Mathematical study of value of Pulverize

Druid
There is a point at which Expertise does get better than Crit though. But it's entirely dependant on your current Crit, Hit, and Exp.
At this point (around 4.2) it's certainly close enough (at least itemization-wise) that you could reasonably argue for going with expertise pieces over crit pieces as a bear - at least when you're below the dodge cap. Going forward I'd have to get some idea as to what our numbers will actually be, but my suspicion is that there's a good inflection point around the 50% crit value for expertise.

That being said, if SD doesn't start getting some fast improvements or bears don't start getting a lot more avoidance SD's going to be pretty irrelevant.
Also, Katja: GEM YOUR DAMN BELT, YOU SCRUB
05/04/2011 06:02 PMPosted by Fellhoof
you could reasonably argue for going with expertise pieces over crit pieces as a bear

The main problem with the crit vs expertise thing is that there is only one 372 Mastery/Expertise item, but many Mastery/Crit items. And mastery is still better than both.
Ok Thanks, after reductions before avoidance.

Now I am trying to understand the following.

Pulverize's results:
161793 average SD hits (hit with SD active)
138291 average SD misses (hit with SD not active)
199914 average avoids
87491 average SD overwrites (SD procced when SD already was active)
34.33% average uptime


Are you calculating the SD average uptime directly from your sim or from these numbers?

To me it 34.33% looks wrong. SD was presumable up the same proportion during avoidance as hits so we can ignore these. Then SD uptime is approx = (hit with SD active)/(hit with SD not active + hit with SD active) = around 53.9% with pulverise and 44.2% for not active.

So about 40% of the time you take nothing. Then another 35% of the time you block with SD, and then 25% of the time you don't block anything at all. The dodge part massively overwhelms the advantage of SD here. That's why it's so low.


ok this possibly explains part of my confusion. When you say uptime you mean "percent of swings that hit SD", not percentage of total time SD is active for.
Actually, the uptime there is literally how long SD is active and sitting around. So if it's up for 40% of the time that means that you've got a shield waiting to be used about 40% of the time.

Are you calculating the SD average uptime directly from your sim or from these numbers?
It's part of the sim. Every tick that the bear has a shield up I add to the uptime. That's all it is. This means the following:

-any time the bear would use a shield the uptime had to have been there
-any time the bear avoids a hit while having a shield up the shield is there as well.

The % of swings that hit SD is easily figurable from this: there are 500,000 total boss swings. Simply take how many shields got used and divide by that 500k and you'll get the percentage.

The main problem with the crit vs expertise thing is that there is only one 372 Mastery/Expertise item, but many Mastery/Crit items. And mastery is still better than both.
Yeah, that's true. And it sucks. But it's something to consider going forward and in looking for gear.
05/04/2011 06:14 PMPosted by Fellhoof
GEM YOUR DAMN BELT

O.o

godamn blizzard.
lol, thanks :) ran the entire raid as DPS the night after blizzard changed that so put off updating my tanking set :P

what's the bet I've completely forogtten about it again by the time I get online tonight...

should probably take a few seconds to actually ask /g who the heck's got agi to bracers... but something about being an enchanter myself makes it that much more anoying relying on others for enchants :S
05/04/2011 06:38 PMPosted by Fellhoof
But it's something to consider going forward and in looking for gear.


I'd definitely prefer expertise/mastery itemisation over crit mastery right now.

pity the offset stuff mmo's mining from the PTR is all haste/mastery :P
Well, everything with agi seems to be haste/mastery at this point. I'd be shocked if that went live. They usually like a little bit of variety. And hey - that just means that the set pieces are going to be AWESOME, right? Right?

sigh.
To be honest, it will be difficult going back to low expertise, since I'm really only reforged this way for top 10 parsing. >.>;
That's the thing - I really want expertise to be more valuable than crit and on par with mastery, because expertise is simply more fun. It's fun to hit things and it's not fun to be parried/dodged 1/5th of the time. I don't want dodge to be removed as an awesome ability, but I'd love it if they rebalanced SD such that expertise became as valuable as mastery or even moreso before the dodge cap.

That really wouldn't be that hard if they wanted to; I suspect the easiest way would be to make it proc 66% of the time or so but nerf mastery by 33%.

Or, ya know, make threat and damage actually matter on a fight.
The % of swings that hit SD is easily figurable from this: there are 500,000 total boss swings. Simply take how many shields got used and divide by that 500k and you'll get the percentage.


Yes ok, but I must be missing something here.

From these numbers.

Pulverize's results:
161793 average SD hits (hit with SD active)
138291 average SD misses (hit with SD not active)
199914 average avoids
87491 average SD overwrites (SD procced when SD already was active)
34.33% average uptime


The shield is active 34.33% of the time. However somehow it is active on 53.9% of all hits? How does this occur? I can sort of see that dodging not using shield may skew the percentage to some degree but this seems to be a lot.
The shield is active 34.33% of the time. However somehow it is active on 53.9% of all hits? How does this occur? I can sort of see that dodging not using shield may skew the percentage to some degree but this seems to be a lot.
I don't know where you're getting the 53.9% number.

AH! Okay, I get it.

53.9% is the total blocked hits out of the non-avoided hits. It's (blocked hits / blocked + unblocked hits). Which is an interesting number, but it isn't to do with uptime. Remember, you can avoid attacks when you don't have a shield active, and any time you do you don't have uptime either.

It is active 34% of the time. It occurs on 32.3% of ALL attacks (avoided, blocked and unblocked). That means there's a very low amount of time when you have a shield and it's sitting around waiting to be absorbed, which given the low crit rate and low avoidance rate is to be somewhat expected.
That's the thing - I really want expertise to be more valuable than crit and on par with mastery, because expertise is simply more fun


I've reforged to get as close to hit capped as possible for regular heroic 5 mans simply because - when you're regularly doing over 40% of the entire group's damage, missing a swing or two can literally be bad for your health with the extra few sings your targets will get in before they die :P

that, and it's oh so much more fun.
Honestly I think the times I can count getting parry/dodge/parry or anything of the like(3 consecutive attacks that didn't land due to no expertise) on one hand. One of them was on Nef with Pulverize(which was absurdly frustrating), but I'm a grown up so I can get over it and know that it's better to live and be frustrated than die and not be frustrated.

Thanks for all the help with your Pulv math! I had some abysmal uptime on it last night(was doing heroic maloriak adds...derrrrrr), and felt really sad! :(
thank you for giving me a reason not to be a terrible bear and let pulverize drop off because of laziness. Now I have a place to go to verbally beat myself when I realize it disappeared 5 globals ago. Bad ssunflash, bad!
53.9% is the total blocked hits out of the non-avoided hits. It's (blocked hits / blocked + unblocked hits). Which is an interesting number, but it isn't to do with uptime. Remember, you can avoid attacks when you don't have a shield active, and any time you do you don't have uptime either.


Yes. I agree with this.

Actually I've just realised what the issue is. Uptime is not the same percent as "shield up when hit". Sorry perhaps I have been slow.

So if for example if we start with no shield, boss swing is 2 Secs, and we Maul once per 3 secs, yellow 1 per 1.5 and melee at 2.5 then our average attack per second is = 1/3 + 1/1.5 + 1/2.5 = 1.4 so in the time between boss swings we have 2.8 hits to refresh the shield, which at say 40% crit is 20% to refresh per hit or (1-(1-0.2)^2.8) = 0.46, and this is increased by there being a 40% chance of missing. The continuing sum boosts the chance up to 0.46/(1-0.4*0.46) = 0.56% chance of being there when hit.

Actually I am sure if I dig through Tangedyns spreadsheet that what he does. More or less.
05/04/2011 07:29 PMPosted by Schifting
Honestly I think the times I can count getting parry/dodge/parry or anything of the like(3 consecutive attacks that didn't land due to no expertise) on one hand.


I was just about ready to murder Halfus last night after

[22:48:05.485] Halfus Wyrmbreaker afflicted by Lacerate (3) from Katjia
[22:48:07.206] Katjia Lacerate Halfus Wyrmbreaker *33722*
[22:48:08.420] Katjia Lacerate Halfus Wyrmbreaker Dodge
[22:48:09.889] Katjia Pulverize Halfus Wyrmbreaker Parry
[22:48:11.585] Katjia Mangle Halfus Wyrmbreaker 68293
[22:48:12.898] Katjia Thrash Halfus Wyrmbreaker Dodge
[22:48:14.393] Katjia Pulverize Halfus Wyrmbreaker Miss
[22:48:15.909] Katjia Pulverize Halfus Wyrmbreaker Miss
[22:48:17.388] Katjia Pulverize Halfus Wyrmbreaker Parry
[22:48:18.900] Katjia Pulverize Halfus Wyrmbreaker Parry
[22:48:20.434] Katjia Pulverize Halfus Wyrmbreaker Miss
[22:48:22.220] Katjia's Lacerate fades from Halfus Wyrmbreaker


although, technically I guess we did kind of murder him and steal all his stuff :P
Actually I am sure if I dig through Tangedyns spreadsheet that what he does. More or less


And yes on checking it is, if his has a little more finess (haste etc). Anyway this explains my confusion. Tangedyn's spreadsheat number of "SD uptime" was much higher than your sim - more like the higher number of percentage of hits with SD active that I was quoting.

The difference being that Tangedyn's number, while entirely relevant for his calculation of avoidance, is not truly "uptime" like in Felhoof's sim calculation. However Felhoof's number of course doesn't really tell you how often it will be there when you are hit, as a fair chunk of down time is immediately after a boss hit when you won't be receiving hits for the next 1.5+ secs or so typically.

Apologies to anyone if my going through this has been a little slow, but I wanted to be sure I understood what was going on.

Join the Conversation

Return to Forum