[:1]As you know, there are a lot of trinkets, jewelry and even some weapons and armours which randomly grant certain buffs while in combat. I didn't care to use them until now, because I don't really like things I can't control, things I can't rely upon (as opposed to trinkets I use myself at the correct time to receive a buff). But, looking at new WoTLK items... it seems they could be quite powerful, even tho their procs are random. I would like to test it myself. Though, it would be rather funny to receive a great buff on hit... in case this is the last hit on dying monster :P You receive a powerbuff for 15 secs, but once you loot and skin the mob, and reach next one - the buff is already gone... That's why I would like to gather some statistics - for example, I want to know percentage of combat time I'm affected by certain buff. I've found an addon which counts buff uptime - but it counts total time, and I couldn't care less about buffs on me while OUT OF COMBAT! Can you suggest an addon to calculate buff uptime during combat only? Would be very good if it was able to count also internal cooldown of these items, as well as average combat time between "internal cd off" and "buff reappeared" events!
I've searched Curse, but there are 177 addons in the buffs section, without much comments... and I can't sort them by rating or hits to discover the most useful ones :( It displays an error when I try it...|||Ehh well... anyone? %)|||I don't know of anything like that. The way most theorycraft models things that have a proc chance is to find out how often it procs on a target dummy (by analyzing the combat log) and then assume the dps increase is spread across a fight.
For example, if a trinket procs 6% of the time and lasts 10 seconds, then you can say over a five minute fight the buff will be up X% of the fight... while it's up you get Y% dps increase... so average that over the five minute fight and the benefit of that trinket is that it yields Z% more dps. Add that to any equip benefits and you can compare it to another trinket to figure out which is better. You can use the same basic approach for non-DPS trinkets
Even an addon that did what you ask wouldn't account for a trinket that procs 2 seconds before combat ends (so that most of its uptime is out of combat), and over enough fights, the averaging noted above will model the effect of the trinket well enough.
EDIT: You might try logging combat (/combatlog) and uploading the resulting log file to WWS to see if it pulls out that information.|||Quote:
I don't know of anything like that. The way most theorycraft models things that have a proc chance is to find out how often it procs on a target dummy (by analyzing the combat log) and then assume the dps increase is spread across a fight.
Theorycraft would not help AT ALL :( Yes I know proccing trinkets have 45s hidden cd; cd of trinkets which you use manually is not even hidden. The problem is that I can't say for sure what is my average fight duration, what is average delay between fights and so on... For example, what is going to be better for me - 60dps buff for 20 secs or 100dps for 10 secs? In the encounter with some static boss (like Reaver or Teron) first one is going to be better for sure (as it grants 1200damage instead of 1000 for the second), but you can't be so sure when you fight common mobs. Maybe my fight ends way faster than 20 secs, and I get an useless combat buff after combat, so I'd rather get a 10 sec long POWER buff? Or maybe I get an add, and use full duration of 20 sec buff? What is my chance to get an add? You never know that until you test it... That's why I need an addon to gather such statistics...
Quote:
Even an addon that did what you ask wouldn't account for a trinket that procs 2 seconds before combat ends
Why can't it? I've already found an addon which stops recording proc uptime once you are out of combat - why can't there be an addon which does the same for "on use" trinket buffs?
Quote:
EDIT: You might try logging combat (/combatlog) and uploading the resulting log file to WWS to see if it pulls out that information.
I guess I'll have to... in case I don't find any easier way (i.e. an addon) :(|||OK, first, small trinket differences just aren't important on 12k hp quest mobs. So I wouldn't worry about a marginal diffence between trinkets for things like that. Small differences will make a difference on high hp mobs... some trash and bosses.
Quote:
The problem is that I can't say for sure what is my average fight duration, what is average delay between fights and so on...
Doesn't matter. You assume infinite fight duration to model a boss because what you want to do is average out the effect of crits, trinket procs, etc. Once you do that a simulation will show you that, in an infinitely long fight gear variation A is better than variation B. That translates to long boss fights pretty well (5min+). For short boss fights I'm not sure the small variation between two trinkets matters much. The effects can be masked by luck on crits, whether you perfectly execute your rotation or are a slight bit off, etc.
Quote:
For example, what is going to be better for me - 60dps buff for 20 secs or 100dps for 10 secs? In the encounter with some static boss (like Reaver or Teron) first one is going to be better for sure (as it grants 1200damage instead of 1000 for the second), but you can't be so sure when you fight common mobs.
That's not how I'd model that. I'd model it on a longer fight and figure out how many times I can have the effect up on average over a fight. For that you not only need to know dps buff and buff duration but proc rate. If we assume proc rate is the same, then you'll ALWAYS want the first trinket on boss fights since, as you point out, it grants more actual damage. For trash the difference too small to matter. Same with quest mobs.|||You didn't get my point... rather, neither of my points :(
Quote:
OK, first, small trinket differences just aren't important on 12k hp quest mobs. So I wouldn't worry about a marginal diffence between trinkets for things like that. Small differences will make a difference on high hp mobs... some trash and bosses.
What is the difference? Indeed, it is not important for a short fight vs a single common mob. Indeed, it IS important for a 10-min long combat vs a raid boss. But it is ALSO important for hours long killing of some trash mobs while farming some regs or rep!
Quote:
Doesn't matter. You assume infinite fight duration to model a boss because what you want to do is average out the effect of crits, trinket procs, etc.
Exactly! I assume infinite fight to average random effects... but this fight is not continuous - as it consists of "in combat" and "out of combat" parts. That's why any model would not work, because I'm unable to estimate time percentage of these parts... their average length and so on. It HAS TO BE TESTED!
Quote:
For that you not only need to know dps buff and buff duration but proc rate.
You didn't get it either... I've already found a working addon (Procodile) which counts proc uptime percentage IN COMBAT. But, it doesn't count buff durations of "on use" trinkets :( And I would really like to test "on use" versus "chance on hit" gear dps bonus!|||I understood your points... I just don't care about testing on random quest level mobs as it's not important. Too many things can happen in a given short fight (crits, etc) that will completely overwhelm the effect of a trinket proccing. But who cares if I kill a given mob in 8 seconds or 7?
Yes, I get that a proc effect might run for some OOC time. I just don't care. No trinket proc is going to boost my dps so much that it will materially affect how fast I kill a quest mob. A random crit on a mob will have a far greater effect.
Look, a really good trinket might make 20-50dps difference. Let's say it's 50. At 1000dps I can kill t he typical Icecrown mob in 12secs. At 1050, I can kill that same mob in... 11.4 secs. /shrug.|||Part of the function of my addon Bang! Proctologist is to do what you are asking.
It may be a bit more than you wanted however.
No comments:
Post a Comment