Market Delta - Transitioning to the New Platform

Discussion in 'Trading Software' started by bolter, Sep 16, 2006.

  1. alann

    alann

    5P,

    As I said, it is Trevor's business to do with as he sees fit.

    My personal view is that the new MD was launched without being thoroughly tested beforehand. If comprehensive tesing had been carried out, there would not be the adverse posts that have appeared here and elsewhere.
     
    #151     Sep 30, 2006
  2. bolter

    bolter

    5Pillars,
    Actually it might be helpful to Chad if you could do some analysis and provide him with some feedback. I have done so for several Asian and European markets using eSignal and given my findings to Chad.

    Here's some simple tests to consider:

    1. Open a 1 min FP from the old platform alongside a 1 min FP from the new platform. The only differences should be slight timing discrepancies.

    2. Open a T&S window alongside a 1 tick FP chart. You might like to filter volume to say > 100 lots on busy markets. Do this for both the old and new platforms. Obviously any differences at this level will be greatly ampilified in higher time-frame FP charts.

    3. On the new platform create a 5 min VB chart of delta values. Export the indicator values to excel. Scroll back thru a 5min FP and compare the total column deltas with VB values. They should be equal.

    4. For a full session, on the new platform. create a chart and add the VB indicator of accmulated delta values. Compare this to the MDI on the old platform, both its shape and it's final value.

    Obviously you can export any/all of this data and conduct more detailed and longer-term testing in excel. This is what've done. Chad uses the S&P emini market for testing so it would add value to test some different markets - esp international.

    FWIW, the only consistency I have found is between T&S (new platform) and the FP from the old platform.

    Good luck,
    bolter
     
    #152     Oct 1, 2006
  3. mindfull

    mindfull

    Ouch, I should have checked out this thread before putting up my 100 bucks. I got this message that this was the last month to get introductory pricing and went for it. The I find out introductory pricing is only for September. My own greed I guess.

    I have used IRT before and you can spend a lot of time taking care of data. This is not fun ....
     
    #153     Oct 1, 2006
  4. Sounds good...I will do some of this testing Monday.
     
    #154     Oct 1, 2006
  5. Sorry but the world of trading software is almost never perfection in the beginning (unfortunately), so we as traders have to be able to deal with these type of issues.

    I'll give you my list of imperfect software that I have used over the years at different stages;

    tradestation

    esignal

    neoticker

    nexgen

    ensign

    qcharts

    ....all of the above products I had to have direct contact with the companies to have problems I found corrected.....some took a few weeks and some took months. The only companies software that I have ever used that I never talked to anyone to have a problem corrected was kwikpop (LOL....they must be the Lexus of charting). Now I wont even get into trade platform software...LOL (where your $$$ are actually on the line).

    So far personally, I have found one problem with Inv R/T that I have already talked to Chad about and he is already familiar with the problem (VB when used in the "accumulating" delta bar mode). I am confident that in time he will correct this issue and I do feel the Linnsoft group are a good bunch (the type that listen to their customers feedback and make adjustments as needed).

    I think that Trevor had a good idea with the merged product working with Inv R/T and I feel that everything will get perfected going forward.
     
    #155     Oct 1, 2006
  6. alann

    alann

    5p,

    I'm sure that some of the software that you mentioned may have had teething problems at the outset. Bear in mind that some of these programs have been around for some years and the testing and debugging process was a lot slower then.

    As an example, a colleague of mine runs a very detailed data mining program, for stock picks, that, 5 years ago, used to take 10 hours to complete it's cycle. That same process now takes 20 minutes. Perhaps, the anticipation of myself, and others, in expecting a seamless transition to the new MD may have been too optimistic.
     
    #156     Oct 2, 2006
  7. I've been testing out MD for the last week. New to both programs, I've found the I R/T part very unintuitive and frustrating and the MD part erratic. Plus the data downloading processes seemed absurdly complicated. However, the potential of the program is clearly very strong.

    Friday I emailed Trevor about serious discrepancies I was seeing in stock footprint charts. Yesterday he told me to try downloading a new release (8.5.5). I did so, and those discrepancies have disappeared. Mind you, I haven't done the checks bolter recommends, which I will do as soon as I get a chance. The data downloading process in the new release has also been much improved and made more bullet-proof (e.g., you have to want to override your tick data to do so).
    If you haven't already done so, I'd recommend getting the new version.

    From Trevor's responses to my questions and problems, I'd guess that all the fundamental issues will be ironed out quickly. But it sure is frustrating doing beta testing when you haven't volunteered for it.

    palinuro
     
    #157     Oct 2, 2006
  8. Everything is working good for me today....no problems that I can find.
     
    #158     Oct 2, 2006
  9. We expect software to work when we sign up for new services and sometimes it does. I am in agreement with you and others that sometimes testing periods are probably not adequate for new services, so I do hope the problems you and others have had will be resolved ASAP! :)
     
    #159     Oct 2, 2006
  10. bolter

    bolter

    Chad is doing a great job getting the delta inconsistencies resolved. The Volume Breakdown now agrees with the numbers from the old platform. He's now looking at the footprints as we speak. Hopefully we'll have this sorted in the next day or two.

    Full credit to Chad for his excellent response.

    bolter
     
    #160     Oct 3, 2006