It is an API bug, but I got no clear information about its reason. I think that it will be gone with the next FXCM API update.

Thanks. Has FXCM said when that next API update will happen?

No. The FXCM information policy about anyting related to the API is a little opaque, to say the least.

If only there were a better broker...

Thanks.

Hi jcl. I now have a reverse-orphan trade that I need to know how to handle. Reverse-orphan means it's closed but Zorro still thinks it's active. Here is what happened, and the questions I have:

a) On Friday, Z12 opened this last trade before Weekend:


b) Today (Sunday), Z12 began to try to close the trade while trading was still closed (per Trading Station) and while Zorro was still in Weekend (at default 2). These messages began to repeat:


c) Question 1: Since FXCM Trading was still closed (per Trading Station), how valid/bogus were the prices Z12 was receiving and on which it based the close decision? Is Weekend=2 the correct value to use for FXCM?

d) Since all Trading was still closed, this trade could not be closed, but Z12 kept trying & logging messages like the above every minute or so. This seems to be normal functioning for Weekend=2 from what's in the manual, except that at each attempt Z12 went "Not responding" during most of this time - Question 2: Is Zorro being hung up by the API?

e) Immediately when Trading opened at 21:00, Trading Station reported the trade as closed. 2 problems:

f) Question 3: The close price recorded in Trading station is 1.61551. The price in Zorro's log at 20:00 is 1.6199. Why the difference?

g) Zorro did not recognize that the trade had been closed. The messages changed and the bogus trade # makes its 1st appearance:


h) At 23:00 Z12 went off Weekend and the logged messages became:


i) The last 4 messages have repeated continuously since then, and Z12 still thinks the trade is open. The bogus order number appears to be time-stamp related, as the 6th digit seems to increment every hour.

j) Now, whenever Z12 tries to close that trade it goes "Not responding" for about 20 seconds. The VPS is also running 2 other Zorro instances, and those 2 also go unresponsive (as judged by their clocks not incrementing) for the same 20 seconds. Question 4: Is this because they're all running through the same API? Is ticks processing for all also suspended & ticks dropped?! Is this a reason to not run multiple instances on the same machine/VPS?

k) And, of course, the final Question 5: How do I get Z12 to recognize that the trade is closed?!

Thanks.

This is a good question: Zorro detects if a trade was closed from its appearance in the Broker's close list. But if the closing was several days ago, it's not anymore in the list and thus not detected as closed.

We had never observed such a case, so none of the Z systems has a script function for discarding a trade. You must ignore that problem for the next days, until the upcoming Zorro update. We'll provide a solution for removing a trade from Zorro's list.

jcl, the closing was yesterday evening less than 24 hours ago, and as of a few seconds ago the trade is still in the Closed Positions tab of TS - is that the closed list? Perhaps there's another bug? Or the trade # API bug is returning the wrong (or empty) closed list to Zorro?

Zorro continues to produce the same error messages - not harmful but certainly messing up the window & log file...

Thanks.

Is 58631732 also the ID in the TS closed trades list?

Yes.

I was just informed by my colleague that Zorro does in fact remove such a trade, after 50 failed attempts to close it. Please check if this really happens with your Zorro - we never had 50 failed attempts so far, so this was only tested by simulation.

The problem is obviously caused by the wrong ID.