Also, you don't really need a 3.4 speed weapon at minimum, although the slower the better. If I use a 2.7 speed weapon, it's very clear that doing 2 steadies between autos is better DPS than one, even though it greatly delays each autoshot. If multi or arcane do more damage than steady, i just substitute them on cooldown for one of those steadys. 3.4 is just the closest speed we have to losing no autoshot time, and i doubt they'll let us get any closer by giving us an even slower weapon.
My experience of it is that the only place your reaction time comes in is firing the first of those two Steady Shots. This is no more demanding than the timing of Aimed Shot many hunters are used to pre-1.13. The additional possible delay is that even though all you have to do to get the 2nd steady is spam the button while the first casts, latency will cause some delay between the two shots.
Correct me if I'm wrong again, but not only is the delay between the first and second Steady Shot due to latency, but also because at 15% haste, you're firing Steady Shots at 1.30 cast speed, and the GCD ends 0.2 seconds after that.
This gives you 2.8 seconds of cast time, not taking latency into account, which gives you 0.16 seconds of latency and reaction time with that haste and a 3.4 weapon. I would argue that this is not a possible scenario to sustain without causing Auto Shot delay.
I guess my real question is, even if Auto Shots are being delayed by some theoretically measurable amount each time, is this still a net DPS gain for that rotation? It certainly gives you more room to work with the GCD if you are adding rotation segments such as:
Auto - Steady - Arcane - Auto
Auto - Steady - Multi - Auto
[Edit] Answered my question already. Is this possible to plug into a spreadsheet, though?
Originally Posted by Apate
Yeah, I'm barely OK with myself being in the room while I have sex
I hate delaying autoshots as much as I can with my 3.0 speed weapon (2.61 after quiver? I'd need to log in to check). Steady -> Arcane (so the GCD goes through releasing the auto) -> Auto ->Steady -> Auto -> Steady2x and Arcane should be ready again or use Multi.
If they made the class work like Rogues or Warriors we wouldn't even have to worry about stuff like this and could just be brainless 1 button mashers.
As far as gear goes, someone made a comment about not wearing DS at 70. I'm still wearing 5 pieces for the pet bonus and because I can't find stuff with Crit+Hit+Agi+AP on it (I don't have T3 since I switched mains after Naxx). I've been picking up what pieces of the Desolation and Beast Lord set I can, but they don't combine enough of what I'm losing for what I potentially get.
My experience of it is that the only place your reaction time comes in is firing the first of those two Steady Shots. This is no more demanding than the timing of Aimed Shot many hunters are used to pre-1.13. The additional possible delay is that even though all you have to do to get the 2nd steady is spam the button while the first casts, latency will cause some delay between the two shots.
If you are going to be weaving two steadies and attempting to delay an auto shot by as a little as possible, mashing the button for your second shot is far from ideal. If you are really smashing the button you might be getting 3 presses a second, you could end up with some pretty length delays.
If your reaction times are fast enough and you have something built into your UI that is a very clear indication of when the GC finishes you are better off trying to time it on that.
My reaction times based on that test are right around 0.15 seconds. So I'm looking at about 0.3 seconds of delay from pressing buttons, I don't think latency will be a factor because GC is the limiting factor and that is all client side.
Once you sort that out I think the calculation is pretty simple:
D = Total reaction delay
Avg = Average shot damage
Spd = Modified weapon speed
Auto Shot DPS loss = Avg * [(3 + D) - Spd]
Steady Shot DPS gain = Avg * [(3 + D) - Spd]
Edit - That really is best case though. In practice I would expect a larger delay.
Targeting at 21/40/0 spcec - 21/36/0 atm, wolf pet lvl65.
1444 rap, 22,50%crit - Hemet Elek gun. (237dps paperdoll)
I found using KC is actually not bad. At least at solo grind.
Usual strat.
Aimed + send pet IF aimed crit (1.7-2.1k) use Intimidation else arcane, auto, steady, auto, steady, auto, arcane... pet (670ap) holds aggro on everything unless his growl resists or i really get crit streak (2-3x 700-1,5k crits)
Now with KC... i found pet is able to hold aggro even trough crit streak or Bladefist+Rapid combo that until now almost always dealt bit more than pet could handle. Also its good filler at kiting, where you don't have time or mana for steady/multi. (was doing Nesingway Masteries and demonhunter chain... arcane/auto/serpent/immolation/KC... & 38k hp legionares
For someone seeking bit of challenge... Solo Mo'arg Master Planner (3rd quest for demonhunter in nangard))
Neat thing about KC... it goes well with "..auto, steady+kc, auto, random special.." if you don't have "good speed" weapon. (meaning 2.4-3.0/3.2).
Hitting 1.4k steady crit+660KC crit in 0.1sec frame sure was the thing of the day yesterday for me...
Last thing... wolf and hunter aggro generating issues...
You know when pet starts tanking 3 mobs and suddenly 2 go for you. Well Furious Howl fixes this. :)
Also, not clobbering an auto shot is almost always the best thing to do and you can clobber one with a multi shot. It has that half second hidden casting time and depending on weapon speed and global cools you can't say you get one multi off every 10 seconds. Actually even in the old rotations its 10.5, the cool doesn't start till the hidden cast finishes.
Is this a fact? Why would Multishot be the only shot with casting time which delays the start of the GCD by the castlength?
If you would say that this "newly displayed but always existent" cast time of multi prolongs the GCD, then the same must be true for autoshot?
I'm sure there is a reason this isnt done, or something I'm missing about it at least, but what are the cons to removing your quiver to get a more ideal steady rotation?
Edit: Or wearing a slower one, inventory space permitting.
I'm sure there is a reason this isnt done, or something I'm missing about it at least, but what are the cons to removing your quiver to get a more ideal steady rotation?
Edit: Or wearing a slower one, inventory space permitting.
Well, the most obvious reason would be that you lose 15% autoshot damage.
Technically, if you are firing two steady shots between an autoshot every time and, by doing so, you always delay the autoshot such that the haste effect of a quiver is negated, you could remove the quiver and not effect your DPS. However, the quiver is never detrimental; you are never going to lose DPS because it is on. Furthermore, the ideal situation is not realistic, and there will always be times when it will not be possible to use two steady-shots in a row because you switch targets, have to move, run out of mana, etc. In every realistic situation, removing your quiver will cause one or more shots to be slower than necessary, because you will be unable to fire the two steady shots in between, and that will be a loss in DPS. Besides which, you need a place to hold ammo anyway.
If you are going to be weaving two steadies and attempting to delay an auto shot by as a little as possible, mashing the button for your second shot is far from ideal. If you are really smashing the button you might be getting 3 presses a second, you could end up with some pretty length delays.
...
My reaction times based on that test are right around 0.15 seconds. So I'm looking at about 0.3 seconds of delay from pressing buttons, I don't think latency will be a factor because GC is the limiting factor and that is all client side.
Even if you're only hitting the button 3 times a second, that's a max delay of .33 seconds, and an average delay of .16 seconds -- about the same as your reaction time. Only hitting a button 3 times a second is incredibly slow, though. Based on a quick test, I can do ~7 presses per second comfortably, which is a max delay of .14 seconds.
Technically, if you are firing two steady shots between an autoshot every time and, by doing so, you always delay the autoshot such that the haste effect of a quiver is negated, you could remove the quiver and not effect your DPS. However, the quiver is never detrimental; you are never going to lose DPS because it is on. Furthermore, the ideal situation is not realistic, and there will always be times when it will not be possible to use two steady-shots in a row because you switch targets, have to move, run out of mana, etc. In every realistic situation, removing your quiver will cause one or more shots to be slower than necessary, because you will be unable to fire the two steady shots in between, and that will be a loss in DPS. Besides which, you need a place to hold ammo anyway.
Wow, momentary lapse on my part. Guess I should post after I drink my coffee, not before.
I apologize if this has been discussed before but I couldn't locate it with some searchs but is ranged attack speed being calculated incorrectly? I just got Hemet's Elekk Gun which has a 2.50 speed. I have 3/3 Serpents Swiftness which should increase that by 20% giving me an attack speed of exactly 2.0 (I'm testing without an ammo bag). However the tooltip lists an attack speed of 2.08. Simple math makes that out to be 16.8% increase in speed. Is this correct? Did I miss something somewhere?
I ask this because I also noticed that with 3/3 Serpents' Swiftness and the 15% quiver my XBow of Smiting has a 2.46 attack speed and I can't figure out for the life of me how to get that number. If I use 16.8% instead of 20% for SS I get real close. Otherwise all of my other calcs have it coming in around 2.3 speed.
I believe that's how hastes are calculated now, instead of the "sum haste percentages then divide" they had before).
Ah that makes more sense, just seems a very bizzare way to do it as I was doing 2.5 * .8 which gives you 2.0. Which seems like the more correct answer than doing 2.5/1.2 since doing the division makes all of the numbers larger.
100% haste is possible in the game through several means. Using your method, that would let you empty your entire quiver in a single attack. Using the method Blizzard uses, it means you make 100% more attacks in the same time period. IMO the second makes much more sense.
It is a subtle difference between "increases your attack speed by 20%" and "decreases the delay between attacks by 20%".
With a 2.5 speed bow your number of attacks per second is 0.4. A 20% increase in that attack rate gives you 0.48 attacks per second.
A 2.08 attack speed gives you 0.481 attacks per second.
A 2.0 attack speed gives you 0.5 attacks per second.
Blizzard's tuning spreadsheets must be based around attacks per second, and average damage per attack to make DPS calcs easy. So we get "increases your attack speed by 20%".
Serious button smashers use scroll wheel. Reaction time plays very little part in my steady shot timing. It is bound to mhweeldown which I can press about 20 times a second if I am really going at it.
Clever idea, but enjoy your repetitive stress injury :)
Speaking of stupid game design ... is there any excuse for blizzard not supporting "cast this spell immediately after the previous finishes" (preferably with support for auto shot itself)?
Sure, not providing it means that only the hardcore will approach optimal dps, but I'd rather distinguish myself from my fellow players via intelligent play, not Decathlon keyspamming.
Finally: it's blizzard-approed to use nonstandard keyboard/gamepad controllers with drivers that can be configured to present keystrokes (think about your 7th mouse button ... you can configure it to send a keystroke and bind that). What about the same software, but triggering off a regular keyboard/mouse press rather than whatever gamepad/controller?
Or, screw legality. Can anybody recommend such software?
Finally: it's blizzard-approed to use nonstandard keyboard/gamepad controllers with drivers that can be configured to present keystrokes (think about your 7th mouse button ... you can configure it to send a keystroke and bind that). What about the same software, but triggering off a regular keyboard/mouse press rather than whatever gamepad/controller?
Or, screw legality. Can anybody recommend such software?
There are a bunch of gaming keyboards that can do this. If you want something on the extreme end of customization you could go with a DX1 - http://www.thinkgeek.com/computing/input/77ba/
Also, not clobbering an auto shot is almost always the best thing to do and you can clobber one with a multi shot. It has that half second hidden casting time and depending on weapon speed and global cools you can't say you get one multi off every 10 seconds. Actually even in the old rotations its 10.5, the cool doesn't start till the hidden cast finishes.
Is this a fact? Why would Multishot be the only shot with casting time which delays the start of the GCD by the castlength?
If you would say that this "newly displayed but always existent" cast time of multi prolongs the GCD, then the same must be true for autoshot?
Sorry I guess I wasn't clear. The half second cast of multi-shot delays the 10 second cool by 0.5 seconds. So the fastest you can fire a multi-shot is once every 10.5 seconds. It has no special effect on the GC.
Feign Death really need to be updated to show a pop-up above a given target if it resists it. In a multi-target environment, especially ones where several targets share a name, it's the only way of distinguising who resisted and who did not. Combat log text would not help with identical names, and manually scanning of combat log is often much worse than simply looking around you.
Edit: In other news, I need to get an avatar.
Editx2:
Originally Posted by alienangel
I believe that's how hastes are calculated now, instead of the "sum haste percentages then divide" they had before).
As far as I know ranged hastes were never any different than what they are now.
Look, Lactose, we'd rather you didn't eradicate the whole human race.
- Sam & Max
If your reaction times are fast enough and you have something built into your UI that is a very clear indication of when the GC finishes you are better off trying to time it on that.
My reaction times based on that test are right around 0.15 seconds. So I'm looking at about 0.3 seconds of delay from pressing buttons, I don't think latency will be a factor because GC is the limiting factor and that is all client side.
Once you sort that out I think the calculation is pretty simple:
D = Total reaction delay
Avg = Average shot damage
Spd = Modified weapon speed
Auto Shot DPS loss = Avg * [(3 + D) - Spd]
Steady Shot DPS gain = Avg * [(3 + D) - Spd]
Edit - That really is best case though. In practice I would expect a larger delay.
This reaction time measure is infact rather irrelevent in the situation you are speaking about, as this is reaction to a randomly occuring event. What is actually happening is a timing issue for an expected event, not a reaction time issue, the precursing events always take the same amount of time. It's more akin to keeping a beat in music or staying in time with a metronome, and as such I would expect in a focused situation to be achieving under 0.1, which isn't really too hard. I don't have a way of proving this but it could be easily done with an application, however I don't have the time to make one, but it's a fact.
I believe that's how hastes are calculated now, instead of the "sum haste percentages then divide" they had before).
As far as I know ranged hastes were never any different than what they are now.
I can swear to having a memory of a blue post a long time ago explaining how quiver haste is calculated, probably around the time they fixed the "quiver stacking with ammo bag" and "quiver providing double haste" problems. I remember it being something of a milestone, after months of whining about the lack of Dev attention on the forums, someone who was apparently a dev instead of a CM showed up on the forums and answered some questions. The formula was along the lines of:
Speed / (1 +(sum of hastes)/100)
I can't offer any proof of this post sadly, or that I'm actually remembering the right formula :(.
01/25/07, 6:46 PM
Similar Threads
