Henry is the resident PP+ expert... maybe he has seen this error before. I have heard of several instances with people having trouble with PP+ MSX errors, but this is the first I have heard of this particular error. Does the error only happen for one chart, or does it happen on many different charts? In the exploration, can you isolate the chart(s) that cause the error? Does the error happen consistently i.e. can you replicate it, or is it just one of those random event things? What other information can you provide that might help with the bug-hunt? wabbit :D

When you run the exploration, please verify that you DO NOT have "load minimum bars" checked. Load say 500 bars and try the scan again. wabbit :D

The answer to that question comes down to whether you want accurate results or a fast scan. There is no point doing a scan of 2 bars per stock if the results are not accurate! There have been many, many, many posts on this forum about the number of bars to use in an exploration, or, to be more precise, how using minimum bars stuffs up the results from explorations. What do you think is better: a scan that takes two minutes to run with results that are the same in the explorer and on the charts, or the scan that takes 20 seconds and the results are all wrong? Please don't skimp on the data just to save a few seconds or minutes. If you are using EOD data, you have all night to do your analysis. If you are using intraday data, your scans will be running continuously anyway. The general rule-of-thumb is to load at least 5 times the maximum data length i.e. if your scan uses a 100 day moving average then load 500 bars. As you're using a password protected scan (I assume) then you will not know the inner workings of the code, and hence should load at least 500 bars, just to be safe. If you find the scan is just taking too long, but the results are correct, try loading 450 bars, then maybe 400 bars, then maybe 350 bars. Find where the limit is between accuracy and speed. Personally, I go for accuracy EVERY time, otherwise the only thing that is going to happen quickly is me losing my money. Hope this helps. wabbit :D

Again, the problem could be caused by how much data has been loaded on the chart, or how much data exists for the chart. If the exploration is working well with 500 bars of data loaded, how many bars of data are you loading on the chart? Is the MSX error generated on any particular chart, or does it happen on all charts? Ifso, whicj charts, so we can try an replicate the error ourselves (in fact I am hoping Henry can, because I don't use PP!) The charts where the error is generated, are they new companies with less than, say, 500 bars of data? Can you see where I am going with this? Personally, I think all experts, indicators, systems and explorations that are "data sensitive" need to have the exceptions handled. Poor testing by the authors means that these errors are left un-handled, when they should be throwing caught exceptions. Its simple to check how many bars of data are loaded (lastvalue(cum(1)) and compare this with the number of bars required: if(lastvalue(cum(1)<100,error, no error) etc. wabbit :D

you will need at least 1000 bars for the add-on to work properly. What time frame are you exploring? how many bars are loaded? write down tha exact error message!

Yes, I see your problem. I can create indicators with the ext function time capsules, plot them on a chart, but once I run an exploration I get the error message, and it then causes all of the other external functions to go out of whack because of the dll now contains errors and will not plot. I think that equis needs to find a fix In the mean time, use time data create, with the offset to create a replacement formula

Both Equis and Sandon Radkowsky "who wrote the dll" should have tested all of the indicators , experts, systems and explorations to make sure that they work. Tfade is one of the packaged systems and should work in all phases of Metastock, otherwise don't include it!

Khandelia,

You've got a few posts on the forum asking about T-Fade and the PP+ addon. I tripped over some code in my archives (from 2007) including what I think might be the original(?) code for T-Fade. I have no idea where it came from, but if I run the code side-by-side with the "proper" version, the results are identical.

The original code was very poorly written, so I cleaned it up to be able to see more clearly what is going on and where errors might be caused, in short, the "smart" part of the code goes like this:


The rest of the expert advisor commentary, highlights and symbols deal with the signals generated by 'tr' and 'season'; the stop values are given by 'hp' and 'lp'

If you've got more questions on how these functions, ask Henry as he's the resident "expert"




wabbit [:D]