Here is my bottom line. I got my 10.6 trading screen fully replicated and the 10.6 efs files fixed to work in 11 (removed playSound). Everything works and starts up on reload. The time scale adjustments are not restored on reload. And there is something weird going on with autoscaling on one of my charts. And I still hate that I can't squeeze the vertical dimension of a study pane down far enough. And to backtest the strategy needs to be on the chart. Otherwise I am only semi-dissatisfied. We'll see how it does when the data starts screaming in at the open. Also I have a page with 23 daily charts of my portfolio running. THAT is a great feature!
I think I mentioned this before, but it is bizarre that the debugPrintln statement doesn't work correctly: Hello, name my is.
Interesting little gremlin. With both 10.6 and 11 up, I ran a backtest in 10.6 and got an XML error, which I never get unless I have run on the order of 50 tests. I killed 11 and the BT worked.
Just a guess, but it could be related to how data manager operation is improved in 11. I believe if you go through <a href="http://kb.esignalcentral.com/display/2/index.aspx?c=12&cpc=40O2r6eeyG0Jfuy2sd5gAr4LG2c3&cid=1&cat=&catURL=&r=0.773735404014587">this setup</a> there might be an improvement to 10.6's data stream. I wouldn't recommend doing it while the market's open -- I could never get it to work myself, and eSignal wouldn't operate if I ran as admin (my UAC is on). Also, maybe Vista/W7's visual display settings could be set to basic if you have a feeling 10.6 is a little slower. Just throwing ideas out there while sitting in a trade. On reflection, I think I was a bit uncharitable in my critique of eSignal, because my life basically depends on their product and I've always appreciated how it looks and operates (save when their servers couldn't handle the data back in... 2005 or so) . I don't think I'm as sharp as you though, and I have a feeling you have a symphony going whereas I've got this one fatass on a rickety chair playing a jaw harp, so these micro changes in appearance are more untenable for your well-tuned set. I'll refrain from venting, although I'm committed to learning NT at the same time now. On a side note, I ALWAYS get XML errors when backtesting. After about 5 runs, the Strategy Analyzer always fails. This is likely due to said 600-line fatass being too much to handle. Plus, recently there have been periods in the backtest where I believe the script stopped running altogether (one bizarre trade spanning 3/20/2010 - 5/21/2010, despite plenty of logic to cover). I'd like to see how it runs in 11, but the core javascript has seriously changed.
Thanks for the link, but it just took me to the version 11 help page. Would you tell me which article to look at, please? I had to kill 11 this morning so I could run a quickie backtest on today's conditions to see if I should trade the morning session. So I did not have a chance to compare the two versions with faster tick flow. Will do that this afternoon during the doldrums. I am ready to deal with it because I just went shooting. Uncharitable? I don't think so. You are the CUSTOMER! As to my skill, I am just an old fool who has written code to see things that aren't there. Boojums and haints. Thanks for sharing your XML error experiences. It must be your monster code, because my longest strategy code is 115 lines long in a not very compacted, readable form. I suppose we could ask customer support questions like that, but I long ago quit calling or writing them because the usual response was "Huh?" I learned to figure it out by myself, and how to work around whatever the probelem was.
The two versions are ticking together to the second and to the tick. The more I use it the more aggravated I get that so many simple things require two clicks in 11 that only took one in 10.6. Cursor on/off, autosizing, data window on/off, etc. 11 desperately needs a toolbar.
I am thinking that if they fix all the shit we've mentioned here that I can live with it, especially if they implement reliable autotrading. Having NT's automated backtesting would be nice, but it might remove some of the insight I gain now doing it the slow way. So code inertia wins. But I'll have to lay in a lifetime supply of barf bags for the puky layout.
Arthur and Samsara: We appreciate your efforts to try v11. If either of you are interested in testing new versions of 11 before they are made public, please PM me and provide your email address. You both are obviously well-versed in the use of our product and while you may not have time to give feedback nor want to be "beta-testers", we want to at least present the opportunity to do so directly with us. Thanks.
Why, that is a very kind offer! How much does it pay? I have loved/hated your product through innumerable versions. But I fear that at my advanced age, death no longer being optional, I may die before it does what I need it to do. And in retrospect, I think some of my bitchiong was overblown, perhaps akin to what my horror would be if my beloved wife, whom I married when I started with you, got cosmetic surgery and came out looking like Joan Rivers.
I forgot to mention a couple of things. Whatever the fuck the source of that XML error is, how soon it occurs depends on how many samples there are in the BT. When I run 120 days of 390 minutes, it crashes in maybe 50-60 tests. A stupid suggestion. Perhaps you have code that exits the trade at the close, and on 3/20/2010 in your data set the close was earlier, or missing? I exit a trade at 12:58 PT if not stopped out, and as I don't trade short sessions (holidays), I have code to exclude those days so the trade won't run into multiple days.