Evaluating Raid Performance with CompareBot
by, 09-11-2012 at 12:45 AM (26297 Views)
Following the recent trend of mine of not doing Elemental specific blog posts, Iím going to write something that should help any DPS player (not so sure about Healers/Tanks, but you never know) evaluate their performance & find ways to improve by using two websites: Raidbots and World Of Logs.
Firstly, we need to find some example parses for you to use. The best way to do this is to use the ďEpeen BotĒ to find your parses and see how they compare with the available rankings at the time. My current preference for evaluating performance is Feng the Accursed because there are no multi-target gimmicks, no damage buffs and allows for a decent period of stand & burn DPS.
This is how my performance on Feng 25N looks on the graph. Thereís a 82 percentile result on the 23rd, which was our first kill, a rather poor 55th two days later, and a fairly amazing 96th percentile on the 8th of November (which clocked in as the 17th highest Feng 25N parse at the time). There was another kill in there, but we forgot to switch on combat logging.
In any case, we can use this to select a parse to compare. Since I have three spread over a wide range, Iíll use all of them as comparisons. Youíll need to go to each log at World of Logs, go into the Kill report for the boss, select Damage Done, and then click into your name. This will give you the individual details for you for that particular fight. For yourself you may find it better to use your best parse compared with one or two of the World of Logs ranking pages
Copy the URL for that page into the Parse URL boxes on CompareBot, add a name to help keep track of the differences, and press Compare Parses.
Press the button and wait for some results to be produced! In the meantime you can twiddle your thumbs or something.
Once the comparison is completed, youíll see a summary of the compared parses at the top. You can make some useful inferences from this section. For example, even though I have a 18,538 dps difference between my best & worst parses, the damage difference is fairly minor at 1.30%, and the 82 parse in the middle actually did more damage overall. As the fight doesnít include any extra targets, it means that my damage in comparison to the rest of the raid didnít change that much.
Incidentally, the only parse where I wasnít in first place in the raid was the 96 parse, where I was third.
From this we can see that while my performance in comparison to everyone else uploading Feng 25N parses varied greatly between parses, my performance relative to my fellow raiders didnít vary that much. Likewise, when comparing to other players youíll be able to see whether your performance relative to your raid was much different to theirs. For example, if my 55 parse only showed 25 million damage, then Iíd know that I had done something wrong that day and would need to fix it. Likewise, if my 96 parse was 35 million damage, Iíd know that not only was I doing really well that day, but the rest of my raid underperformed a bit too.
The other thing to note here is that the duration of the fight has a massive impact on your DPS, which should be fairly obvious given that itís Damage Per Second. Itís impact is a little different than just reducing the number youíre dividing your damage by though, as a lower fight duration can result in higher relative uptimes for damage buffs like Bloodlust.
The first tab you should look at is the Damage tab. This will show you all the information about the various abilities used, total damage/hits/misses/crits/etc.
The most important column to look at is the Hits/Minute column. This will tell you how your action use compared to the comparison parse. The 96 parse (column 1) will have a higher action count than the other two purely because Iíd gained more gear, and thus more haste, but we can see that the hits per minute of the 55 parse (column 2) was lower than either of the other two by a considerable margin, even though it was the second of the recorded parses.
From the screenshot above, you can see that my Lava Burst & Lightning Bolt casts were slightly lower and I had a much lower Lava Burst Overload count (thatís the second Lava Burst in the list, for all you non-Elemental people out there). My Elemental Blast use was pretty good between all three parses, so I know I donít have to work on that, but Iíd need to up my LvB & LB use a bit.
Most of the things youíll be looking at will be specific to classes, specs and abilities. Things like Average Hits, Crit % etc will all be dependant on your gear, which can be improved by using SimulationCraft, reading guides, or even reviewing the armory of the player youíre comparing yourself against (note that they probably will have different gear when you have a look).
Donít get too focused on the Total Damage, Ability DPS and Average Hit columns. These are all products of other things youíll be looking at, and you can only really compare them between similarly geared players.
On the buffs tab you can see the uptime, amount & amount/minute values for each buff. The amount value can be ignored for the most case, as the count of buffs will be fairly irrelevant. Itís the uptime & amount/minute values that youíll be more concerned with. For buffs like Savage Roar youíll be concerned primarily with the Uptime value, making sure itís as high as possible, while Amount/Minute is more useful for charge based procs like Lava Surge where you want to look at the frequency of the buff.
What Not To Do
Itís important to note that the Epeen Bot name is intentionally ironic. Doing really well is good, but itís definitely not good to try gimmick tactics to boost your own performance at the expense of the rest of the raid.
You should also take things in context. A relaxed 10 man group running once or twice a week may never reach the sort of numbers that Paragon is putting out, but it doesnít mean that you canít compare things Iíve mentioned above. Nor will the fight duration be the same length for similar reasons, which again will skew some of the numbers.
Total Trackbacks 0