 |
03/20/09, 2:43 PM
|
#1
|
|
Piston Honda
|
Hunter WWS Spreadsheet Thread
Since the days of steady shot spamming are over, analyzing our combat results do more to increase our understanding our DPS and how to improve it. The intent of this thread is to better understand and make use of these results, similar to Moonkin WWS Thread or Shadow Priest WWS Thread.
To start, in order to assist myself (and hopefully some others) in tuning performance, I have been working on a spreadsheet to breakdown WWS parses in more detail, and I now feel good enough to at least provide an initial version for general consumption. The idea here is to determine how good a hunter is doing at maximizing his/her cooldowns, keeping dots ticking and overall casting activity (including use/overuse of steady shot). Results are derived via input from WWS parses and relevant character info, and reflected in percentage values (this is not a DPS spreadsheet, it is an activity spreadsheet, and as such, it is entirely possible for a character in greens to do better than Best in Slot characters).
The spreadsheet can currently be located here:
Hunter WWS Spreadsheet
OK, now here is an example of how this works:
First, a WWS: Faerdael v Patchwerk
And a the result using the Hunter WWS Spreadsheet:
I'm not terribly impressed by my results: Some things to point out are my Explosive Shot usage, Serpent Sting uptime and Steady Shot usage. You can see that I am at 115% of ideal Steady Shot usage; this basically means I was 15% over my optimal filler cast time, and you can see that Explosive Shot usage suffered as a result (< 80%). Overusage of steady is likely a common occurance and the way I see it, I should strive to get my Explosive Shot as close to 100% as and bring down Steady Shot without letting total cast time suffer. I am clobbering my ES cd's with steadies probably more than I want. Now generally, I feel like I am pretty good about tracking my DoTs, however I never seem to reach 90% uptime using a ES > Aimed > Serpent > Steady priority and not clipping dots. This has actually currently compelled me to spec out of Improved Stings for now. It is also making me wonder if I should worry so much about not clipping dots, and play to keep serpent up all the time, but I'm not so sure.
This was a trap dance fight, which may help you understand how I ended up with only 92% casting time on Patchwork (I spec for sniper training, but had to rely on battle shout for the night, which I don't trust getting from 30+ yards (I am sure I am not the best trap dancer you will see).
Some caveats about the spreadsheet:
*If you use Chimera Shot, Serpent Sting cast time will considered one cast, and always refreshed via Chimera. If you have to re-serpent, its going to look like wasted casting time (which is kind of what it is anyways).
*DoT cast time is calculated by the number of ticks on the mob plus misses.
*Steady Shot speed is calculated differently than on Gear spreadsheet. It calculates the controllable uptime of haste buffs and seperates them, then divides out IAotH uptime among all (non)hasted states, and then averages this out. This means if you are not soft-capped for haste, you will have an averaged steady time over 1.5s where gear spreadsheet may show you at 1.4x. I feel this is a bit more accurate model; you can see this in the calculations section at the bottom of the spreadsheet.
*PTR changes should be up and working now; make sure to turn on the Kill Shot GCD option.
*Pets are not yet implemented. would like to to see this happen, but all the different specials will make it take some time.
Thanks to Esoth and others who have given me feedback on the betas of the spreadsheet.
Last edited by Faerdael : 04/14/09 at 11:21 AM.
|
|
|
|
|
03/20/09, 2:53 PM
|
#2
|
|
Piston Honda
|
How do we get the data from wws into the sheet? Manual input, or import? I can't find an import function...
|
|
|
|
03/20/09, 2:54 PM
|
#3
|
|
Bald Bull
|
I know this is a limitation of WWS, so I don't know how you'd get around it, but seeing parsed uptime on Unleashed Rage/Abom's Might/Totems/Ret Aura/Horn of Winter would be great too, to see when you hurt yourself by something being out of range. The combat log does generate events for these buffs (well, not sure about totems) fading, and other sites like WoWMetersOnline even use them to parse Uptime, but WWS for some reason filters out buff fade events.
Given that you'd actually have to parse the full log for this, it'd be more useful to just get it from a log import though, but still, it's info I'd love to see as part of this kind of analysis :S
Number of LnL procs gained would be useful to see though, and you can get that off WWS.
|
|
|
|
|
03/20/09, 3:07 PM
|
#4
|
|
Piston Honda
|
The spreadsheet requires manual input from WWS and some limited character info. About the only thing I had to look up on my toon was Haste Rating and the rest I pretty much always know. I'm not sure I'm interested in the working on an import process for haste rating and a few talent points. Importing combat logs or WWS sounds less appealing, and I've never worked with XML from inside Excel before (though not a big deal). If enough issues come across, I might consider digging into it.
As far as tracking uptime, yes, I've considered this. Frankly as far as buff uptimes go, the 3% haste and IAotH are the only ones relevant to shot usage. Unleashed Rage, for example, will affect how hard you hit, and not really how good you are doing at using your shots. Changing 3% haste buff to a percentage rather than a true/false is something I can look at. It does appear that WWS is now showing a percent value for gains, if you look at my buffs and debuffs page you will see percentages; probably to catch up because WoW Meters has been doing buffs better for a while.
|
|
|
|
|
03/20/09, 3:33 PM
|
#5
|
|
Bald Bull
|
Originally Posted by Faerdael
As far as tracking uptime, yes, I've considered this. Frankly as far as buff uptimes go, the 3% haste and IAotH are the only ones relevant to shot usage. Unleashed Rage, for example, will affect how hard you hit, and not really how good you are doing at using your shots. Changing 3% haste buff to a percentage rather than a true/false is something I can look at. It does appear that WWS is now showing a percent value for gains, if you look at my buffs and debuffs page you will see percentages; probably to catch up because WoW Meters has been doing buffs better for a while.
|
Buff uptime is relevant to how good you are at staying in range of buffers, which is why I think it belongs in this analysis, but if WWS is showing uptimes finally (awesome, thanks for pointing this out!), it's good enough.
So to use the spreadsheet (I'm at work and don't have excel handy :S) you actually have to manually enter all the hit/crit counts for your shots, one at a time? Or can you at least copy the whole attack breakdown page and have it parse that out for you?
|
|
|
|
|
03/20/09, 3:52 PM
|
#6
|
|
Bald Bull
Worgen Hunter
Whisperwind
|
A couple other caveats:
*Perhaps the biggest is that it assumes Kill Shot time is equivalent to 20% of the boss fight duration. This is not necessarily true, depending on heroism/bloodlust usage and fight mechanics, but it's the closest estimate I know of. If anyone knows if this is tracked somewhere on WWS or something, that would be great.
*I have found myself tweaking around the time of the fight by a second or two, as a fight ending near the periodic cooldown of some abilities can skew the results slightly.
For instance on this fight I had 151 seconds of DPS time, which is going to be 30.2 seconds of KS time. Technically, I could have pulled off three, but I only did two which I don't think is unrealistic. So it shows my KS time as 66%, and my aimed and explosive shots drop considerably because of this. If I cheat and change the time to 150, I get much more realistic numbers across the board.
Anyway, the main use I see from this is not solely looking at your numbers in a vacuum and wondering why it's not all 100%. Some potential uses would be:
1. Comparing people of the same spec in the raid. It's something to look at other than the final DPS result, and it's some way of comparing performance aside from gear.
2. Comparing yourself on different fights. What's my average uptime on serpent sting on boss X? Steady usage on boss Y instead of boss Z?
3. Personal performance could probably be measured better if we had a lot of data run through this and saw what kind of numbers people were actually getting. For instance, to know that 85%-90% (as a made up example) explosive usage is a realistic goal to aim for.
Edit -
@alienangel: There's really not a whole lot of data to enter. The player, glyph, and weapon settings are mostly a one-time thing, and it's just maybe 10 cells of data to enter after that.
|
|
|
|
03/20/09, 4:19 PM
|
#7
|
|
Bald Bull
|
Originally Posted by Esoth
A couple other caveats:
@alienangel: There's really not a whole lot of data to enter. The player, glyph, and weapon settings are mostly a one-time thing, and it's just maybe 10 cells of data to enter after that.
|
I wasn't objecting to entering the static things though, just the shot counts. I used to do the "load up the wws page, make it display the columns for number of crits/hits instead of just the count for one and % for other like it does by default, etc etc, see how I did compared to the other hunters" for my reports fairly regularly too, I know it doesn't take long to do, it's just a pain to do manually. Was hoping the spreadsheet above would automate that with a little html scraping ^_^
Maybe I'm just lazier than everyone else, but doing that manually or not is the difference for me between using this every raid by pasting in the WWS url, and only doing it once in a blue moon when I feel like plugging in counts into the correct cells. It looks like it'll be a great tool either way, thanks for the work, but automating the input to it at some point would make it even better.
|
|
|
|
|
03/20/09, 5:01 PM
|
#8
|
|
Piston Honda
|
I'll look into importing WWS data, however it may be slow for me to get it done, I don't know. Frankly never tried it (though I really should have) so its something good to practice. I can't promise any kind of timeline for that though.
Esoth, I see that > 150s is the threshold where you would gain one more theoretical ES, one more Aimed shot and one more KS. Thats why taking 1 second off adds 5% aimed usage and 2.something% ES usage. If you were to move up to 152s for example, nothing would change save for auto and steady usage, and changing the Kill Shot usage is not going to alter any other usage unless you put it on the GCD.
This is conceptually accurate, save its not technically possible to add 1 more Aimed plus one more ES in one second. It doesn't do any type of GCD collision management, which is another limitation here. Heroism at the end of the fight I'm not sure I'll be able to deal with, perhaps make it an option under combat settings, I don't know.
|
|
|
|
|
03/20/09, 5:03 PM
|
#9
|
|
Bald Bull
Worgen Hunter
Whisperwind
|
Originally Posted by Faerdael
Esoth, I see that > 150s is the threshold where you would gain one more theoretical ES, one more Aimed shot and one more KS. Thats why taking 1 second off adds 5% aimed usage and 2.something% ES usage. If you were to move up to 152s for example, nothing would change save for auto and steady usage, and changing the Kill Shot usage is not going to alter any other usage unless you put it on the GCD.
|
Right, that's what I was trying to say about it being around times when all cooldowns are lining up, I just didn't express it very well.
|
|
|
|
03/23/09, 10:31 AM
|
#10
|
|
Glass Joe
|
Chimera Shot Calc?
First of all, thank you for the spreadsheet. I will be using it to maximize my rotations, and already see how watching my cooldowns more effectively can boost my DPS.
One question though: The Chimera Shot Usage calculation in cell J7 is using cell B74 (potential Chimera Shot Cast Time) instead of cell C74 (Potential Chimera Shot Shots). Should it be pulling from cell C74?
|
|
|
|
|
03/23/09, 3:12 PM
|
#11
|
|
Piston Honda
|
Originally Posted by DawizNJ
One question though: The Chimera Shot Usage calculation in cell J7 is using cell B74 (potential Chimera Shot Cast Time) instead of cell C74 (Potential Chimera Shot Shots). Should it be pulling from cell C74?
|
Indeed it should, same with multi-shot. Some leftovers from early passes I made at this. Fixed in v0.1.0.5 now.
|
|
|
|
|
03/24/09, 8:42 AM
|
#12
|
|
Von Kaiser
|
Originally Posted by Esoth
Anyway, the main use I see from this is not solely looking at your numbers in a vacuum and wondering why it's not all 100%. Some potential uses would be:
1. Comparing people of the same spec in the raid. It's something to look at other than the final DPS result, and it's some way of comparing performance aside from gear.
2. Comparing yourself on different fights. What's my average uptime on serpent sting on boss X? Steady usage on boss Y instead of boss Z?
3. Personal performance could probably be measured better if we had a lot of data run through this and saw what kind of numbers people were actually getting. For instance, to know that 85%-90% (as a made up example) explosive usage is a realistic goal to aim for.
|
I have devised a way to measure performance apart from gear. You might want to check out my thread on Hunter Statistics. I have indentified certain ratios that will enable us to compare hunters in DIFFERENT raids and adjust for buffing. I did this during BC, so there are a few things that I will need to update to adjust for WOTLK (like aspect of the viper). I did two studies, one of Gruul and one of Lurker, and those gave me benchmarks for hunter performance. I have posted a copy of my work in the hunter subform at Elitist Jerks. Statistical Reference Guide for hunters
I have done a pilot study of patchwerk in WOTLK, but I have not wanted to present it at EJ yet. There are too many problems with sample selection. I will probably do another study after patch 3.1 which I will release.
The big problem with doing this work is data collection. I would really like to get these ratios built into a log parser. It takes way too much time to manually compute them.
|
|
|
|
|
03/24/09, 9:04 AM
|
#13
|
|
Von Kaiser
|
Originally Posted by Faerdael
I'll look into importing WWS data, however it may be slow for me to get it done, I don't know. Frankly never tried it (though I really should have) so its something good to practice. I can't promise any kind of timeline for that though.
|
I have asked more technically oriented people to download data from WWS, and they were not able to. They said that the data is kept in data base format, and recalculated each time. For reasons that I don't fully understand, that makes it very difficult to download.
I have manually copied WWS reports. It takes me about 6 hours to copy everything that I need for a population of 25 hunters (that includes visiting the armory as well).
|
|
|
|
|
03/24/09, 2:50 PM
|
#14
|
|
Bald Bull
|
You wouldn't need to scrape the entire combat logs off the WWS pages (that sounds a ridiculously inefficient way to get a huge amount of data), wws already summarises all the information you want for this spreadsheet, since it just needs the counts of various shots/buffs - you'd just need to scrape the damage breakdown page and buffs/debuffs page for the particular hunter you want to look at.
|
|
|
|
|
03/24/09, 5:34 PM
|
#15
|
|
Glass Joe
Blood Elf Hunter
Executus
|
Player Settings, Combat Settings questions
Faerdael, thanks for the great work. I have a couple questions.
The screenshot (first post) shows 0 for Explosive Shot under Player Settings. Armory stalking you shows you are 5/15/51, so you have ES, shouldn't there be a 1 in the ES column in Player Settings?
Latency something you guess at? Manually track via WoW client while in game? And what's the units? I typically run 55ms latency, put 55 into Combat Settings->Latency?
Thanks.
|
|
|
|
|
|