[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: AW: To MarkBrown: Real-time datafeeds in Python



PureBytes Links

Trading Reference Links

> ....and then you discovered Wealth-Lab Developer 3.0 ... :)
> With all respect most of it is taken care of in WLD3. 

Volker, while WL has some good features, and WLD3 has some huge 
improvements over WLD2.1, it is hardly the perfect platform.  
Some weaknesses include:

* Nothing is updated until the close of a bar.  This is a 
critical weakness for many discretionary traders.  (I know the 
whole argument about "systems can't react until the close of a 
bar."  TS does the same thing, but that doesn't prevent it from 
updating indicator calculations in realtime.)

* Optimization and portfolio testing use different interfaces.  
WLD3 is a major improvement over WLD2.1 in this respect, but it 
is still possible to get messed up with different results on your 
optimization vs. your portfolio test.

* I find the WLD model of using series (which is fairly 
pervasive) to be obtuse and a pain to deal with.  As an example:  
there's no good way (that I know of) to reference historic values 
of an arbitrary variable, so unless you want to save the previous 
bar's value, the only way to do do a "crosses over/under" 
operator is to use the CrossOver/Under functions.  But in order 
to use those, you have to have built the entire series before you 
first reference it -- which means you either use the series 
operators beforehand, or you need a separate loop to build the 
series before you reference it.  It's certainly workable but I 
personally find it very cumbersome.

* Similar things could be said about many GUI design issues in 
WLD.  It works, but I think (given 25 years of software and GUI 
design experience) many areas are badly designed.

WLD works, and many people love it.  But it drives me crazy every 
time I use it.  Generally I pick it up to test out some kind of 
portfolio idea that's difficult to test in TS.  After a while 
remembering how the WLD language & order model works, I get 
things working.  But even then I get so frustrated and ticked off 
at WL that I drop it as soon as I finish the experiment, even 
though it does some things much better than TS.  I think that's 
telling.  I'm not so enamored with TS that I would stick to it if 
there was a better answer.  I just haven't seen one yet.

Gary