PRT did it again…

Forums › ProRealTime English forum › ProRealTime platform support › PRT did it again…

Viewing 15 posts - 1 through 15 (of 15 total)
  • 09/04/2017 at 10:11 AM #45343
    Despair

    What a lovely monday morning. And again I login to PRT and have to find that the strategy I wrote during the weekend is gone without traces. But it is my own fault, I did not export it and relied on PRT to save it.

    Don’t do that folks! Export always your work before exiting PRT!!!

    09/04/2017 at 1:09 PM #45360
    Andre Vorster

    What PRT does is they reboot/reset their servers every 18hours.!

    I don’t understand why any company needs to reboot/reset their platform to update it. Especially systems that other handles other people’s money.  Being in IT and installing, maintaining and running mission critical systems that runs across 100s of servers that includes big databases’ it does not make sense.

    I found that running backtest with WFA also gets cancelled because of PRT’s reboot. As we all know this process can take many hours. So then, the next day you have to start all over again with the backtesting.

    So what to do…?

    IGs response to this:
    “With activity on ProRealTime, it still needs to shut down/log off after 18 hours to restart their servers to make sure data is coming through correctly. Unfortunately this also can’t be changed, however, I would suggest re-logging into your ProRealTime if you have a lull in your trading during the day, or wish to start a new system.

    ProRealTime are also always working towards improving their platform and I believe they are constantly working to try and make the log off requirement longer.”

    09/04/2017 at 1:33 PM #45362
    Despair

    I’m aware of this resting. At least they increased it from 12h to 18h. BUT this time that wasn’t the reason. I had closed the platform and saved when exiting as usual. Anyway when I started PRT today the strategy was simply gone without traces. Not the first time…

    Luckily I had most of it fresh in my mind and could rewrite it but this is REALLY annoying.

    09/06/2017 at 8:51 AM #45568
    juanj

    @Despair, could it be that PRT didn’t actually save the strategy upon exit? I have noticed that sometimes PRT becomes unresponsive to the point where it no longer executes command such as Save etc.

    1 user thanked author for this post.
    avatar Andre Vorster
    09/06/2017 at 10:25 AM #45593
    Andre Vorster

    What I have seen is:
    You have to close the strategy after you made any changes. Just normal Ctrl+S does not save it. Make sure it is listed with it’s own in the list with other strategies.

    My experience:
    Opened a current strategy – made some changes to it. Saved it (Ctrl+S) and ran it. When PRT reset their servers the strategy was gone.

    Duplicate a strategy – open it – make changes – change the name of the strategy – save it. Duplicated the one strategy 4 times and saved all of them with different names using Ctrl+S keys.
    But I did not close them to get confirmation to save it.

    One must close the strategy to get the Save prompt [yes]-[no]-[cancel]. Only then is the strategy saved. Pressing Ctrl+S is not good enough.

    09/06/2017 at 11:33 AM #45606
    Despair

    It now happened 3 times to me that PRT crashed (became unresponsive) when I exported a strategy. When I then stop PRT and restart it, the exported strategy is completely gone from PRT but (luckily) is on my HDD. So PRT crashes obviously after exporting.

    I pointed out this problem to PRT and they responded that if I have problems exporting it must be my firewall settings what I think is BS. Most often exporting works fine but sometimes it freezes (just happened to me last night).

    PRT also seems to make mistakes when you open a strategy to modify it during you run a backtest of the same strategy – then all gets screwed up. Obviously PRT is then updating the edited version but still has the old version in memory on which it runs the backtest. Results seem to be chaotic and hard to predict. 🙂

    Anyway for the issue of long backtests, I just want to point out that it also freaked me out in the beginning when I tested mostly intraday strategies on low TF. If you run strategies on daily TF it is much better. Another option that makes the backtest duration more easy to handle is to separate the long and short side and test them separately.

    09/06/2017 at 1:34 PM #45625
    Andre Vorster

    @Despair

    A Firewall has nothing to do with files exported/copied/written on the same system. That is such a lame answer. Like you say BS answer.!

    I have support tickets open with IG/PRT regarding PRT that stops responding. I can’t get an answer that explains the issue nor resolve it.

    Since it is not easy to replicate the issue, I offered to help test PRT software and have an application logging program that will log what happens to PRT every millisecond to help catch the bug but my assistance was not accepted.

    Anybody else found different results running the same strategy back-test on Demo and Live accounts?

     

    09/06/2017 at 3:57 PM #45646
    Despair

    I experience definitely differences between live and backtest. Demo I’m not sure (haven’t used it so much; started using it more recently).

    But I think little differences are unavoidable. For example in backtest your order will always be filled on touch, live this doesn’t happen so often. Live the price often really has to penetrate your limit and traded beyond it to get a fill. I even saw live that price exceeded my limit and the order was not executed anyway.

    Then it happened to me several times especially with BTC (therefore I stopped trading BTC) that i got a fill at a price that was according to PRT never traded. For example price traded between 100 and 105 during one candle. Then I receive a fill on this candle for say 98. 🙂

    09/06/2017 at 7:55 PM #45661
    Derek

    You can adjust slippage levels in PRT backtests to account for the overall effect in real trading conditions. Of course, that is even more helpful if you think the deck is unfairly stacked against you?

    09/07/2017 at 10:23 PM #45756
    Despair

    And how would you do this – adjusting slippage levels in backtests? I’m afraid I don’t understand.

    09/09/2017 at 12:03 PM #45836
    Derek

    It’s just an increased spread for the backtest.

    If you want to estimate expected slippage levels I suggest dividing the current average volume by the highest volume that ever occured. The lower the value the more slippage is expected to occur. Then you adjust your order prices dynamically by adding (long) or substracting (short) a few points to your orders in the backtest.

    You can also include events: if a volume spike occurs the spread is increased x minutes before and after the spike during the backtest.

     

    Edit: it may be easier to restrict trading to certain trading hours.

    Edit2: essentially it’s substracting only the downside of increased noise levels to your backtest results. My initial thought was to just increase the average slippage but the IG spreads are so volatile these days, this needed a better answer. We can build this for robustness testing?

    09/09/2017 at 12:21 PM #45840
    Andre Vorster

    When I work on a strategy I copy&paste the code to a text editor or MSWord after editing. These files I keep as a backup.
    All these I then make backup copy to another drive and keep at a different location. Just in case one loses all due to fire/theft/flood/computer crash, etc.

    The code is worth Gold.! Literally. Worth making and having seperate backups 🙂

    09/09/2017 at 3:02 PM #45846
    Despair

    Yes, I also made it a habit to always export my work before I exit.

    09/09/2017 at 5:03 PM #45847
    Inertia

    Got scared by this thread gents so I just did what is an habit for you : word file saved in a different Cloud…

    Feel reassured… Thx ;).

    09/12/2017 at 12:50 PM #46011
    AutoStrategist

    As @Andre does I copy and paste my strategies into MS Word which are then stored on disc and cloud backup. I have lost work a couple of times when I got the message that I had reached the limit for the number of strategies (100?), even though I went and deleted strategies I no longer needed the current strategy was lost on exit.

    Regarding differences between demo and live, I don’t use demo but have noticed differences with back tests across 2 live accounts (1 for manual and 1 for auto). They never seem too significant and you have to remember if you are not running them at the exact same time then the amount of back testing data will be different and could account for the difference.

    1 user thanked author for this post.
    avatar Andre Vorster
  • Author
    Posts
Viewing 15 posts - 1 through 15 (of 15 total)

Find exclusive trading pro-tools on ProRealCode Marketplace

Similar topics:
is it possible to open a trade manually and then have it close automatically?
24 june 24 – cannot login to prt
6 jan 24 prt not available – too long to respond
a qustion about the median line in prt bands
again moneymanagement calculate
candle stick what price is it based on?
cannot get my interactive brokers account back after cancelling prt
code to find when prt bands short term line changes colour
conversion from trading view to prt : fibonacci golden wave indicator
conversion from trading view to prt :buy and sell alerts using vwap

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