esignal's backtesting capability

Discussion in 'Strategy Building' started by frank123, Aug 4, 2003.

  1. frank123

    frank123

    I began to use Esignal's backtesting capability recently and feel it is rather user friendly. The language it is using is rather self-explanatory. I especially like its formula wizard, which speeds up the process a lot.

    I understand most of the backtesting job is done using Tradestation in the past and most books about strategy trading is in EZlanguage, so there is certainly advantages in using TS. However, TS’s subscription is twice as expensive as Esignal.

    I am wondering if anyone who have both used Esignal’s and Tradestation’s backtesting function can comment on the comparative advantages and disadvantages of the two.

    I started this thread to see if I should switch to TS in the future for backtesting.

    Thanks,

    Frank
     
  2. frank123

    frank123

    come on, any one?

    Is it because Esignal's backtesting function is so new that very few people have used it and TS both?

    To Tradestation people: why do not you convince me TS is better than Esignal here on a forum sponsor by you?!

    To Esignal people: Come on, make a case for yourself of why I should stick with you?!
     
  3. frank123,

    Saw in the September "ACTIVE TRADER" that you can get esignal for "free" if you open a MB Trading account.

    I don't know much about MB.

    Fine print:

    Have to do 50 round trip trades/month to get a $100 credit and you have to pay exchange fees.

    DS
     
  4. It basically sucks. In my opinion.

    But... considering the software comes "free" with the E-signal, I think it's good.

    It's not a tool for system traders.
     
  5. maxpi

    maxpi

    Could you elaborate? I just switched from TS2000 to Esignal. Esignal looks much better in many regards. It has global variables, a formula wizard, good data is included, and the EFS language seems well documented and very easy to use.
     
  6. I've been with esignal for about 6 months, and tried TS for a month. I really think esignal is better in every way except it's EFS programming language.

    I'm not a programmer, yet with easy language and their great online manual, I was able to put together a few studies in minutes. Esignals scripting language on the other hand is impossibly difficult. I'm sure it's easy enough for a programmer, but I'm not a programmer. I dearly wish esignal would come up with something FAR more user friendly in that respect.

    Here's an example: in Easy language you can use a simple, one line builtin command to reference a a different time frame. In esignal, it takes dozens of lines of code that makes no sense to me whatever. As it is, for me EFS script is unusable.
     
  7. I agree on the EFS issue, Bundle.

    I believe eSignal is equal to, or slightly better, than TS in most regards, EXCEPT for their EFS scripting language. I'm not a programmer, but I found it relatively easy to create custom indicators and strategies in TS. It's a pain in the posterior to program equivalent code in eSignal. For one example, try a multiple time-frame study in eSignal and in TS and you'll see what I mean.

    If TS had not abandoned their software strategy to focus on being a mediocre brokerage I would still be with them. And, yes, I know I can subscribe to TS without being a customer, but it's much more expensive than eSignal and other charting packages. Besides, I have my old, trusty (sort-of) TS2000i to do any back-testing I need.
     
  8. maxpi

    maxpi

    And then when you use data 2 in a chart in Tradestation to reference the other timeframe 1) it can't be tick data 2) so many weird bugs will pop up you will give up eventually. I had to resort to building my own longer timeframe data in the arrays with tradestation and then testing on that.
     
  9. Could you be more specific with the weird bugs you saw? I never saw anything that couldn't be explained by the way TS handles orders at the end of the bar.
     
  10. maxpi

    maxpi

    I just saw so many posts on the forums regarding backtesting problems with more than one data in a chart that I never went that way after I saw a couple myself. It was about 3 years back, I don't remember the details but I do remember the frustration. I assumed that TS just never debugged that part of the code that well.
     
    #10     Aug 5, 2003