Acknowledging that this is in part an old topic, in browsing the new mail and leather items from 2.1 rep and Black Temple, I've noticed a "new" component to itemization: Haste Rating. Now, I know its not new...but the fact that it is a constant effect rather than something you trigger (as with Abacus) or wait to proc (as IAoH and the new meta) means it can be used to simply...increase the speed of your bow by a fixed amount. The advantage of this is that it can be used to plan your "standard" shot rotation.
Were someone to systematically collect all these (after waiting for rogues, etc, as politically necessary) and ASSUMING that these haste ratings all really do apply to ranged (as reported by wowhead) the question will eventually come up:
For each of the given standard shot rotations for the various trees...what is the IDEAL bowspeed?
#1 - Beastmastery. Auto-steady rotation
#2 - Marksman/survival. Full Rotation with threaded shots
I'd love to know if people have experienced a "delay" instituting Kill Command as well...that would require a slower bowspeed to fully take advantage of.
So what bowspeed is the fastest each rotation works at without clipping? Since I've read conflicting reports, I thought I'd simply ask the question.
I feel like this is a relevant discussion again since these items are essentially a "volume knob" for your bowspeeds. Which to me is an elegant solution to the "do I give up the higher dps bow for a cleaner shot rotation" dilemna. With these items you could scale your speed up accordingly, if you wanted to.
As I understand it (and please correct my understanding if it is flawed) 10.53 haste rating equates to 1% haste. So, these items all equipped would add 122 Haste rating, or 11.59% haste. As I've seen it calculated, a 2.9 speed bow with 15% quiver and these items would therefore yield a 2.9/(1.15*1.1159) = 2.26 speed for a marksman/survivor.
While I know this is incorporated into spreadsheets...where the rubber meets the road is in our shot rotations. So before one can intelligently approach the question of "are these items worth the stats you'd give up" one must fully understand the sweet spots in bowspeeds for shot rotation.
As a thought for later in the discussion...it occurs to me that the fact that beastmasters have by far the most efficient rotations at this moment, but these items (esp mixed with rapid fire/bloodlust/etc) may generate speeds "too fast" for them. A marksman or a high dps survivor hunter (if you believe in such a creature) with a 2.26 bowspeed...may eventually give the BM a run for his money on the dps charts.
But this is speculation at this point.
Anyone with a wicked knowledge of shot rotations want to help me with this?
And of course...I'm sure there are errors in the above math. Please help me get it right.
As I understand it (and please correct my understanding if it is flawed) 10.53 haste rating equates to 1% haste. So, these items all equipped would add 122 Haste rating, or 11.59% haste. As I've seen it calculated, a 2.9 speed bow with 15% quiver and these items would therefore yield a 2.9/(1.15*1.1159) = 2.26 speed for a marksman/survivor.
I seem to remember each haste item being converted and calculated seperately (multiplicatively per item instead of additively). I don't have any items with which to verify this, though.
Combat log stamps have shown Kill Command to be able to delay Auto Shot.
Generalizing which Hasted Speed is the best is impossible, especially with such high amounts of haste. Latency and reaction time is too important (and individual) to really make any broad statements, in my opinion. This is exasperated by the non-constant hastes, some of which you mentioned, basically screwing over most x speed is better than y stuff.
Look, Lactose, we'd rather you didn't eradicate the whole human race.
- Sam & Max
The problem with haste effects for hunters is that when we use a special shot we get a CD on 0.5 second before autofire can fire and this isn't scaling with haste effects.
I think most will say that 2.7 or 2.8 are the ideal speed for a BM atm leaving a little room for lag and such ( no matter if you use a macro or not its really impossible to trigger steady as soon as it should be to utilize faster weapons except on paper )
So without any kind of haste effects with a 2.8 speed weapon this looks like this:
2.8/(1.15*1.20)=2.03 weapon speed.
SS = 1.5/(1.15*1.2)=1.09 cast speed.
auto CD 0.5
leaves a total of 2.03-1.59=0.44 second to lag etc.
Now haste it with IAotH+rapid fire:
2.8/(1.15*1.2*1.15*1.4)=1.03
SS=1.5/(1.15*1.2*1.15*1.4)=0.68
Allready here we dont get the full gain from haste since 1.03-0.68=0.35 seconds, so the haste benefit have to "wait" for 0.15 second and thats before even considering latency and perfect triggering of steady.
When I was MM I found that when I used rapid fire I gained the best result by simply using my SS macro and it kinda makes sense with the above calculations.
So while haste effects on gear could be nice for MM/surv its incredibly bad design for hunters from my point of view when our mechanics work as they do.
2.9/1.214 = 2.39
SS = 1.5/1.214 = 1.26 cast speed
Auto cd .5
2.39 - 1.26 = 1.13
Which is still plenty of time to add in special shots as they come up.
So we've gotten a 5% speed increase on our auto-steady rotation (which at last peek accounted for about 80% of my dps...that about what everyone else gets?)
Increase speed by 5% on 80% of dps = roughly 4% dps increase.
What we've essentially done here is make our 2.9 speed bow into a (2.39 * 1.15) = 2.75 bow. Without losing, it appears, anything in the process.
Perhaps this is blizzard's answer to our complaints about slow bows...they've given us volume control knobs.
Anyone have a problem with any of this math before I take it to the next level?
Doesn't Steady Shot activate the 1.5s GCD? If that's the case, you're never going to get the effective time to cast (i.e. time in which you can't do anything else) under 1.5 seconds (the notable exception being while using heroism, which decreases the GCD by the same amount as the haste effect). Using your first example,
2.9/1.214 = 2.39
SS = 1.5/1.214 = 1.26 cast speed (1.5 effective time to cast)
Auto cd .5
2.39 - 1.50 = 0.89
Which is enough time to insert a shot, but with such a small window you've got to be considerate of latency issues (my average raid latency is 200-300 ms, YMMV). Likewise, you have to consider the GCD effects on the following rotation block. If cast Multishot in this window, it's going to activate the GCD that will spill past my auto shot and into the next intra-auto shot window, delaying the next steady shot.
Let us assume I cast my instant shot perfectly and shot as soon as the steady shot had gone off and have the entire 0.89 seconds available after the steady shot.
1.50-0.89=0.61
I've now consumed 0.61 seconds of the 0.89 seconds I had 'spare' in the next window. This has also delayed my steady shot, such that it now starts 0.61s later than it would have. Timeline:
1.5 GCD--------------------------------|
0.89 s 'spare'--------|
auto+0.61-------|
steady/1.5 GCD------------------------|
0.28 'spare'--|
auto
Forgive me, that's sloppy but the best I can do in this medium. However, note that I've effectively reduced my window into which I can insert a shot on the following window to 0.28 seconds. Thus, you have to either skip inserting a special in the following window or run the very real risk of delaying an auto shot. You can forgo inserting a second special every other window and make this work, but then you run into cooldown issues on multi shot. You can't insert a 10 second cooldown every 4th window if the windows are smaller than 2.5 seconds. Also keep in mind that I didn't even account for latency in those calculations, which is going to cause havoc with the small windows into which you are now trying to insert shots.
At this point you're pretty much reduced to using steady/auto with arcane and multi on availability, which somewhats limits the usefulness of haste. Honestly, if you're going to go the haste route you may as well just drive your intra-auto shot window down to around 1.5 seconds and just use a priority replacement rotation, because that would PROBABLY net you a larger increase than trying to insert 2-1-2-1 specials. It's worth noting that this is essentially why serpent's swiftness is such a good talent, but I digress.
As I see it the new haste items will make it possible for MM/Surv to get down to something similar to what is happening for BM atm in regards to haste attack speed, which means that when they get to that point they will either have to use a 1 special shot between auto's or a steady+arcane or multi and delay autoshot every so often.
Yes...I agree. I feel like this disadvantages beastmasters.
Which...on one hand I don't mind. You guys don't need the help, by all appearances.
This might "look" good, but its really bad for all hunters and I have a feeling that when people start getting stacked up on haste they will start to respec away from BM ( if they are BM ofc.. ) and after a while I have a feeling that they wont see near the DPS increase that they had counted on and thats when we will see on outroar.
Unless off course someone is reading stuff like this and realises that they have to fix hunter mechanics if they want to use haste as a viable mod.
Which i suppose brings us back to my original question.
Forgetting the issue of haste for a moment...what's the fastest bowspeed a marksman would be comfortable with that would not sacrifice the ability to perform a "full" shot rotation, that is hitting arcane approximately every 6 seconds, and multi approximately every 10 seconds?
We were doing the job with our 2.8 bows...could we do it with 2.7? 2.6? Whatever the answer is...we could model taking our 2.9 and 3.0 endgame bows and scaling them down.
But I know 2.9 is too slow...we were doing full rotations with faster.
In the extreme, perhaps, we're pushing what BM's do...but I don't even want to go that far, myself. Just a few, perhaps to take the dead air out of my rotation.
But how much dead air can I afford to give up...once we know that, we can discuss the viability of such a choice.
So...no effort to undermine my current rotation. Just aiming for a higher efficiency.
Guys you don't need to add in 0.5s in your shot window like you are doing. The 0.5s starts at the beginning of a cast.
So if steadyshot takes 1.1s to cast, then the autoshot "loads" during the last 0.5s of that steadyshot cast. When the SS fires, the autoshot fires with it.
Where this 0.5 comes into play is mostly just dealing with arcane shot. It fires instantly and then a 0.5s windup comes afterwards.
So far, it doesn't seem worthwhile to use kill command much at all in a raid environment.
Imo, the best speeds are going to be the fastest that you can possibly maintain a single-shot rotation with. The fastest weapon you can auto->steady consistently will give the best dps imo. Of course you need to include quickshots imo since it is up so often...that is debatable though and it would take a lot of work to determine anything conclusively.
A 2.0 speed though would be ideal for me if you consider all of the default hastes, and not quickshots. qQickshots brings it down to 1.7 from there.
If there were weapons like these, suitable for MM and survival specs, then I think we would see them doing equal or more damage than BM is doing.
guys you don't need to add in 0.5s in your shot window like you are doing. the 0.5s starts at the beginning of a cast.
so if steadyshot takes 1.1s to cast, then the autoshot "loads" during the last 0.5s of that steadyshot cast. when the SS fires the autoshot fires with it.
Then it wouldn't possible to spam Steady Shot without getting a single Auto Shot fired, which it is.
EDIT: Also, combat logs show (again and again) ~0.5 seconds between Steady Shot hitting target and Auto Shot hitting target, regardless of hastes. Testing on basically any mob, letting your pet tank it makes this easy to verify.
Look, Lactose, we'd rather you didn't eradicate the whole human race.
- Sam & Max
which is nearly optimal for an BM, depending on your latency.
If I made any mistakes, I'd be glad if anyone could explain them for me. Thanks!
I agree...I think adding the "fixed haste" into our mechanics goes hand in hand with tight shot rotations for ALL trees. Your bow doesn't care where you put your talent points.
For me...it comes down to 2 timings. What is the fastest shot rotation you can have without clipped shots for:
1. Full rotation of steady/multi/arcane
2. Auto/Steady
Also...
3. How long does a kill command add to to either rotation?
Once we have clean numbers on this (which to my knowledge has never been conclusively determined) then we can look at the next steps.
Eventually, we'll be able to look at someone's bowspeed and say, "Ah, see...you could take on 12% haste without clipped shots in your rotation, and that'd up your dps by X%"
In the long run, this advantages BM's the least...they seem to succeed by virtue of not having much "dead air" in their shot rotations. If they get a slower bow...they can use haste items to get back to their ideal. Overdone, their rapid fire becomes useless. (which would be an interesting study...life w/o rapid fire possibly just fine?)
But take some of that air out of survival and marks rotations...and no question they'll see a huge boost.
I'm gathering data on this from various sources...if anyone has any hard numbers regarding lag or kill command, that would be particularly helpful.
I'm gathering data on this from various sources...if anyone has any hard numbers regarding lag or kill command, that would be particularly helpful.
All I've seen regarding KC timing it the time it takes to press the key, and the 0.5 second window before auto shot you want to avoid with it. Other than that it seems to be a true instant that doesn't use the GCD.
I'm new to these forums (and the hunter theorycrafting in general) so forgive me if this already have been covered elsewhere and/or if it's not completely on the topic of haste effects.
I've been playing around with some triggered haste effects and got into a discussion about Steady Shot and it's effect on Auto Shot's cooldown and if there is any.
I'm currently using Sunfury Bow of the Phoenix which with quiver gets a 2.52s speed. Using a Steady Shot/Auto Shot rotation it seems the Auto Shot shoots right after the Steady Shot. Removing my quiver and trying the same it still seems like the Auto Shot comes right after the Steady Shot.
Now the question is if Steady Shot resets the Auto Shot's cooldown or I'm too sleepy to notice the difference?
Feels like a silly question but all feedback is appriciated.
Which i suppose brings us back to my original question.
Forgetting the issue of haste for a moment...what's the fastest bowspeed a marksman would be comfortable with that would not sacrifice the ability to perform a "full" shot rotation, that is hitting arcane approximately every 6 seconds, and multi approximately every 10 seconds?
We were doing the job with our 2.8 bows...could we do it with 2.7? 2.6? Whatever the answer is...we could model taking our 2.9 and 3.0 endgame bows and scaling them down.
But I know 2.9 is too slow...we were doing full rotations with faster.
In the extreme, perhaps, we're pushing what BM's do...but I don't even want to go that far, myself. Just a few, perhaps to take the dead air out of my rotation.
But how much dead air can I afford to give up...once we know that, we can discuss the viability of such a choice.
So...no effort to undermine my current rotation. Just aiming for a higher efficiency.
Actually I do get delays for my Auto Shots even with a 2.9s Bow with my full shot rotation considering all mechanics involved and 100ms latency/reaction time. This shot rotation inserts one Arcane every other Auto Shot (with the talent to support it) and one Multi every fourth Auto Shot instead of Steady (Auto-Steady-Arcane-Auto-Multi-Auto-Steady-Arcane-Auto-Steady). The 3.1s Glad X-Bow works fine with this rotation I´m just at loss how you guys manage to use a full rotation with 2.9s Bows or even faster.
I've done some testing on the subject and came to some pretty odd conclusions: http://www.megaupload.com/?d=IAHF7OMJ My conclusions were that with only a 15% quiver haste, the ideal unhasted weapon speed is ~2.36. With serpents swiftness and/or quick shots the ideal unhasted weapon speed becomes ~2.54. Strange imo.
That file is either corrupted or a format I'm unfamiliar with - *.xlsx? Excel can't open it. Can you post your methodology here so we can discuss it? The unhasted speeds you posted seem a little low to me (rather, low if you account for variable latency and response time). More info would be great!
*.xlsx is the new format Microsoft Office saves Excel files to. Sadly, it isn't backwards compatible, so Xeno will have to explicitly save his file as a *.xls file in order for you to check it out. Or you could go buy the new Microsoft Office =p.
Ah ok, thanks for the heads up. I'm at work now, alas I have no access to the new excel (unfortunately, I don't have it at home either, but at least I have the option of upgrading my software there). Sorry for the confusion!
1. What your ping like?
2. Have you run this with a full rotation, ie, auto/steady/multi/auto/steady/auto/steady/arcane/auto/steady ?
While I certainly do have a lot of interest in beastmaster speeds, I do feel like this study is ultimately going to be of more use to marksman and survivalists. I feel like they have further to go before they hit a speed which is on the verge of "too fast".
By knowing the benchmarks (and testing them for different ping rates, etc) we can really plan which bow/spec combinations bear improvement through haste effects. And really nail down which procs will force you to adjust your shot rotation.
I am definitely confused by the fact that an auto/steady speed study came out with two different conclusions with or without serpents swiftness...unless gcd's and cast times are affected differently by different haste effects/talents...seems like the steady shot speed barrier should be the same regardless of if you got there through gear or through talents.
When you say steady doesn't reset auto's CD, you're talking about the hasted bowspeed between-auto time, and not the ~.5 sec cast time of auto shot itself, right?
Or are you saying that auto shot in fact always comes out with steady, with no .5 sec interval? If true, I should never be able to move (or esc) to interrupt the following auto but not the steady, assuming my between-autoshots timer is ready (e.g. opening with steady). Also, if I get a hasted speed slightly faster than 1.5s, I should be able to spam steady shot with autos always getting through.
My understanding was that there's a .5 sec gap between steady firing and auto (if ready).
Ping is 100-150 ms. Unfortunately I haven't looked into a rotation with multi and arcane, I might get around to it eventually, but it'd require some respecs.
As for the different ideal speeds, I've been pretty much stumped by what could be causing it too. It's one or more of the following: 1. An error in testing or parsing on my end, it'd have to be pretty fundamental. 2. Cast time's are somehow affected differently by different types of haste. (This is what I figured it might be and that's why I did the tests in the first place.) 3. GCD. 4. Latency!
05/29/07, 3:22 PM
Similar Threads
