Hi,
I just downloaded the new release:
I have the following problem(s)
- when trying it out with the simulated trading account instead of a
real account,
I sometimes get "invalid" data. This hardly occurs with the real
trading account
(Some brief fluctuations of invalid sometimes)
From my understanding this implies that it is not possible to
"forward test" a strategy except if
I am really trading it?
- perhaps even more severe: I introduced a dummy strategy to trade EUR
(it was more about testing
whether the framework works than about a real strategy.
Independent of simulated or real trading, within seconds I get
"invalid" and it stays this way.
Perhaps the analysis for "valid" data is too rigid for some
instruments?
Klaus
On 29 Jul., 00:32, nonlinear5 <[email protected]> wrote:
> Latest release
> download:http://jbooktrader.googlecode.com/files/JBookTrader-7.02.zip
> Release Notes:http://code.google.com/p/jbooktrader/wiki/ReleaseNotes
> User Manual:http://docs.google.com/View?id=dfzgvqp4_10gb63b8hg
> Sample market data (based on the 10-level
> depth):http://jbooktrader.googlecode.com/files/ES-sample.zip
>
> There are substantial changes in this release:
>
> 1. Market depth validation is stricter and more consistent.
> Specifically, market depth
> would be considered valid if and only if all four of the following
> conditions are true:
> a. The number of bid levels equals the number of ask levels and is non-
> zero
> b. The bid price of level N is smaller than the bid price of level N-1
> for all levels
> c. The ask price of level N is greater than the ask price of level N-1
> for all levels
> d. The best bid price (at level 0) is smaller than the best ask price
> (at level 0)
> If market depth is detected as invalid, the last known valid balance
> would be used to update the indicators. If market depth stays invalid
> for more than 60 seconds, the corresponding message will appear in the
> event report.
>
> 2. Fixed the methodology of dividing the parameter space in the
> "divide and conquer" optimizer. It was not working as intended. The
> D&C optimizer is now more focused.
>
> 3. Added the "divide and conquer coverage" preference, which controls
> the density of the search.
> The denser the coverage, the slower optimization is, but also more
> likely to find the top regions
> in the search space. This option is useful when working with the data
> sets of varying sizes
> and running optimizations on machine with varying CPU power. On fast
> machines and with small data sets, the "coverage" preference may be
> set to "dense", while on slow machines and with large data sets, it
> may be set to "sparse".
>
> 4. Added the "brute force strategies per processor" preference. This
> does not affect the results, but it may make optimizer run faster.
>
> 5. Both BF and D&C optimizers are faster
>
> 6. Web console more dynamic, all information is refreshed every 5
> seconds
>
> 7. A range of dates can be specified for both back testing and
> optimization
>
> 8. Fixed the scroll bar and zooming issues in performance chart
>
> 9. Added "user manual" and "release notes" to main menu. The user
> manual is no longer distributed in the /docs directory, but is
> reachable online from the JBT "Help" menu.
>
> 10. Better substance L&F (frame decorators and status bar changed)
>
> 11. User manual is updated to reflect the recent changes
>
> Many thanks again to Martin for the numerous contributions to this
> release.
>
> As always, I am asking everyone to test a lot and test with prejudice.
> That is, even if you only suspect something may not be right, please
> do not hesitate to mention it.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"JBookTrader" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/jbooktrader?hl=en
-~----------~----~----~----~------~----~------~--~---
