V12 for PRT-IG now live on Live

Forums ProRealTime English forum ProRealTime platform support V12 for PRT-IG now live on Live

Viewing 15 posts - 61 through 75 (of 98 total)
  • #227661

     

    Still no movement on the 1sec data ….

     

    What you said makes perfect sense there Peter. Many years ago when I was younger (ha), we’d trade very tiny margins (1/4 etc) for full size Ft100, talking $AUD300 with per side at $15 but the L2 and access fees were quite big (levied by the exchange, pre-ice). This was through a prop outfit in AUS that was rolled over a few times before Corzine got his hands on it as it was known back then MFGlobal. Was mid 90’s from memory (the lowest margins) and the Indexes were quite hot then too comparatively. Back then you’d have a pager on a PC, big tech, lol, you wanted charts, you drew them up of pushed them though prop.software.ugly. But the world was a kinder place then too, maybe.

    TBH with a full-sized firm, futures desk etc, access in AUS was never a problem (mid 90’s) but regulations and the advent of CFD being available to the gen.pop means that it’s all changed and govt and regulators ‘intervene’.

     

    I mean since when did access become so hard and convoluted for us, AUD acc. sizes of 10k and AUD300 day trade margins mid 90’s for a 10pound tick was very good value for a service that went L2 @ ~140/mth for ALL EU futures and APAC, if you wanted NA it went ~AUD200. So while PRT’s offering are still inline with that and the value is there, the local service desks are closing down (in AUS) and you’re stuck with IBKR etc with a few hoops to jump through depending where you live and whether you could even Day Trade at one point. Early days MM were seen as a kiss of death, now they are the BIG companies.

    So yes, trading blocks could exist further more so but that was what MIFID  and v2 was about, est. the framework to interop and it works well when you have big accounts, no dramas, go with a big commodities desk and off you go but you’ll want north of AUD2.5mill to go wholesale in a family office. Can of worms but I can see further challenges down the road.

     

    I’m not unpleased with what’s on offer for me with PRT as a part-timer, I know very little about tech but because it was so fluid here in AUS a long while back to the extent I could call my broker on a Sunday while he was having a BBQ with family and while I still have that contact the worlds a very different place now, staff turn over is big and we’re being over regulated in an effort to save us from ourselves apparently.

    TBQH some of what you say has a net neutrality ‘feel’ to it, aka, depending on who you are with, based on domicile will depend on access – more trade blocs here we come, assuming I live long enough lol. So yeah agree.

    I never accessed PRT with an API so not sure what was done then (2019).

    I’m sure I’ve missed the point somewhere in there but I think we’re on the same page ….  🙂

     

     

     

     

    #227663

    At least we would be on the same page of us now having a BBQ ourselves with the trade phones on the (table-)desk. Or at least have a nice view on it (I use remote tablets as Windows monitors).

    Can’t you use IG (with PRT) ? I think that gives you more possibilities for markets.

    #227692

     

    They we just fun times, everyone has there own story.

    Yes, PRT is doable but if you want more than a few days 1s data (maybe as of late, not sure) it’s lacking for IG customers and the standalone PRT doesn’t have APAC I think. Realistically speaking if I get any more serious about it (and can’t live with ~7days of 1sec) I have an old licence of MarketAnalyst that supported IQFeed and Bloomberg-term underneath, they now roll with monthly access fees I think plus the exchange cost, so while you own the license there’s a fee for real time data feed ‘enabling ‘ aside exchange royalties. (EOD is free).

    It’s not the end of the world but I’m sure there was more than a few days 1sec data for IG customers a little while back. V12 has introduced some oddities and I haven’t done a complete uninstall of the client to test but a long while back when I looked at it, when 1s did play up, it’d build the ‘window’ and you’d see data add in but that was v10-early v11 I think. V11 worked and it could be all the massaging underneath that’s rolled the 1sec data component.

    I first looked at PRT as a supplement to the old PATS J-trader setup I think, java charts, and while agricultural it got the job done for me, so PRT was a refreshing upgrade back then and while it’s so much better now, broker-PRT interactions (as you have seen yourself) can be tricky to work through. My leaning is that the MM care less about the smaller tweaks whereas the developers have the real skin in the game and whenever I’ve asked for help the PRT team have been forthcoming – it’s hard to complain about an almost zero cost platform haha.

    Can you remember whether there’s any localised caching on a Windows PC that can impede the data build and a simple deletion allows for a full download to process, giving you more than a few days of 1sec … ?

    Apologies for the TMI on my part.

    Thanks for taking the time to correspond!

     

     

     

     

    #227694

     

    Starting to wonder whether this didn’t start to wobble with a Java update 24/1.

     

    Running Version 8 Update 401 here. (build 1.8.0_401-b10)

    Dow now only showing data from last session at 1sec ….

    #227698
    I’m starting to think I’m triggering to some of these issues because of a new template I’m working with and perhaps confirms that 1sec has been upside-down for a while now (see below). For example, my 225 chart has ASX there but also DAX FTSE as we roll into EU but DOW was hidden from view.
    We’re also running with OpenJDK underneath here and probably have been for a while now.
    To be honest, other than increase avail. ram to 12MB I’m not at all too keen on having to force another rev. of Java as a test, not sure if anyone else has taken this path???

    I recognize all you say there. I don’t think this has changed; the “seconds” Timeframes always behaved weird compared to Minutes.

    #227700
    I get about 3 days history on 1 sec TF on 200K bars … which is about right as there are 86,400 seconds in 1 day.

    1sec has been upside-down for a while

    What does upside-down mean?

    Also no attachments in your last post.

     

     

     

    #227702

    but you can’t ‘zoom-out’ and have it fit in a window all nice and proper like.

    Is attached what you mean … see date in bottom left hand corner … Monday, so 3 days worth of data showing for 1 second on 200k bars … which is correct for 3 x 24 hour days.

    I know it’s early and my coffee has not worked through yet, but why am I misunderstanding what your problem is? 🙂

    1 user thanked author for this post.
    #227707

     

    I think it’s because I had quite a bit more 1sec data historically, or so I thought, I’ve not pulled the calculator out lol to see if I’m trippin’ but when I compare multiple indexes it’s a mess and I’m not sure if it’s custom time zones underneath impacting or else.

    So, I’m a WIP atmo, when last week and a bit it was all G.

    Come to think of it, it was the inclusion of weekend data that hoses the amount of  1sec viewable. If you ‘uncheck’ this you see more 1sec but without the Saturday mornings close. It was a trade off and it was this I ticketed, from memory.

    So it is SOP and I’m barking mad 😛

    I am starting to wonder IF I’d not seen / used 10sec charts as a trade off but the more I’m thinking about it, it’s the weeks closing session (sat morning), that’s causing it.

    Have a look and see if what I’m seeing or am I crackers….ha!

     

    edit: a quick test shows same data here, which wasn’t what I saw when ticketed.

     

    There’s also the issue that I can’t limit DAX and FTSE to there close time on v12.

     

     

     

    1 user thanked author for this post.
    #227709

    Monday, so 3 days worth of data showing for 1 second on 200k bars … which is correct for 3 x 24 hour days.

    But then on AUS 200 we have data back to 30/1 …. so a mixed bag. But you need to use the scroll bar to view it all and it’s here that I can’t zoom out to view all, whereas on 10s it’s doable.

     

    Odd or me maybe.

    #227713

     

    Yeah, it’s become unusable with mem. usage atmo, even with 12GB set……it’s as if whatever was fixed a little while back and was working for me, from memory 18/12 release date was all good.

    We’ve since also gone to a late JAN release which is a problem maybe with some of the studies.

     

    I’ve gone to an older template and it’s borked at best right now.

    #227720

    I know it’s early and my coffee has not worked through yet, but why am I misunderstanding what your problem is? 🙂

    All good Gra never enough coffee for me … :))

    I’m still unable to kill-off DAX session at 22.00 as I used to and still can under v11 – IG know this.

    In your image I think I can glimpse a scroll bar along the bottom, I think, if so, does that mean you have data going back a little further? If so, this is what I mean by being able to zoom out proper. I see something similar on ASX and also DJIA, as I think you might be too. Does 200k units = viewing area? 200K units @ 1sec = ~2.2days ….. looking for rabbits down this hole atmo

    Most interesting is I can NOLONGER reproduce the effect ‘weekend’ data had on ASX – w/o it we’d get more data 1sec going back, now it’s static. 30-31/1. So the problem’s been resolved by limiting the viewable 1sec data ….

    I can’t be certain when it went RS but the latest rev dropped 24/1 and I’ve been  caring for a family member and it’s not been till now that as I’m documenting more that I’m finding variation (that could yet be me and my memory).

    This was the last rev. that I was running until recently (and all was well)  I think???

    Version : v12.0.20231121030018

     

    First thing I’d do every preopen is max the units to 200k and then zoom out – I don’t think I’m imagining it all…

    TBH I might just roll with 10sec for now, assuming the platform doesn’t leak to death.

    Thanks for hearing me out and perhaps helping me reason it out a little too!

     

     

     

     

     

    #227726

    if so, does that mean you have data going back a little further

    Below done at 10.45am 9 Feb 24

    1.  Select 200k bars and Mon 5 Feb 23. 32 shows.

    2.  Scroll / zoom out and Fri 2 Feb 19.21 shows

    platform doesn’t leak to death.

    v12 performs fine for me … I am not aware of RAM leakage, I have 16 GB RAM.
    How many Charts do you have open at any one time when you experience leakage?
    I would try and duplicate same on my Platform to see if I get leakage under same conditions as you?

    I’m still unable to kill-off DAX session at 22.00 as I used to

    What does above mean?
    If you have a Chart showing DAX, do you mean you can’t close the Chart?
    OR
    Are you referring to an Algo and the code below doesn’t exit / close open trades?

     

     

    1 user thanked author for this post.
    #227727

    May it help anywhere :

    1. 200K data can’t be in one window (viewport). Possibly this depends on the monitor’s resolution, but over here it can’t (resolution-width is 2560). Btw, see below for 1M bars 1-second data in V12 (PRemium).
    2. When you load 200K bars, and assumed of 1 second, then more will be added on the fly. Thus, leave your platform and particular chart open for 18 hours, and 18x60x60 will have been added to the 200K.

     

    Further, although the data is cached at fetching it (it is read in chunks), there is no storage space other than internal memory where bar data is stored. It is Java which applies the fetching.
    But as GraHal told/implied, with a theoretical 86400 seconds in a day, you won’t get more than 3 days anyway. Something else is : I have a quite strong feeling that less liquid Instruments may receive more data in seconds (bar is one second) because not every bar contains ticks (hence data). I have not sorted this out yet.
    Another fact (as it seems) is that things are relatively collapsing when backtesting with “seconds”  data. And for IG this is far worse than for IB, already because IG generates 10x more ticks within a time period. Anyway, backtesting with 200K seconds data is relatively undoable, while with the same amount of 1 minute data, all is (again relatively) fine.

    1 user thanked author for this post.
    #227730

    1 second graph.

    Voorbeeld van afbeelding

    1 user thanked author for this post.
    #227731

    What does above mean? If you have a Chart showing DAX, do you mean you can’t close the Chart?

    Yeah apologies there, getting sloppy …. the night session closes (or used to at 22:00), on v11 I’d set the closing time w/in timezones as 22:00 (via the timezone slider), I don’t leave it at 24hrs for example but it’ll revert there upon restart (will not save). So effectively on the chart you’ll chart till 22:00 or try too.

    How many Charts do you have open at any one time when you experience leakage?

    All up ~25charts etc, plus the studies underneath.

    It’s mostly when customising ‘colours’ (this evening in particular) and saving indicator tweaks, when left alone, aka do it once and get over it it’s generally ok and is ok right now because I’m done with tweaking colours but I’m also allocating 12GB to Java. on a 32Gb PC. So today when poking around ‘setting it up’ and given the templates are bigish it’s been a bit of a pain which is an outlier for me.

    What ver of v12 are you running Gra?

    When you load 200K bars, and assumed of 1 second, then more will be added on the fly.

    Yes, have seen evidence of this too but historically it would ‘back fill the charts’, even IG said as much while on a  call with them or it would for me. Now whether this is an SOP or occurs in a break-fix more specifically, I can only say that in a break-fix
    scenario it would or has done and I’ve not had 1sec data problems since I went off tick because there were issues there.

    Thus, leave your platform and particular chart open for 18 hours,

     

    Makes sense as I generally start at 6am and finish up 11pm, so would make sense to ‘fill’ data – thank you for pointing this out.

     

    Regarding resolution I’m 2560 here too and have seen more data in a window but that marries with your earlier comment.

    Starting to think that it could ‘recover’ if I stopped reinstalling and alternating Java etc (pointless).

    But as GraHal told/implied, with a theoretical 86400 seconds in a day, you won’t get more than 3 days anyway. Something else is : I have a quite strong feeling that less liquid Instruments may receive more data in seconds (bar is one second) because not every bar contains ticks (hence data).

    Thanks for this.

    I’ve cleared the install and java (a few hours ago) and have ‘data’ back to 31/1 and yes the view port is 2.5-3days however scrolling left exposes the data back to 31/1.

    Your point on dubious data (when illiquid) isn’t lost on me, so thanks again both you and Gra for digging in here.

     

     

     

Viewing 15 posts - 61 through 75 (of 98 total)

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