0 Second Transaction

Forums ProRealTime English forum ProOrder support 0 Second Transaction

Viewing 15 posts - 1 through 15 (of 32 total)
  • #200253
    B&L

    I’m experiencing a lot of 0 seconds transaction lately, where PRT triggers an operation BUT closes it automatically on the same instant, usually generating a loss of the spread cost ( With IG ).

    It always happens when the operations starts on the lsat second of the candle of reference. Let’s say that if a candle lasts 15 seconds, if the price is hit ( Stop or Limit orders ) on the last second, BUT the operation is executed on second 1 of the new candle the operation is triggered and then closed without letting me decide nor operate it.  Seems that PRT servers recognize it as a mistake and forces the automatic close of the operation to IG

    It also happens when I try to turn directions and the sell stop order is hit BUT the sellshort positions is opened a second after. In this case PRT trigger a even weirder order and re-opens a position similar to the one that was closed AND deduces it from the sell short position that was opened a second prior.

    Is anyone here experiencing  similar issues and/or has found a way to avoid it?

     

    #200284

    Does it happens on demo or real trading?

    IMO, it seems related to a code issue rather than a bug. Would you share the code with us to understand what’s happening here?

    #200302
    B&L

    Happens Live,

    across with all codes that I may create. Whether on 10s, 6s or 15s timeframes. If the triggers happens on the last second of the candle but it’s run by IG on the next candle the rejection happens.

    I’ve been reporting it to PRT and IG since already a few weeks but seems that they have no fix yet

    #200311

    All code is read at the close of a bar and resultant trades are executed at the open of the next bar (a millisecond or so after close of previous bar).

    Would above account for what you are seeing?

    #200312
    B&L

    Let me try to explain it a little bit more in detail with an example.

    Candle length: 10seconds

    candle[1] ends, reads code and launches a BUY stop order for the candle that starts. Target is 3 points above current price.

    At second 9 of the current candle the price reaches the stops and triggers. Sistem execute de order BUT between back and forth PRT-IG-PRT the executions happens on the Second ZERO of the new Candle[+1].

    PRT or IG recognize this as an error as it was supposed to be triggered on the previous candles and closes the transaction in the same instant.

    An example here in the picture. The candle finishes at 48sec but the transaction is triggered on second 49

    #200315

    So you say that the pending stop order that has triggered which is now a market order is automatically closed by the platform at the opening of the next candle?

    I do not see any reason why the platform would close a market order if there is no code to make it so, such as “SELL AT MARKET” or “EXITSHORT AT MARKET”.

    #200323
    B&L

    I do not see any reason why the platform would close a market order if there is no code to make it so, such as “SELL AT MARKET” or “EXITSHORT AT MARKET”.

    I do not see any reason too, but it keeps happening anyway… The stop of those algos is set minimum at 20pts behind the price. When it happens I just get a simultaneous notification from IG with open and close. On the log in IG is also at zero seconds

    #200341

    Do you have the ticket number from IG please?

    #200351
    B&L

    The one on the picture was the DIAAAAKANENYJAP.

    This morning I had partial close on a position, due most probably by the same error. The strange thing is that the ID on PRT on that case after the  DI********_C for “close” in this case adds a  DI******_c_1

    #200374

    I need the ticket number for your support demand to the IG support desk, not the ID of orders 🙂

    #200391
    B&L

    Some of the tickets for the Zero second transactions:

    Ticket#2022090110012767

    Ticket#2022072910018589

    Ticket#2022080810002874

    Ticket#2022072910018589

    Ticket#2022080810001599

     

    This ones are related to similar issues but that happened when turning the positions:

    Ticket#2022082610016932

    Ticket#2022080310023191

    Ticket#2022080610016872

    1 user thanked author for this post.
    #200461

    Hi B&L,

    I feel this is related to a new order in the opposite position first cancelling the running position (no matter the quantity of it) before the new order can be given to the exchange.

    Please read the above a couple of times, because it is hardly visible (and not intuitive) that it happens like this.

    Crucial for understanding : there can’t be a Short and Long position at the same time in one System.

    When you’d have 1 position Long and a Limit is triggered for a Short of 1 position, a Market order will first kill the Long position. Most probably this Market Order will lag (because a buyer needs to be found) so it will (a kind of coincidentally) show up in the next candle when the event of the Limit was triggered close to the end of the bar (like last second).

    When you’d have 1 position Long and a Stop is triggered, all bets are off because that implies a Market Order in the first place. At least this is so complicated that I would not dare to write it out (but Google Limit vs Stop and you will see it is described everywhere).


    I’d like to bet you (but I may lose) that all is perfectly normal but that it is about misinterpretation of what actually happens at the exchange (vs broker, I’d say). I am also pretty sure that you know very well what you are talking about. However, miss one detail and things may become incomprehensible. 🙂


    The strange thing is that the ID on PRT on that case after the  DI********_C for “close” in this case adds a  DI******_c_1

    Although this won’t be the situation you largely described, you should try to call the IG order desk and ask them what the _1 addition means. You will get an answer and from there you will start to understand what is happening. I think I can already tell you that this is about what I described in my first lines above : an IG injected order to first set things straight (can’t have Long and Short at the same time).

    I hope you keep us posted with what you find.
    Regards,
    Peter

     

     

    #200467
    B&L

    Hi Peter,

    Yes, indeed in the case of the partial close it is what I reported to IG, that for one second they are running a short and a long position while this should not happen on PRT. What came out is that even though the order to sellshort stop ( Opening Short) is met and sell stop (closing long)  is met at the same time if one is triggered with delay a second order from PRT arrives asking to reopen the position of the long contract as it sees it as a mistake. Here comes the big problem, for IG the order of reopening comes from PRT servers.

    ( the _1 tag it’s from PRT not from IG. Even to IG desk that _1 has no meaning )

     

    While this can be an explanation of the “partial closing” it does not solve at all the incauntable time when PRT closes  a trade just on the same instant while is triggered.

    #200491

    While this can be an explanation of the “partial closing” it does not solve at all the incauntable time when PRT closes  a trade just on the same instant while is triggered.

    At least these I have seen too but I don’t recall the exact situation. To me it felt like *because* a trade was triggered  the other one was closed.
    Small problem for me at this moment : I save these situations (numerous per day indeed) but the server I use for that is currently down. So now I don’t know when this was, which also implies that I don’t know which (kind of) code triggered that situation.
    Currently I don’t run into this situation any more because I quite explicitly took care of  no more than one execution within one bar happening (for a reason I don’t know at this moment as well).

    a. Is this a situation that ran well in Backtests ?
    b. If so, can you show a screenshot of a backtest equity curve as how the PRT Statistics would show it ? See example below. If it’s the same as my ever back situation I will recognize it.

    PS: For IG I can’t tell by 100% guarantee, but what you see below are the Order-IDs from IB (from PRT-IB); I know because I worked with the IB-API as well. And I see no real reason why PRT would create her own order numbers for IG while she doesn’t do that for IB.

    #200499
    B&L

    Running on Backtest the operations always show that they should have been run on a previous candle. ON the images there is one from yesterday. The yellow tick is when it was supposed to trigger ( price was there) but was on market the candle after with the red tick. There is also the “close position” with a green tick which of course was not taken from my code as it happens on the same candle

Viewing 15 posts - 1 through 15 (of 32 total)

Create your free account now and post your request to benefit from the help of the community
Register or Login