Again 6 (!) strategies stopped!
Forums › ProRealTime English forum › ProRealTime platform support › Again 6 (!) strategies stopped!
- This topic has 86 replies, 7 voices, and was last updated 1 year ago by PeterSt.
-
-
02/22/2023 at 9:51 PM #210193
I’ve just found another / additional 19 Systems stopped … across 2 x Demo Accounts!
Another 17 Systems stopped today, all on Demo Accounts.
Reasons according to the error message were 16 x ‘exceeds the number of retries’ and 1 x ‘status of last order is unknown’.
Same Old Same Old?? 🙁
02/22/2023 at 11:50 PM #210196The systems on the demo accounts are stopped. happens to me all the time, not often, but irregularly.
Mostly mentioned problem is “not enough capital available”, although there is really enough money lying around on the demo account and would have been enough for 50 more trades….
I do not ask why and why, I restart the systems patiently, the final results of the system then no longer fit properly, but change? how can I change it? Does the support do anything? For demo accounts? All initial attempts years ago have failed.
Demo accounts are made for beginners, they do not even notice or accept it at first and are comforted with the words on the live accounts does not happen … hehe1 user thanked author for this post.
02/23/2023 at 1:26 AM #210201Demo accounts are made for beginners, they do not even notice or accept it at first and are comforted with the words on the live accounts does not happen …
Mostly mentioned problem is “not enough capital available”
haha, well, it would be quite wrong if someone would comfort an other with saying “that won’t happen in Live”.
This is a general remark of course. 🙂
Anyway, I never heard of that happening in Demo unjustified, but I sure saw that message some times on Live … And I can tell you, it was true and it is a good thing (when true indeed – and it always is true when it happens).02/23/2023 at 1:36 AM #210205Reasons according to the error message were 16 x ‘exceeds the number of retries’
GraHal, you’re mentioning this more often, but that in itself never is a reason. Thus, something has been tried for x times (and that for 16 Systems as you say), and the reason can always be found. Just look some deeper or harder to find that reason. Thus example below. The why is in the message itself but the data will differ. See second screenshot. Same sequence of attempts, but the data has changed because the price changed (but did not solve the issue). The attempt (usuaally 10 times) is an attempt of the same.
02/23/2023 at 2:08 AM #210209Scrolling through the recent data, I was astounded to find what you see below. This in itself happened 3 times today and this did NOT incur for a Stopped System. This means
- that this BS is also happening unnoticed;
- that I apparently miss out on (good ?) trades I am not even aware of.
The reason in this very cases is the same as talked through yesterday – they are non-existent but occur anyway. There is no such thing as a minimum distance “wrong” or anything. So for PRT’s information, the 25,24 number you see in the 2nd pic in the middle, is the number of executed trades (in and out) on average per day (over 1500 trades registered which 1500 is the limit which PRT can cope with for the query accordingly (grrr)). IOW, these underlaying Systems (always the same) work all right. But IG fails not-so-all-right on … what ?? 10 percent of occasions by now ?
I’d wish I did not see this. And maybe it is even new because I never saw it before. So again : today for 3 different occasions no Entry occurred because it was rejected for the same stupid reason (I added the 3rd attachment to show a 2nd occasion). But I did not notice because the System was not stopped for this. To be complete, there also was no Executed Entry – see 4th attachment for the executions around that time. Thus, one attempt for Entry which failed for BS reasons, and away is my opportunity.
One more crucial remark : My local support knows that since last week I filed complaints about for me totally unknown situations that backtesting in Forward mode is not in line any more with Live. I now suddenly can see why because Backtesting (which is thus Forward testing when just running during Live) will obviously not reject this.
And oh, I have collected dozens of *videos* by now. Time to post them ? … good for PRT because they are not allowed to receive links to videos unless YouTube.
Maybe a small warning : by the time I start to post these videos, PRT can better stop advertising itself, because these videos really don’t justify anything good; the very very contrary.I restart the systems patiently, the final results of the system then no longer fit properly,
Please understand what John says here. If you don’t, re-read it until you do or can envision what this is about. A derivative of this is that the system itself can’t work because it builds up history which is destroyed each time the system is stopped – while the system depends on this history; It is too stupid for words that I am already trying to plot current data (forward testing graphs) so I can know this history when the System is thrown out and I can enter that for base data when I restart it. Don’t believe me ? see next post because this one is full for attachments.
02/23/2023 at 2:08 AM #21021402/23/2023 at 8:53 AM #210228Thus, something has been tried for x times (and that for 16 Systems as you say), and the reason can always be found. Just look some deeper or harder to find that reason
My (maybe wrong?) assumption is that another Order from the same Account on the same Instrument in the same direction are being attempted? Like many folks all trying to squeeze through a single turnstile gate??
In my case, I run several similar versions at same time and so the above scenario may be worse for me … hence me getting more Rejections (for > than allowed number of retries) than most folks on here?
If above is the case, then it will keep occuring for 10 attempts until Rejection.
Currently there is a 1 second delay applied to all Orders, but then after 1 second, the several same Order Systems, then cause congestion again, then 1 second delay, then congestiond again, then 1 second delay … etc etc until Rejection of the Systems after 10 seconds or so?
Surely all it needs is a ‘staggered delay‘, or similar creative solution, across similar prohibited Orders?
If above were built into PRT, then all / most Orders may get through and not cause the prohibited … same Order from the same Account on the same Instrument in the same direction all being attempted at the same time??
02/23/2023 at 10:07 AM #210240In my case, I run several similar versions at same time and so the above scenario may be worse for me … hence me getting more Rejections (for > than allowed number of retries) than most folks on here?
I got a lot of strategy stops before when I had several of the same strategy, but with small variations. But the error message was different, something along the line “Stopped due to duplicate order”
02/23/2023 at 10:24 AM #210242something along the line “Stopped due to duplicate order”
Yes I used to get above loads also (on similar Systems) but not for ages now, so I think IG have selected the … ‘exceeds the number of retries’ … as the catch-all error message!? 🙁
02/23/2023 at 10:54 AM #210247(Not) Coincidentally I have been thinking the very same this morning. I too run many systems in parallel and on the same Instrument. But my thoughts for the cause were – and are a little different :
I know from PRT-IB what a gigantic mess it is going to be when more than one system is let loose on one Instument. From there you can see that nothing is made for it.
ehhhNo wait, with PRT-IG we can work with that ?
To me it is quite clear that PRT itself can deal with it, or else nothing would work with PRT-IG. Mind you, with IB it really is a mess because just nothing works. Orders stay In while they should Exit. Opposite orders can’t work at (cause is IB) all but *are* allowed. And …
… Take the opposite orders as an example. PRT surely and definitely can work with them. But not IB. And now things need to be prohibited or else all messes up. This is all about the communication with the broker, and my ongoing more knowledge about this because of so many issues with manual trading (the 200+ tickets for that, all unsolved).
Okay …So now we have these several running systems, and they communicate over the order IDs. We can see on the failed Entries that there’s no Order ID applicable (it is just not there yet). You can see the assigned Order-ID’s in the top part of the screenshot below and the not-yet-assigned in the bottom half; you can see this by the right-most column where instead of the Order-ID your Account-ID shows.
Now what if the Account-ID as the communication means mixes up the orders and from there the PRT software fails badly. All it requires is two orders at the same time (same beginning of the next bar). For me it now seems logical that with the account-ID as the key (think database now), this just can not work.
This not unique key over functional several orders looks to be in order definitely for not yet put-through orders (your question to the broker “can I place this order please”). But what about the already running positions then ? they goo out (are stopped) just the same for various uncomprehensible reasons;
From IB I know that in a 6 second Strategy the time required to place the order, is often too short (I have the videos of it). Thus, IB answers too slow. The cause thus is IB ? well, Yes, but PRT can’t cope with that situation and starts to mess up. Envision : No response from the broker after 6 seconds and there comes my next pending stop order already, that too handed to the broker. But that can’t be because PRT first kills the order from the previous bar. Well, by that time the broker responds finally to the order of the previous bar but for PRT that order does not exist any more. Unexpected situation and BOOM out is my System. With this story, the videos clearly show this. It is only that I have to make the story regarding what I see. But it is almost by guarantee that my story is correct, already because I can predict the situations and know via the code how to avoid them.
Now don’t you ever think that PRT solved this even the slightest. Beta tester for PRT-IB Autotrading, all made clear by video and all (not seen in Paris) and I hear nothing back about it in even (on estimate) 2 months of time.
OK, that was IB. Now I wouldn’t know why the PRT software suddenly would work fine regarding this communication between provider (PRT) and Broker (IG) . I know for many facts what IG all can incur for, and I most certainly know what PRT “refuses” to cope with (IB story above).
And now the intermittent conclusion :
Because the both don’t communicate very well all sorts of strange messages are thrown at us. Or maybe not strange messages per se, but what about messages applying to orders which are the wrong ones. If that communication does not understand – now we even understand less.
IG applies her new trickery to earn some more money, but IG is known for living on her island and ProRealTime is not told about this. Next for us another thing stops working (well).
But GraHal, in the base you are right, I think. I don’t know what right today makes us think about the several systems in parallel for a cause, but that in itself will be no coincidence. I don’t know what made me think so this morning; maybe psychics of human can let this surface. Then probably that would be a good starting point.
02/23/2023 at 12:07 PM #210268‘exceeds the number of retries’ … as the catch-all error message!? 🙁
That *is* no error message. And otherwise show one (in full !) ?
02/23/2023 at 7:16 PM #210328I didn’t mean a catch-all error message to cover every Rejection / stoppage, but just to cover “Stopped due to duplicate order” (as paisantrader referred to) as well Rejection for ‘too many retries’.
03/06/2023 at 10:55 AM #210988Another 17 Systems stopped today, all on Demo Accounts.
12 days passed (includes 2 weekends) since NO Stoppages for ‘exceeds the number of retries’. (No more than a few Stoppages for any reason over this same period!)
I am still running the same Systems that gave 50 Stoppages for ‘exceeds the number of retries’ over the 4 day period up to 22 Feb 23. There has been no update to PRT since 16 Feb 23.
So what has changed? Market conditions not been same / similar to the period 18 Feb to 22 Feb (during which I had 50 Stoppages) or has IG sorted their Server / Network Issues??
I guess we will never know, but I just wanted to document this apparent improvement in functionality so that we can all rejoice!! 😉
03/07/2023 at 4:44 PM #21104803/27/2023 at 8:45 AM #212175 -
AuthorPosts
Find exclusive trading pro-tools on