Try My EA! Need Group Feedback... | Page 16 | Metals Mine

Options

Go to
Permalinks
Print Thread

Related Instruments

Gold/USD

Similar Threads

USD/TRY and EUR/TRY 12 replies

Fibo Group Feedback 23 replies

Need feedback on my trading breakouts system 8 replies

One World Capital... Need feedback 2 replies

Need feedback on FXCM Educational workshops 3 replies

Attachments: Try My EA! Need Group Feedback...
Exit Attachments

Try My EA! Need Group Feedback...

  • #301
  • Dec 5, 2014 8:10am Dec 5, 2014 8:10am
http://www.kitco.com/images/live/nyg...07787682455684
Ok, My trade exporer is up.
NFP was ok, we're in a retrace, but I think price will make a continuation going down later today.
My EA is setup with micro demo on Finfx, VPS is server in frankfurt, Germany. The Latency pinged at 29 ms...pretty good!
I'm not sure if it's the fault of the VPS server being interally slow, or what, but the EA took too much time to close orders once in profit...example trade:

Sell 1,192.57
Buy 1,193.66...this on close of trades, lost $1.09.

This is ok, the EA is about long term, dig in the trenches and wait it out.

I see some of you who were down are now going up. Patience is key.

Also, the latency to the broker seems to be less important than the speed MT4 is running at on a server\computer. Shiva has a kinda high ping, but I'm guessing the time it takes for all his trades to close is small, which is good. Hang in there everybody!
Too close for pips, switching to points.
  • #302
  • Dec 5, 2014 8:10am Dec 5, 2014 8:10am
I have a problem. The EA hit the equity gain it was aiming for (I think). It then closed all the pending orders and some of the open orders but kept the rest open. Why would that happen?
  • #303
  • Dec 5, 2014 8:11am Dec 5, 2014 8:11am
Hi all,

@Thepipster: You've got me interested; thanks for sharing the EA. Here's my TE: Steady Gold, started today.

This is an Armada Demo, set up on a free Amazon VPS, with latency of 160ms. I took the liberty of tweaking some settings: buffer 2.0; equity 2.0; noofsteps 60. Most importantly though, I'll only be testing it during the London session. So far, it's done well on the NFP spike down (47% gain since LO today), but let's not be greedy, it's turned off now until next week.

It's a fantastic tool. In responsible and experienced hands it can be a real gold mine.
mind over matter
  • #304
  • Dec 5, 2014 8:13am Dec 5, 2014 8:13am
So, looking at what the EA is doing during this spike, some observations. My account on default is trading a much larger lot size than others seem to have with the London exchanges (0.4 lots on $8k) . What this does is to cause it to hit the targeted amount much quicker when it moves, and close and reopen a lot more than others do as well. There may be some method to that madness, but I guess it could crater the account as well. It's just that I can't envision a zigzag scenario in which it wouldn't close and reopen in the midst of to collect the gains at the extremes, since the profits come so quickly as it moves.

I'll open another demo in a clean account when I return from vacation. But this does seems to work as intended. I'm a US customer using ATC brokers, and my latency is pretty good at 70ms - mostly the trades close within 5 seconds as they are supposed to. Wish I could use Armada like Shiva - no chance for me there : (
A good man knows his limitations ~ Clint Eastwood
  • #305
  • Dec 5, 2014 8:15am Dec 5, 2014 8:15am
Quoting kashTrader
I have a problem. The EA hit the equity gain it was aiming for (I think). It then closed all the pending orders and some of the open orders but kept the rest open. Why would that happen?

I had that same trouble, I found some orphans left when it cycled on my demo too. There should be some error checking to clean up trades that didn't execute in the code, but I haven't tried to find it.
A good man knows his limitations ~ Clint Eastwood
  • #306
  • Edited 8:44am Dec 5, 2014 8:21am | Edited 8:44am
Quoting kashTrader
I have a problem. The EA hit the equity gain it was aiming for (I think). It then closed all the pending orders and some of the open orders but kept the rest open. Why would that happen?
It shouldn't leave any orphans, that, I think, is a broker problem. The EA is programmed to look for orphan trades and close them before restarting.

~Kosmo...all I can say is wow! that's more return than we should be making in a whole month! Can you please share with us, when the info becomes available to you, how long the EA is taking to close a set of trades? Your equity curve has no spikes in it, except right at the news release. Very impressive. What setting did you change when you say equity, do you mean target, or lotpercent? Can you show the lot size and profit on your TE, please? I'm also using free tier AWS, so I would think you and me have same playing ground on the server side, but you have way more latency, and you're using 60 steps...that's 120 pending orders that have to be closed! Yet, your EA had trades that closed sets every minute for 7 minutes after nfp. You got all of the spike down, and all of the retrace up. The only other difference I see is the broker. I have tried limiting the EA to sessions before, and even only traded news with it...it's ok, but back in the spring there was a lot of russia vs. ukraine news that kept happening in the middle of nowhere that I wanted to take advantage.
Too close for pips, switching to points.
  • #307
  • Dec 5, 2014 8:30am Dec 5, 2014 8:30am
My Armada demo with a £50k bank racked up a profit of £3855 (7.7%) and then didnt open anymore trades. Its still in live mode but no trade action,with one untriggered sell stop order on screen. Is that whats supposed to happen. Reaches a target for the day and stops or is something adrift ?
By the way Pipster , Many Thanks for an amazing piece of work, cant believe we,ll get away with using it live in a few months time, if its still delivering.
  • #308
  • Dec 5, 2014 8:36am Dec 5, 2014 8:36am
Quoting go4
My Armada demo with a £50k bank racked up a profit of £3855 (7.7%) and then didnt open anymore trades. Its still in live mode but no trade action,with one untriggered sell stop order on screen. Is that whats supposed to happen. Reaches a target for the day and stops or is something adrift ? By the way Pipster , Many Thanks for an amazing piece of work, cant believe we,ll get away with using it live in a few months time, if its still delivering.
You'll have to close that last trade. It's called an orphan, and it should have closed with the others, I don't know why some people have that problem...I think it's broker related.
Once you close it, the EA will start over again.
Too close for pips, switching to points.
  • #309
  • Dec 5, 2014 8:55am Dec 5, 2014 8:55am
Quoting Thepipster
{quote} You'll have to close that last trade. It's called an orphan, and it should have closed with the others, I don't know why some people have that problem...I think it's broker related. Once you close it, the EA will start over again.

Thanks for that. Just out of interest I tried to delete the order, modify it ,execute it; non of them would get rid of it . The only solution was to shut it down and restart armada. Now its repopulated and off again.
Thanks
  • #310
  • Dec 5, 2014 8:58am Dec 5, 2014 8:58am
I know why its leaving orders there. I have updated the closeOrder code to ensure that it does close all the orders and waits until all of the are actually closed as opposed to giving up.

However, I think that if broker is causing issues then this may just wait too long to close all the orders. I think I have to look deeper into the a and think about setting TP as well as this closing method. That will hopefully get a more solid closing system. However, setting the TP will require some clever maths and a bit of coding.
  • #311
  • Dec 5, 2014 9:09am Dec 5, 2014 9:09am
Quoting go4
{quote} Thanks for that. Just out of interest I tried to delete the order, modify it ,execute it; non of them would get rid of it . The only solution was to shut it down and restart armada. Now its repopulated and off again. Thanks
Definitely broker problem...I've never seen or heard of that before.

Just because some people are getting ridiculous returns with armada, doesn't mean there the best overall

Quoting kashTrader
I know why its leaving orders there. I have updated the closeOrder code to ensure that it does close all the orders and waits until all of the are actually closed as opposed to giving up. However, I think that if broker is causing issues then this may just wait too long to close all the orders. I think I have to look deeper into the a and think about setting TP as well as this closing method. That will hopefully get a more solid closing system. However, setting the TP will require some clever maths and a bit of coding.
If you think you can figure it out, I'm all for it...please test it for awhile, at least a few days, before releasing it to us.
Too close for pips, switching to points.
  • #312
  • Dec 5, 2014 9:32am Dec 5, 2014 9:32am
Quoting Thepipster
{quote} Definitely broker problem...I've never seen or heard of that before. Just because some people are getting ridiculous returns with armada, doesn't mean there the best overall {quote} If you think you can figure it out, I'm all for it...please test it for awhile, at least a few days, before releasing it to us.

Hi Pips,

I like this EA. I think I am going to spend some time over the break improving some of the logic in the code such as closing of orders and stuff. I won't change the underlying idea as I think this is its strength. I will simply change code where I see it will be helpful with execution of the idea.

I will be in touch with you if i need to understand any of the code in more details.

Although it will take some time, I think the key will be to set the correct TP on each trade so that there is broker side closing with additional monitoring on the EA side to make sure that the broker didn't miss our closes because of a slip in the price. This will also possibly help with the 30 trades you now have. We should be able to increase that to lets say 50. This combined with small profit targets should do the trick.

I will keep you posted.
  • #313
  • Dec 5, 2014 9:36am Dec 5, 2014 9:36am
[quote=Thepipster;7923981]{quote} Just because some people are getting ridiculous returns with armada, doesn't mean there the best overall {quote}


As I said in an earlier post, I have the Goldmine EA running since this morning on 3 Broker demos simultaneously,each with a 50k bank. At this moment Finfx is up £2116, Alpari up £748 and Armada (with a timeout for the orphan of about 2 hrs) up £4144. Amazing ! Long may it continue
  • #314
  • Dec 5, 2014 9:52am Dec 5, 2014 9:52am
Pips,

A quick question. Is there any reason why the EA looks to gain a certain Equity? Is it not the same if we look for a certain P/L?
I ask this because, if you have multiple accounts, or doing something else then the global variable to read the previous equity can screw up the values picked up by the EA.

Is it not sufficient to say we want a $100 P/L instead of saying my desired equity if $5100 when starting with $5000?
  • #315
  • Dec 5, 2014 10:01am Dec 5, 2014 10:01am
Quoting kashTrader
Pips, A quick question. Is there any reason why the EA looks to gain a certain Equity? Is it not the same if we look for a certain P/L? I ask this because, if you have multiple accounts, or doing something else then the global variable to read the previous equity can screw up the values picked up by the EA. Is it not sufficient to say we want a $100 P/L instead of saying my desired equity if $5100 when starting with $5000?
No, Equity Profit and Trades P/L are two different matters. Global variables are bound to the current MT4 instance not all the MT4 running on the same machine.
My Pips Stolen By The Broker
  • #316
  • Dec 5, 2014 10:06am Dec 5, 2014 10:06am
quote=kashTrader;7924148]Pips, A quick question. Is there any reason why the EA looks to gain a certain Equity? Is it not the same if we look for a certain P/L? I ask this because, if you have multiple accounts, or doing something else then the global variable to read the previous equity can screw up the values picked up by the EA. Is it not sufficient to say we want a $100 P/L instead of saying my desired equity if $5100 when starting with $5000?[/quote]

Good Question. It may not matter anymore, but originally I needed to make sure if any trades closed incorrectly, or a TP option I used to have in there was used, that the EA wouldn't change how much profit was needed from original balance. The floating p\l doesn't work, because it could change, without being true to the original account balance.

Example: Lets say I had 30 trades sitting open at various prices\directions, and the balance was $1000 with target to increase it to $1010. Let's say the floating P\L was negative, but I had 10 trades that had a lot of profit themselves and decided to close them, and hope the losers would come back to net the whole account some profit. When I do that, the floating P\l goes far more negative, but my balance increases...even though the true original P\l didn't change at all. The EA would just see the floating P\l, not what is happening in the big picture. Now, if we took that example and instead of closing some profitable trades, closed some negative trades. Now, the floating P\L shifts to in the green-profit, but the total account is the same actual P\L as before, and the balance is negative! If price moved either way at this point, and the EA closed all trades, it would have no idea what the original balance was, or the required overall target. It would either cause a loss or profit at random. I'm worried this might happen after the EA gives close all orders command, but before all orders are actually closed. Otherwise, since we are not intervening and no actual TP, this isn't a real problem to us.

~go4...great! I'm just pointing out, despite the great profit with armada, them leaving a trade open, AND not letting you close it could be a real bummer if you went live. Link up the accounts to trade explorer!
Too close for pips, switching to points.
  • #317
  • Dec 5, 2014 10:07am Dec 5, 2014 10:07am
Quoting Thepipster
Kosmo...all I can say is wow! that's more return than we should be making in a whole month! Can you please share with us, when the info becomes available to you, how long the EA is taking to close a set of trades? Your equity curve has no spikes in it, except right at the news release. Very impressive. What setting did you change when you say equity, do you mean target, or lotpercent? Can you show the lot size and profit on your TE, please? I'm also using free tier AWS, so I would think you and me have same playing ground on the server side, but...
Yes, it's an impressive start, but I'm under no illusion that it'll be like that every day. I still believe, though, that it has the potential of making up to 10% per session, provided that the PA is favourable and there are no technical glitches. Some manual intervention would probably be necessary to make the best of it, as with any autotrader...

As for your questions:
From what I've observed, it takes around 10 sec to close all trades and pending orders. I don't mind that as long as the market is moving; that's the main key to success here, imho. With Armada's tiny spread and Gold's ATR, I believe it's only logical to be aiming at a higher span of the orders, hence 60 steps.
Equity = EquityGainMultiple.
Sorry, didn't realise the TE is not showing everything, I'm more used to myfxbook... Going to make everything public now. It's a humble £1000 account.

By the way, it's a well known secret that Armada is one of the most EA-friendly brokers out there.

Regards,
Kosmo
mind over matter
  • #318
  • Edited 10:36am Dec 5, 2014 10:19am | Edited 10:36am
Quoting kosmo
{quote} Yes, it's an impressive start, but I'm under no illusion that it'll be like that every day. I still believe, though, that it has the potential of making up to 10% per session, provided that the PA is favourable and there are no technical glitches. Some manual intervention would probably be necessary to make the best of it, as with any autotrader... As for your questions: From what I've observed, it takes around 10 sec to close all trades and pending orders. I don't mind that as long as the market is moving; that's the main key to success...
http://www.kitco.com/images/live/nyg...07787682455684

You may be on to something with using 2 for equity gain...it may be small enough because of the incredibly small spread, but more profit each time.
It closed over 400 trades in LO, though! and you cut if off right after NFP, didn't even let it run the rest of LO
that's kinda insane! I only got 41 trades, most of which are still open, and that's since the beginning of Tokyo last night!

To compare, Finfx seems to be floating around 0.50 spread today. This means Armada accounts could, theoretically, beat it 4X by two ways:

1. The equity target could be doubled, giving effective spread of 0.20, still less than half!

2. You could double the number of steps, like you said, to keep from getting "boxed in" with just hedged orders, as long as margin allows.

Too bad I don't have a prayer of using them, in US.

Somebody mentioned likening this to HFT...well, sorta, I guess if the market moves really, really fast in one direction for a long time...but HFT has the ability to fire market orders, and then modify them 10,000 times in a single second. That's worlds away from MT4 capability. If only I could do that....
Too close for pips, switching to points.
  • #319
  • Dec 5, 2014 10:56am Dec 5, 2014 10:56am
Quoting kjaved
{quote} No, Equity Profit and Trades P/L are two different matters. Global variables are bound to the current MT4 instance not all the MT4 running on the same machine.
Could you please explain with a difference in case of this EA. I fail to see a difference!
  • #320
  • Dec 5, 2014 11:12am Dec 5, 2014 11:12am
Quoting Thepipster
You may be on to something with using 2 for equity gain...it may be small enough because of the incredibly small spread, but more profit each time.
In my mind, it's all about finding the right balance for the three components here: the broker, the tool and the nature of the market traded. Once we get that right, we should be fine...

Let's not forget, either, that there might easily be a dark side to today's fairy tale; if it can make this much in a matter of minutes, it might as well lose as much. It's all good and well in demo environment, but I do wonder what the spread/slippage/execution would be like on an NFP day on live...
mind over matter
1 trader viewing now