we now its not perfect and correlations can and do change at times. I didnt like the origanal pairs as there was some negative correlations which I felt just didnt work as well. I used mataf.com...
I trade the following 6 pairs (Phantom 6_01x)
Longs: GBP/USD EUR/CAD EUR/GBP
shorts: EUR/USD EUR/AUD AUD/CAD
What I meant about correlation is the are + correlated, (ex: if EUR/Cad is moving up EUR/AUD probably is too) so if I have a EUR/AUD short and price starts to go up then I will start to get longs in EUR/Cad to balance the basket. Of course we now its not perfect and correlations can and do change at times. I didnt like the origanal pairs as there was some negative correlations which I felt just didnt work as well. I used mataf.com...
Ignored
oanda is offering MT4 on practice accounts, i will try this....
oanda is offering MT4 on practice accounts, i will try this....
Ignored
Yes, that is awsome. I am going to set one up monday. Are you able to set up multiple mt4 sub accounts? I know on fxtrade you could have multiple demos running at once. That would be ideal for testing multiple ea's on one broker.
Yes, that is awsome. I am going to set one up monday. Are you able to set up multiple mt4 sub accounts? I know on fxtrade you could have multiple demos running at once. That would be ideal for testing multiple ea's on one broker.
Ignored
i have created 2 sub-accounts and installed MT4 in 2 locations (same computer). i am planning to test AddSix_03bx_CL_TH_PS. and Phantom6_01x with clm and your pairs.
i need to test the lotsize, they offer minimum 1000 units. i created sub-accounts with 3k account, this is good for mmtu 0.4, pp 0.5, 50:1 leverage. i will trade 1000 units.
I have re-read your previous post and made some calculations too. Correct me if I made a mistake but this demo account gained 71 pips/day as average. It means more than 25,000 pips/year: is it right?
May the Gain be with you.
/achab
Ignored
I'm not counting pips which would be pretty difficult to figure out. The account has grown about 15% over the last two months. But that's over $6000 with a 40,000 account.
i have created 2 sub-accounts and installed MT4 in 2 locations (same computer). i am planning to test AddSix_03bx_CL_TH_PS. and Phantom6_01x with clm and your pairs.
i need to test the lotsize, they offer minimum 1000 units. i created sub-accounts with 3k account, this is good for mmtu 0.4, pp 0.5, 50:1 leverage. i will trade 1000 units.
Ignored
Current phantom Im testing with since 11/24/10. Feel free to change the settings. This one does not use pip step or trading hours. Since most brokers have high spread on aud/cad, eur/cad, eur/aud, Oanda should work out well with the lower spreads.
Thorn,
I understand the correlation aspect, but the ea itself picks the biggest winner of the three pairs on each side of the hedge. In a sense your picking three correlleated pairs (on the buy and sell side) to ensure a good basket ; but yet the pairs also have to be somewhat positive correlated within the group.? is that accurate.
Current phantom Im testing with since 11/24/10. Feel free to change the settings. This one does not use pip step or trading hours. Since most brokers have high spread on aud/cad, eur/cad, eur/aud, Oanda should work out well with the lower spreads.
Something I noticed with the Phantom series of EA's.
They are designed to open the same number of lots in each pair, e.g. 0.25, regardless of the value of each trade.
I thought the idea was that if you had exactly one trade open in each pair then you should have a full synthetic hedge, i.e. nothing going anywhere regardless of market movement.
But with one lot of each pair, you won't get that; you need to adjust.
Hypothetical example.
Suppose you had one lot of EURUSD and one lot of the (non-existent) USDEUR. Do those balance? No, because one lot of the first is 100,000 euros which is 135,000 dollars, and one lot of the second is 100,000 dollars.
This extends to the three and up pair case.
I noticed that in running the EA's also if you consider phantoms in $ terms, then you get more GBPxxx trades simply because one lot of GBPxxx is "more money" than one lot of say USDxxx. I thought that might not be intended either.
So my new EA (attached) recomputes the lotsize, adjusting it to be normalized to USD, e..g for a GBPxxx trade it divides the nominal lotsize by the value of GBPUSD, and similarly for EURxxx and CHFxxx trades.
When it gets a fractional lot size, e.g. 0.2541 but you can only trade to two digits, it uses a stochastic approximation. E.g. if it wants 0.2541, it will actually use 0.25 (100-41)% of the time, and 0.26 41% of the time, using a random number generator.
Also I noticed that at some point the ability to restart the EA without resetting the phantoms (using truly global variables) was lost. I reimplemented it again. The EA also reports in the comment block the time that the phantoms were last reset. If you want to reset upon restarting the EA, then hit F3 and delete the global variable (it should be obvious from the name).
Because of the balancing this will obviously have different results than the traditional EA's, who knows if it is better or worse over time practically (maybe being GBP volatility trader is good!), but it seems that it is at least cleaner theoretically.
Because the typical lotsizes with balancing are a little smaller, I increased the default MMTU from 0.5 to 0.65.
please try it and report results.
Also it computes phantom PL in terms of basis points, i.e. a relative percent measure instead of $ terms.
