此站点大量使用JavaScript。
请在您的浏览器中启用JavaScript。
正式服
PTR
10.2.7
PTR
10.2.6
Beta
Best way to test DPS
发表回复
返回主列表
发布者
104767
This post was from a user who has deleted their account.
发布者
524425
This post was from a user who has deleted their account.
发布者
R1TeR
the dummy is your best friend
发布者
Porcell
Unless you are going to get a bunch of friends together to stand with you for buffs, and unless you are going to stand there and attack the dummy for a very very long time, you won't get much meaningful data from the target dummy.
On the other hand, you don't have much of a choice otherwise.
What exactly are you "testing" here? You can't just go to a dummy, swing for a while, and walk away saying "I can do 7k DPS!" because it will not in any way be representative of what you will do in a raid.
发布者
131584
This post was from a user who has deleted their account.
发布者
104767
This post was from a user who has deleted their account.
发布者
Foopeemoa
Dummies are not a good way to judge gear upgrades, since relative stat-weights change from unbuffed to the full raid-buffed environment.
There are basically three ways to judge gear upgrades:
1) Get half of your guild to come to the dummy with you, and put up the same buffs/debuffs you would have while raiding. This works for some classes but not for others, since the dummy is usually at 1% health and some things work differently below 35/30/20% health depending on the class.
2) Use a prediction tool. Generally that means Rawr, Landsoul's Spreadsheet, or SimulationCraft. All three are fairly accurate for comparing pieces of gear in a controlled environment. You can also just use them to generate stat-weights and then use those stat-weights to judge gear separately, however the stat-weights change as you equip different gear, obviously.
3) Use your gut. Just eyeball the gear and go with whatever you feel. Obviously this isn't what you want, if you are trying to compare gear choices on a dummy.
发布者
560025
This post was from a user who has deleted their account.
发布者
brcondor
You could build a statistic model. With 30-50 osservation of a fact you can build a gaussian approsimate model... and for example you can test if the mean of your dmg is bigger or not with an accuracy of 90 95 or 99%... or you can do this with a skill. You have to spam this skill almost 10-15 times and then you can build your model.
发布者
439646
This post was from a user who has deleted their account.
发表回复
您没有登录。请
登录
发表回复或
注册
如果您还没有账号。