Well you don't want to analyze the entire log because you may have data from a previous mission with old or different equipment. You wouldn't want that to skew the results.
I don't want to limit the size because someone may have a 20 MB log that is really an entire match. I know CCA was 7 MB for me. Maybe a No Win is over 10 MB.
From what I can tell, other readers try to break up the log into matches by using a certain number of minutes to split it. So let's say it splits the matches if there is a 2 minute break in the logs. One entry is 11:24 and the next one is 11:26, it'll consider them separate matches. I really didn't want to do that extra work lol also I don't feel it's accurate.
The last 5 minutes (and yes it really does check only the last 5) was most likely the majority or half of an entire match. Think how long a CCA actually lasts, or ISN/ISA.
Another option is before uploading, the user could choose the maximum number of minutes to read. Maybe between 5 and 20 minutes with 5 minute intervals (5, 10, 15, 20). I could even go up to 30 minutes, or after 20 juts have an option for the whole log.
I am by all means quite naive when it comes to things like this, but my only thought is if there could be a keybind put in to tell the log that a match has ended and a new one has started, would that be feasible?
(12-31-2014, 04:33 PM)Attilio Wrote: Another option is before uploading, the user could choose the maximum number of minutes to read. Maybe between 5 and 20 minutes with 5 minute intervals (5, 10, 15, 20). I could even go up to 30 minutes, or after 20 juts have an option for the whole log.
That would be interesting.
(12-31-2014, 06:23 PM)Jstagg Wrote: I am by all means quite naive when it comes to things like this, but my only thought is if there could be a keybind put in to tell the log that a match has ended and a new one has started, would that be feasible?
I have no idea lol the command to start/stop record is /combatlog 1 and /combatlog 0 but I don't know if anything else can be added
Ok, tested it again and increasing time would surely be helpful, at least for comparison.
I'm using that German CombatLog reader that's built in Java and difference in DPS is almost 7k, program showed 20k, site 13k DPS.
I often use Plasma Lance as one of first moves against Tac Cube in ISA (with all buffs that I have), so it could be that it's not read by analyzer on site, but can't be sure until you "fix" it and give us option for bigger time.
Looking forward to further testing!
01-03-2015, 10:39 PM
(This post was last modified: 01-03-2015, 10:45 PM by Attilio.)
Wow I thought I uploaded this and made a post about it, I guess I got distracted and forgot lol
Anyway before I overwrite the current code, go here and let me know your results http://www.stoacademy.com/tools/combatlog_new.php
Basically in this new one, it calculates each player's combat time and then does the "per second" math based on that. This way it's specific to the player instead of having one global time (e.g. 5 or 20 minutes) and using that for all players. This new method should be more accurate... but then again I always sucked in math lol
It also reads the entire log, which means if you have a combat log with an hour worth of data, it'll do the math for everything. I could add a time limit as mentioned earlier. When the user uploads they can choose 5, 10, 15, or 20 minutes, but it would still be player specific.
EDIT: If you tried it before this edit, give it another shot. I forgot to make one tiny change before uploading. Should be fine now.
01-04-2015, 02:28 AM
(This post was last modified: 01-04-2015, 02:32 AM by Xhopeon.)
I've tested it and found same result, but also I've found reason why it does that
If looked at first table of results (from site) match lasted for 333 seconds and my DPS was 13,281.32 (total of 4.422.679,56 damage)
Now when looked into second table (program), for match that lasted 216 seconds IN BATTLE
, DPS is 20.475,36 (total of aprox 4.422.678 damage)
So to conclude, there's only difference of "battle time" and "total time", which is not that bad at all, this way it counts effectiveness of individual for entire match, not just during time person fought (IMHO, DPS cheating because of waiting for buffs to cool-down).
^^EVERYTHING ABOVE IS TESTING OF NEW ONE!^^
P.S. ok, now I saw that you posted about combatlog_new.php so I've tested to see difference between old and new one.
New one is more accurate/close to results of program. I'm not talking about DPS but damage in total (old one gave 400k less) and all other stats.
And I love the detail of "total match time", it's good to know real time of battle!
Definitely an improvement
Cool, I just want to make sure I understand because it sounds like you came to the same results as me.
The combatlog_new.php has a different amount of combat time when compared to the program, but the damage is accurate. The DPS is different because of the obvious combat time difference.
I just started playing around with it and I believe I may have figured out why there is a discrepancy. I'll work on combatlog_new.php and reply when it's updated.
01-04-2015, 04:25 AM
(This post was last modified: 01-04-2015, 04:28 AM by Attilio.)
Ok I'm at a loss to explain it. TP I NEED YOUR HELP!!!! You're my math-a-magician lol or if someone else can help that's good too. Take this
log which only holds my combat entries from start to finish. The first column is the time in the form year : month : date : hour : minute : second . tenth of a second
You could do simple math by subtracting the start and end time, but that doesn't give you an accurate "in combat time" because it won't account for travel times between enemies, dying/respawning/traveling, etc.
According to the CombatLogReader, I was in combat for 252 seconds. No matter what I do, I can't calculate that same result. If anyone has a clue please let me know. I'm going to be now but I'll try to work on it some more tomorrow.
(01-04-2015, 04:25 AM)Attilio Wrote: Ok I'm at a loss to explain it. TP I NEED YOUR HELP!!!! You're my math-a-magician lol or if someone else can help that's good too. Take this log which only holds my combat entries from start to finish. The first column is the time in the form year : month : date : hour : minute : second . tenth of a second
Boy, that's a tough one. Neither Google Spreadsheet nor MS Excel seem to like that format for a date-time format at all. So I wonder, how you can calculate with them at all. Does the Reader convert those values into a usable number? What is your current approach?
Would be nice to know. Just so that I do not repeat it accidentally.
Otherwise, assuming that you are breaking down those values and put them back together to have an all-second value, then I guess it's plain simple
seconds + (minutes*60) + (hours*3600) + (days*86400)
However, if combat passes midnight and the next day is the first day of the following month, then this formula won't work. LOL.
But I guess, it is not that simple, right?