Oh!! Golemagg's dogs! You can set it up so you can attack from either the back or the front. Golemagg with no DPS on him would take forever to kill. But again... would take a while to clear to him.
I don't remember an enrage timer ever popping on Yauj, and we had some VERY long bug trio fights back in the day.
Major point of concern with Ebonroc would be the detaunt - eventually you'll hit the aggro ceiling with him unless someone taunts the wing buffet, which adds another layer of complexity to it (and also changes the way he faces, which will alter parry statistics while he's turned).
Last edited by Shaker : 05/21/07 at 5:54 PM.
Reason: N/m, 5 man bosses aren't skulls.
Oh!! Golemagg's dogs! You can set it up so you can attack from either the back or the front. Golemagg with no DPS on him would take forever to kill. But again... would take a while to clear to him.
I think the bigger problem would be trash respawns; clearing Mag to stop dog respawns and then having a couple rogues and druids stealth back to Golemagg and clearing only his room shouldn't take overly long; but trash respawn is 2 hours, which will put an upper limit on the length of your sessions.
I think the problem with Ebonroc would be aggro; if you have two warriors/paladins/whatever hanging out in front of him, they're both going to get wing buffeted quite a bit such that eventually the healer would catch them on aggro.
A test at two other points of weapon skill, say +5 and +10, may be more useful than running more swings.
From what little I remember from stat class, the standard deviation and your confidence interval shrinks ever more slowly as you pile more data points in.
With tests at different points you might be able to fit a trendline and narrow things down faster. 2c from somebody who hasn't done a stat analysis in a long time.
I think the bigger problem would be trash respawns; clearing Mag to stop dog respawns and then having a couple rogues and druids stealth back to Golemagg and clearing only his room shouldn't take overly long; but trash respawn is 2 hours, which will put an upper limit on the length of your sessions.
Golemagg doesn't have a leash - and there's places you could take him where there's no trash repop, but the puppies aren't skull level, so you'd eventually kill Golemagg regardless.
What about Sulfuron? His adds don't ever run out of mana, do they? I don't know that we ever tried killing him first to see if they healed HIM though.. I know they heal eachother.
A test at two other points of weapon skill, say +5 and +10, may be more useful than running more swings.
From what little I remember from stat class, the standard deviation and your confidence interval shrinks ever more slowly as you pile more data points in.
With tests at different points you might be able to fit a trendline and narrow things down faster. 2c from somebody who hasn't done a stat analysis in a long time.
I don't personally see how this would help, unless we're concerned that the answer is not linear. What would be useful is to test with a larger variance in weapon skill - for those of you tracking the calculations, we get to divide the error by the amount of weapon skill being used. In order to narrow the uncertainty by a factor of two, it would work to either a) take 4 times as much data or b) test with weapon skills that differ by twice as much (or some combination of the 2). While tests at intermediate values wouldn't be *totally* useless, if we assume the effect of weapon skill is linear, testing at smaller offsets is strictly less useful than testing at larger offsets.
Golemagg doesn't have a leash - and there's places you could take him where there's no trash repop, but the puppies aren't skull level, so you'd eventually kill Golemagg regardless.
Hmm, point. I was trying to avoid clearing trash all the way back to a safe area, but that may not be avoidable.
The fact that the dogs are level 62, however, is a problem.
I think the bigger problem would be trash respawns; clearing Mag to stop dog respawns and then having a couple rogues and druids stealth back to Golemagg and clearing only his room shouldn't take overly long; but trash respawn is 2 hours, which will put an upper limit on the length of your sessions.
I think the problem with Ebonroc would be aggro; if you have two warriors/paladins/whatever hanging out in front of him, they're both going to get wing buffeted quite a bit such that eventually the healer would catch them on aggro.
Well, my plans on an Ebonroc test would put the healers in front too. Shadow Flame isn't going to kill anyone outright with today's hp totals. So you would just need to guarantee a certain level of healing. It would have been more healing than my test on Venoxis for sure, but I think I could have set something up for it. So an option down the line.
Originally Posted by The Grog
A test at two other points of weapon skill, say +5 and +10, may be more useful than running more swings.
From what little I remember from stat class, the standard deviation and your confidence interval shrinks ever more slowly as you pile more data points in.
With tests at different points you might be able to fit a trendline and narrow things down faster. 2c from somebody who hasn't done a stat analysis in a long time.
Yes, that would be worthwhile as well. Not only in helping lower the variance some (it is essentially more data), but we could also try to eventually find a cap to it. My hunch is that the cap is 5*level difference for weapon skill. But more data points can help tighten down all the points.
Another reason Ebonroc is good would be for this reason. We could pile on a ton of characters for the test and there would still be no way in hell of killing him. Since the simple test on him would keep him locked on one tank the entire time, he would pretty much be healing full time. Unless the level difference resist stuff causes his debuff to be resisted too much.
I think Venoxis (or maybe Jeklik phase 2) is likely to be the easiest to do ... add another lifebloom though, and you could probably do Thekal, phase 1... kill him, he resurrects... kill him again, he resurrects - the shaman lady heals quite a bit as well. The only issue would be a slightly longer clear to get to him, and the higher DPS from the two adds. I suppose the Gouge from the rogue helper would be a bit annoying too, but I doubt that he could kill a tree druid before the gouge wore off.
... you could probably do Thekal, phase 1... kill him, he resurrects... kill him again, he resurrects - the shaman lady heals quite a bit as well. The only issue would be a slightly longer clear to get to him, and the higher DPS from the two adds. I suppose the Gouge from the rogue helper would be a bit annoying too, but I doubt that he could kill a tree druid before the gouge wore off.
We use to run ZG once a week to have a look at Bloodlord Mandokir and High Priest Thekal for the mounts if we get the time. You can 100% skip to Thekal, 1 pull for Mandokir.
Their gear was 100% identical and had zero crit / hit rating on it, so I am fairly certain that the only thing that would affect results is the 3 agility difference (0.12% crit).
Their gear was 100% identical and had zero crit / hit rating on it, so I am fairly certain that the only thing that would affect results is the 3 agility difference (0.12% crit).
I note that the paper doll for Therasiv is showing 0.6% extra crit from weapon skill difference. That is 0.04% crit per weapon skill. So they just nailed one of the differences for us with the paper doll.
It's well established that tooltip crit goes up by .04% per skill. The question is whether it's correct; for instance, some patches back, they said that there was a bonus .1% crit per skill that applied vs level 73s, which appears to be MIA. The current data makes .04% somewhat plausible, but we'd need more testing to be sure.
It's well established that tooltip crit goes up by .04% per skill. The question is whether it's correct; for instance, some patches back, they said that there was a bonus .1% crit per skill that applied vs level 73s, which appears to be MIA. The current data makes .04% somewhat plausible, but we'd need more testing to be sure.
Well the paper doll is for even level targets. So I am guessing the 0.04% is guaranteed no matter what level your target is. The missing 0.1% for above level targets is a bit baffling, though it was worded really weird as well as I remember.
Hmm actually, if you you get 0.1 % crit per skill against a higher lvl target, that would mean you get more %crit if the target is higher than you because you only loose 0.04 crit etc per point of defense but you gain 0.1 %crit .
So
1 wep skill counters 2.5 def skill for higher lvl mobs
1 wep skill counters 1 def skill for lower/same lvl mobs
With increasing lvl of the target this would mean first a big jump and then a slow lowering of your %crit (0.06 crit / lvl or 1.5 skill/lvl)
This seems kinda weird to me. So i thought that they maybe did it in an other fashion.
Maybe you dont loose 0.04 % crit per def skill with a target higher that your lvl, maybe you loose 0.1 % crit per def skill.
So 1 wep skill counters 1 def skill for higher/same/lower lvl mobs
Maybe its not 0.1 but 0.14 (0.1 + the base 0.04) but this is can be found out later. The important thing is that 1 wep skill counters 1 def skill.
Just a theory, but if that thoery was true we would have to check crit rating against higher lvl mobs with no additional +skill.
Hmm actually, if you you get 0.1 % crit per skill against a higher lvl target, that would mean you get more %crit if the target is higher than you because you only loose 0.04 crit etc per point of defense but you gain 0.1 %crit .
So
1 wep skill counters 2.5 def skill for higher lvl mobs
1 wep skill counters 1 def skill for lower/same lvl mobs
With increasing lvl of the target this would mean first a big jump and then a slow lowering of your %crit (0.06 crit / lvl or 1.5 skill/lvl)
This seems kinda weird to me. So i thought that they maybe did it in an other fashion.
Maybe you dont loose 0.04 % crit per def skill with a target higher that your lvl, maybe you loose 0.1 % crit per def skill.
So 1 wep skill counters 1 def skill for higher/same/lower lvl mobs
Maybe its not 0.1 but 0.14 (0.1 + the base 0.04) but this is can be found out later. The important thing is that 1 wep skill counters 1 def skill.
Just a theory, but if that thoery was true we would have to check crit rating against higher lvl mobs with no additional +skill.
The parry/dodge discrepancy throws out the idea that weapon skill <-> defense is a 1 to 1 relationship. The 0.1% crit for higher level targets was what Blizzard said when they got rid of the effect weapon skill had on glancing blows. It was suppose to help make up for it. Now a lot of the results we are seeing here may be related to this, but it is likely they just coded in something special for weapon skill values on mobs in pve that are higher level than you are. Theras didn't give me numbers, but he said he parsed +weapon skill against an even level target and really didn't see much effect. I know we can't say we didn't see much effect in our Venoxis tests, because we did see a big change.
The parry/dodge discrepancy throws out the idea that weapon skill <-> defense is a 1 to 1 relationship. The 0.1% crit for higher level targets was what Blizzard said when they got rid of the effect weapon skill had on glancing blows. It was suppose to help make up for it. Now a lot of the results we are seeing here may be related to this, but it is likely they just coded in something special for weapon skill values on mobs in pve that are higher level than you are. Theras didn't give me numbers, but he said he parsed +weapon skill against an even level target and really didn't see much effect. I know we can't say we didn't see much effect in our Venoxis tests, because we did see a big change.
Well, in the rogue dps spreadsheet threat, there was extensive examination of the effects of weapon skill on hitting same-level mobs, and the conclusion was that it seemed to be .1% hit per skill, so there's at least some effect on same-level mobs.
Theras didn't give me numbers, but he said he parsed +weapon skill against an even level target and really didn't see much effect. I know we can't say we didn't see much effect in our Venoxis tests, because we did see a big change.
I tried some light testing against a lvl 54 mob in blasted lands, 10k white swings with 350 and same with 360 weapon skill. Paperdoll showing 5.25% crit with 350 skill and 5.65% with 360, 38 hit rating in both tests no precision (screen/screen).
Seems to support the 0.04 crit per skill against even/lower level mobs shown in earlier parses, noticeable difference at least. I think tests against true lvl 73's is needed to know if the patch notes are correct or not, as we don't know what the "skull level" mechanic really does.
Oops, I fell to exaggeration. What I meant to say is that the effect was less than what we saw in the boss mob logs. However, since I didn't run any numbers on it nor received any logs on this, I will defer to you guys on any tests already done there. So was the effect for same level mobs less than what the logs currently show for boss? Or was the tests too short to know for sure?
I would say the glancing results alone support that "skull mobs" behave as Level+3 for combat table results. Though I heard that Gurg said he had logs from Karazhan showing +weapon skill affecting parry as well, which is counter to our results.
The general conclusion is that we need more testing to make a general conclusion. Crit is probably +0.04% / weapon skill, % to hit and % to be dodged are still up in the air.
Well as the thread says the testing done in the rogue dps thread pretty much concludes each point gives 0.1% hit to equal level mobs aswell as higher level mobs.
What's not settled is the extra crit chance and the supposed reduced avoidance of mobs. The char screen add 0.04% crit per skill, but we've yet to confirm that information as correct.
Well as the thread says the testing done in the rogue dps thread pretty much concludes each point gives 0.1% hit to equal level mobs aswell as higher level mobs.
What's not settled is the extra crit chance and the supposed reduced avoidance of mobs. The char screen add 0.04% crit per skill, but we've yet to confirm that information as correct.
So this is per point, meaning which of the following:
05/21/07, 5:39 PM
Similar Threads
