Is Waldo's still compatible with todays chat logs? If not, what are the current compatible parsers everybody is using to parse melee/spell data? I've seen a few mentioned on VN, data output from Waldo & Silicon Dragon didn't seem to correlate or match, Tornalis wouldn't DL correctly from Gamefront, and Nerfs didn't have an option to see data for each opponent.

 

-----signature-----

Did you find a parser that does what you want?

 

-----signature-----

Waldo's does everything i need function wise, i just don't think its parsing the current chat logs correctly. Haven't found any other parsers that give me the data im looking for and can read current logs.

 

-----signature-----

If you just need combat statistics, I could add a chat parser that does that to Mojo. But I probably don't want to bother making it update automatically second by second like Waldo. It will parse a complete file when you click a button.

 

-----signature-----

Wasn't needing sec by sec updates, i just need to have confidence the parser is spiting out good melee/spell data (both mine & enemies/mobs) from current chat logs. Hopefully it also can parse/display data from individual enemies/mobs. I would appreciate it very much if your parser was able to do these things. Send me a link so I can DL it, thanks.

 

-----signature-----

Okay, I'll try to post a new version of the program that contains a parser in the next day or two.

 

-----signature-----

I'm having trouble figuring out how the parser should interpret certain messages about styled attacks. Can anybody help? The following examples were logged while fighting a test mob on Pendragon.

 

-----signature-----

The number of misses should not be able to exceed the number of total attacks.

My theory of styles is this:
1. An attack is only "unstyled" if no styles are queued.
2. If one style is queued, it either succeeds or fails based on its prerequisite. No style bonus damage is applied on failure, though the to-hit bonus may still be in effect.
3. If two styles are queued, the prerequisite for the first is evaluated. If that is not met, the backup style is attempted according to rule #2 above.

Note that there are a couple of ideas in there that I only have anecdotal support for, but that can be tested and verified if you have the time and interest:
1. To-hit bonus may still be applied on "failed" styles
2. The back-up style is always attempted when the prerequisite is not met for the primary style. (I.E., you will see a failure message for the backup style if neither style meets the prerequisites -- it is not an "unstyled" attack.)

My anecdotal evidence for to-hit working on failed styles is PvE-related. I used to farm the Tartaros' Gift mob with my hero for torches, and it seemed that spamming Annihilation improved either my chance to hit, damage, or both (I can't remember for certain) over using anytime or front positional styles.

To answer your questions 1, 2, and 3:
1. It is impossible to know which style was attempted from this fragment of log. You would have to know whether the prereq's for Brutalize had been met or not.
2. Same here. You do not have enough information from the log alone to know what style was attempted.
3. It looks to me like Cross Slash was ignored, because it was the primary style but your prerequisite was not met. Thus Side Slicer was attempted, and hit, but for no bonus damage because the style failed. To-hit bonus may have been in effect, however, so this should not be treated as an unstyled swing.

Figuring out what to do with the "unattributable" misses is tough. For someone doing a controlled test, the lesson here is that you should not queue styles. For a real combat log, there is no way to use those data points that doesn't distort the results one way or another. Assigning the miss to both styles will underestimate one style's to-hit bonus; assigning it to neither will inflate one style's to-hit bonus. One way to do it without distortion would be to ignore all queued attacks for per-style hit/miss calculations. Queued attacks could still be counted in global hit/miss data, and for landed style bonus damage, just not for per-style hit/miss numbers.

 

-----signature-----

Nothing beats Excel (or your choice of alternative spreadsheet app) for parsing, but you have to know your way around it, obviously.

 

-----signature-----

Asilithiel, thanks very much for your comments. They are extremely helpful. I'll have more to say in a day or two after I've done additional testing and the first version of the parser is working.

 

-----signature-----