While working on the MM hunter guides for the Warcraft Hunter Unions, I thought it might be fun to show on here some of the processes we go through to come up with our rotations/priority systems. Well, fun for me. Perhaps incredibly boring for others!
Anyhow, there are several ways come up with a shot priority system. The best, of course, is spending years testing out various rotations, until you have hundreds of thousands of data points on each one, and seeing which worked best. Since it's hard to do that before an expansion releases, however, we have to rely on other methods.
On of the most popular, is using coded or spreadsheet simulations. This is a way you can test out many iterations of a rotation, without needing to actually do the rotation over and over. The downside here is that we're generally working with shot co-efficients that we have reverse engineered, and may not be completely accurate. At this point, blizzard has been very gracious in recognizing the value of theorycrafters to the game, and is generally pretty forthcoming with information about values, mechanics and coefficients. Still, with all the information we have right now, you can't always 100% trust a sim, and we see some that give odd suggestions, every once in a while.
Another way I like to look at Shot Priority is through a metric I call D/F/CT. That is, Damage per focus per cast time. This is not a perfect method by any means. It struggles to compare abilities that don't have a focus cost, and it doesn't give good considerations for DoT type abilities. Still, with a little logic, we can use D/F/CT to narrow down which rotations we want to simulate, using the two methods together should result in a pretty good picture of the best priority system we can come up with.
Gathering Data
So last night, I spent a couple hours out in Shattrath killing some training dummies. And after logging and compiling all of that information, I ended up with a bunch of numbers that looked like this:
| Hit | Crit | Per cast ave: | Focus Cost | Cast/GCD | Dam/CT | Dam/Focus | D/F/CT | Focus/KDam. | |
| Kill Shot | 39595.1 | 88532.6 | 65610.4 | 0 | 1 | 65610.40 | n/a | n/a | 0.00 |
| aMoC | 2596.3 | 5754.9 | 64015.4 | 30 | 1 | 64015.38 | 2133.85 | 2133.85 | 4.69 |
| Aimed | 22491.4 | 51735.1 | 44081.6 | 10 | 2.19 | 20128.60 | 4408.16 | 2012.86 | 2.27 |
| Chim | 41515.8 | 96025.4 | 59615.4 | 35 | 1 | 59615.38 | 1703.30 | 1703.30 | 5.87 |
| Aimed | 22491.4 | 51735.1 | 44081.6 | 30 | 2.19 | 20128.60 | 1469.39 | 670.95 | 6.81 |
| Aimed | 22491.4 | 51735.1 | 44081.6 | 50 | 2.19 | 20128.60 | 881.63 | 402.57 | 11.34 |
| Barrage | 2464.4 | 5380.8 | 56666.7 | 60 | 2.63 | 21546.27 | 944.45 | 359.10 | 10.59 |
To clarify what you're seeing there, the three Aimed Shots represent the three different focus costs Aimed Shot may have, one just normally casting it, on with TotH, and one with TotH that is also a Crit (which we can count on during Careful Aim). I was using Lone Wolf, TotH, Barrage and aMoC for the tests, and had the following stats:
| Agility | 4566 |
| AP | 5023 |
| Crit | 31.97% |
| Haste | 14.12% |
| Mastery | 10.20% |
| Multistrike | 14.79% |
| Versatility | 3.59% |
Now, after compiling all of that, some of the numbers seemed off to me, so I went back, killed another target dummy, just to double check. It gave me these numbers (which aren't included in the above averages).
| Hit | Crit | Per cast ave: | Focus Cost | Cast/GCD | Dam/CT | Dam/Focus | D/F/CT | Focus/KDam. | |
| Kill Shot | 40439.2 | 90233.9 | 62857.1 | 0 | 1 | 62857.14 | n/a | n/a | 0.00 |
| aMoC | 2675.0 | 6062.8 | 63913.3 | 30 | 1 | 63913.33 | 2130.44 | 2130.44 | 4.69 |
| Aimed | 22031.9 | 54123.3 | 44081.6 | 10 | 2.19 | 20128.60 | 4408.16 | 2012.86 | 2.27 |
| Chim | 41262.2 | 96387.5 | 61521.7 | 35 | 1 | 61521.74 | 1757.76 | 1757.76 | 5.69 |
| Aimed | 22031.9 | 54123.3 | 44081.6 | 30 | 2.19 | 20128.60 | 1469.39 | 670.95 | 6.81 |
| Aimed | 22031.9 | 54123.3 | 45391.7 | 50 | 2.19 | 20726.80 | 907.83 | 414.54 | 11.02 |
| Barrage | 2391.4 | 5312.1 | 55757.6 | 60 | 2.63 | 21200.60 | 929.29 | 353.34 | 10.76 |
While the numbers weren't exactly the same (RNG makes for good times in data compilation), they do appear to have similar relative value. So I'll go ahead and say I didn't make a mistake.
There are a few bits of information I can gleam from this. The first, most glaring one, is Barrage is not useful as MM with Lone Wolf. The buff to Aimed Shot is enough that the cost of Barrage (and by cost, I mean the amount of time you'll have to spend casting Steady Shots) will make Barrage do less DPS than Aimed Shot with no procs. And since Aimed Shot has no CD, there would be no reason to ever cast Barrage. Of course, there are plenty of other reasons to use Barrage, like for bursting down adds, or AoE in general, but before I start thinking about considerations, I like to have a pure single-target priority system well defined.
One of the misleading bits about this spreadsheet (which, if it's not obvious, is sorted by D/F/CT), is that it shows a Murder of Crows as a higher priority than Chim Shot. This is tricky, because while it's true that aMoC does more damage, and is a more valuable use of your focus, the one minute CD, and the DoT effect of aMoC means we'll still consider Chim Shot a higher priority shot.
Let me clarify that. If I delay aMoC by one GCD (for Chim Shot) and a few seconds (to build up focus from Steady Shots), it's not going to change the total number of aMoC's I can cast during the fight. Because it has a relatively long CD, I can delay it a little bit without worrying about losing Damage from it. On the other hand, if I delay Chim Shot a few seconds, and then I do that a few times throughout the fight, I start lowering the total number of Chim Shots I can cast during the fight. This is what I meant, when I said we needed some basic logic to go along with the D/F/CT method.
The next anomaly I'd look at is that Aimed Shot (during Careful Aim when TotH is proc'ed) is showing up as a better use of our focus than Chim Shot. Conventional Wisdom would tell you that your signature shot is generally the most important shot. However, I have the benefit of having talked to Celestalon (systems designer for WoW) a few months ago about this topic, and he stated the devs were open to Aimed Shot being more valuable than Chim Shot during certain buffs. So I'm assuming this isn't a mistake.
Another issue, which is somewhat more expected, is that Kill Shot is crazy powerful. That's just something we'll have to get used to. Where in MoP, it was less valuable than Chim Shot, Kill Command or Explosive Shot, for WoD, it's going to be the top shot (which is sad for SV hunters). This makes sense, since it is supposed to be executing people.
It is worth noting with Kill Shot, that it's got that "N/A" in D/F/CT. That's because, unfortunately, using elementary algebra (which WoW is based on), we can't divide a number by zero. That doesn't mean that it's necessarily our most important shot. But that it's free and has the highest average damage per cast does put it to the top of our Priority List.
It is worth noting with Kill Shot, that it's got that "N/A" in D/F/CT. That's because, unfortunately, using elementary algebra (which WoW is based on), we can't divide a number by zero. That doesn't mean that it's necessarily our most important shot. But that it's free and has the highest average damage per cast does put it to the top of our Priority List.
So my next task, is to test this priority system using simulators like Wowhuntertools.com, and SimC. Of course, this doesn't cover things like when to use Rapid Fire, when to use Potions, etc. And it doesn't get into our Opener, or any of the many other things that need to be decided. But it should give anyone who's interested a look into how I decide my shot priority, when I'm playing.
No comments:
Post a Comment