WLP Crashing While Running Optimizations
Author: sdbens20
Creation Date: 6/21/2011 7:49 PM
profile picture

sdbens20

#1
WLP (both the newest prior versions) is crashing after 5-6 optimization runs. This is happening on both my 32 bit Vista Pro laptop and my Windows 7 Pro 64 bit desktop.

I'm running manual Monte Carlo and genetic optimizations using simple 4 rule strategies. After 5-6 runs on one symbol at a time, WLP crashes and I have to restart the program. One year's daily data is used. $25000 fixed dollars is the initial position.

Any idea why this is happening? What additional info do you need?

Thanks,

Sherm
profile picture

Eugene

#2
Since this, as well as the mysterious optimization slowdown, and the GA crash from Nov 2010 (probably couldn't be reproduced by the GA developer), seems to be coming from the same source - your computer(s) - then if you want "any idea", I'd say it's something with your configuration.
profile picture

Cone

#3
QUOTE:
What additional info do you need?

1. Error messages with any details
2. Script - the xml version attached
3. Provider, DataSet, and Scale
4. Position Sizing
4. Screen shots

Before creating a ticket with all that information, the first thing I'd do is check your data for errors. Although we haven't "caught it in the act" yet, Fidelity static data (if that's what you're using) notoriously may become corrupted during updates. Use the "Bad Data Check [Rev A.]" script in the Utilities folder, but read its Strategy Summary first to understand how to use it.
profile picture

abegy

#4
I have the same problem. I have 2 computers with different configurations (and hardware) and the problem appears.
I see that Monte Carlo crashes all the time if it's the first run after launching WLD.
Moreover, I see that if some symbol has no data because of the data range selection, WLD crashes also.
profile picture

Cone

#5
QUOTE:
Moreover, I see that if some symbol has no data because of the data range selection, WLD crashes also.
Under what conditions, specifically? Only during MC optimizations?
profile picture

Eugene

#6
QUOTE:
I have the same problem.

Whenever I hear this phrase, I know: this will end being a different problem.
profile picture

abegy

#7
I will create this weekend a ticket with all information and data to help you to create the crash.
profile picture

hlh

#8
I second the slow down issue. I did document that already (in a ticket) by providing data showing that execution per symbol started quick and did slow down more and more. I was told this was coding error which (in this case) it clearly was not. There for sure is a slow down bug in WL. I do not agree to the tone of support staff here and specially I do not agree with the general assumption that we users are always wrong and dumb (I admit the latter is true in my case, but not for the other users - they are smart too, dear WL staff, and they trying to use WL for real trading)

I also have a problem that the very same strategy over the very same dataset takes 1 minute when started manually but between 15 and 25 minutes when startetd from Strat Monitor. This may or may not have something to do with the above issue(s). Ticket has been opened and closed by WL -> again my codes fault was stated (again not true; see also thread /Forum/Posts/Strategy-Monitor-Automation-31621).

WL6.1 as well as WL6.2 CRASHES since today, Wed. 2011-06-22, when starting BBFree data update! This is critical (I have two open tickets where this issue is mentioned).

Thank you for starting taking your customers serious!
profile picture

Eugene

#9
The BBFree provider must be updated to version 2011.07. Previous versions are not supported.

Current version updates the data and returns SM updates fine. If your experience is different with version 2011.07, then please completely delete all accumulated BBFree data and reload it from scratch to fix possible data file corruption:

"c:\Users\hlh\AppData\Roaming\Fidelity Investments\WealthLabDev\1.0.0.0\Data\B........Provider"
profile picture

sdbens20

#10
Eugene,

I am a little disappointed in the tone of your response. Also, looks like I am not the only one experiencing the slow down issue. Must be their computer, too. Months ago, I worked the with WLP support people and provided extensive information om the problem. I didn't sense a deep desire to help get it fixed. It may be that my new computer is the problem, but I'll bet that it isn't.

Cone,

I am on vacation and will provide the information you requested when I return. I'll do a data check then, too.

Sherm
profile picture

Eugene

#11
I bet that hlh's slow down issue is different. It's related to Strategy Monitor. His workflow is truly a different thing (I do have an idea but it's their private information).

You provided extensive information. Thank you. It's nobody's fault that our computers and the GA developer weren't able to reproduce the issue. Not reproduced = not fixed. It has nothing to do with intentions, so I don't think that claiming that the author of RL, NL, MCV, GA etc. who got your ticket, did not intend to fix a possible issue in his work. It just couldn't be reproduced. On the contrast, the same customer (hlh) pointed me yesterday to a crash in Bloomberg Free which I immediately was able to duplicate. This issue has been fixed within an hour; on the other hand, their slowdown issue is still being troubleshooted. As you can see, nothing personal.

P.S. I'm a little jealous since I haven't taken a vacation since my employment here in 2006 but I hope that my intentions are clearly visible with the products of my work.
profile picture

Eugene

#12
Note re: intentions. I was just able to replicate and fix a bug in Portfolio Inspector that I couldn't even reproduce before. I lost many hours on troubleshooting and even installed a different operating system just to completely reflect the customer's setup who reported it. This is a striking illustration how judging by "intentions" can fail with intermittent bugs.