In this article I am analyzing the Marksmanship hunter rotation. I will demonstrate how we develop the optimal rotation from a very basic rotation. Also, I will quantify the value of different rotation optimizations.
For a graphical version of some of the rotation tests, scroll down to the “4T20 simulation results + a look at legendaries” section, or click here.
Optimal DPS is rarely determined by a player’s ability to perform complex rotation logic - it is most often determined by their ability to Always Be Casting (ABC). Marksmanship hunters essentially play like a ranged spell caster, constantly using a spell with a significant cast time, Aimed Shot, for the majority of their damage. To put this article’s analysis into context, one Aimed Shot is worth about 0.5% of your total DPS in a 5 minute fight. In shorter fights, it is worth proportionally more. What does this mean? Planning ahead and avoiding any interruption of your Aimed Shot casts is the most important thing you can do to increase your DPS. Excessive movement can also cause you to cap focus or waste time during the vulnerable window, also losing out on Aimed Shot casts.
I mention this because one of my underlying points in this and similar articles is that the WoW theorycraft community has become obsessed with simulation to the point that the value of small rotation complexities is vastly overstated. Almost every spec in the game can be played with a very basic rotation. If you execute the basic rotation perfectly and maximize your actions during the fight due to solid fight mechanics, you will do competitive DPS, even at the mythic raiding level. Just because we can make a simulator do 2% extra DPS with some clever programming doesn’t mean that all good players must play like that to do high damage.
For the rotation analysis, I will start by using a character with no set bonuses or legendary items and a stat distribution that is typical of what MM hunters would prefer: A lot of mastery and crit and some haste. After we find the optimal base rotation, I will add in set bonuses and legendary items that impact the rotation and analyze tweaks related to those effects.
First, we look at the most common single target spec right now. The relevant talents from a rotational point of view are Lone Wolf, True Aim (often Lock and Load is available too from Soul of the Huntmaster), Patient Sniper, A Murder of Crows, and Trick Shot.
At the most basic level, this is how you play a MM hunter:
1.) Use Trueshot on cooldown.
1a.) Use Potion, Trinkets and Racials when Trueshot is up.
2.) Use Aimed Shot if Vulnerable is up and will not drop off before you finish casting.
3.) Use A Murder of Crows on cooldown.
4.) Use Windburst on cooldown.
5.) Use Marked Shot.
6.) Use Arcane Shot.
Using this rotation, we observe 1210k DPS.
Add Pool for Marked Shot Logic
Now we will explore and quantify different rotational optimizations. The DPS reported is without tier 20 equipped, and later in the article we look at T20 simulations.
1.) Since our main damage source is Aimed Shot, we want to maximize the Aimed Shots being cast in the vulnerable window. This also means that we don’t want to spend Focus outside of the vulnerable window unless absolutely necessary. We want to save some focus before casting Marked Shot. But how much? We can do a test using a Rotation Parameter. I tested pooling from 50 to 120 focus in increments of 5:
Pooling for Marked Shot Test
We see that 60-70 focus pooled comes out as best*:
Pooling before Marked Shot, we increase DPS by 2.5% to 1241k DPS.
*There is a formula that can be used to test if the difference between two simulation results is statistically significant. When looking at a batch report, click on one row, and then click on a second row you want to compare to. This pops up a window that shows you the calculation. If the difference range crosses zero, the difference between the two simulation results is not statistically significant, by definition. This is helpful to compare results when you aren’t sure if the difference is due to margin of error or some actually significant result.
Add 1 Arcane Shot after Marked Shot
2.) Patient Sniper causes the Vulnerable debuff on the target to ramp up in value over the 7 second duration. We can’t spam Aimed Shot for the entire Vulnerable window so it makes sense to try to push our Aimed Shots that we do cast towards the back end of the debuff. One easy way to do this is to always cast an Arcane Shot right after Marked Shot. It is common to see guides recommend that you cast 1x Arcane Shot followed by 2x Aimed Shot after Marked Shot. But, this may affect how much Focus we want to pool before Marked Shot as well, so, we need to redo our focus test from the Marked Shot rotation tweak:
Pooling for Marked Shot Re-Test
We see that 65-70 focus pooled comes out best. We will use 65 as our number for now:
Adding an Arcane Shot after Marked Shot, we see DPS increase about 0.75% to 1250k DPS.
Add optimization for cooldowns
3.) Marksman Hunters have a substantial execute phase when targets are below 20% health because of the Bullseye artifact trait. For this reason, we want to try to save our last use of long cooldowns like potions, trinkets, racials, and Trueshot to be near the end of the fight, but without losing out on any uses. It can be hard to judge exactly when the fight will end in-game, but, with practice you can make an educated guess.
Saving last cooldowns for execute increases DPS about 1% to [1262k DPS] (http://www.askmrrobot.com/wow/simulator/report/230bec9f228a4811becd031e53973ab5).
For the rest of this article, I am calling this rotation number 3 the “simple” rotation for Marksmanship Hunters. (This is the purple line in the graphic).
Optimize Patient Sniper (default AMR rotation)
4.) We could further refine the optimization from step 2. If we have high haste, sometimes we could fit in two arcane shots before our Aimed Shots. Or, maybe sometimes we could actually cast 3 Aimed Shots. This change will take us to the default rotation in the simulator which uses a variable to track how many Aimed Shots can be cast in the vulnerable window. With that information, it can optimally choose when and how many Arcane Shots to cast before we cast our Aimed Shots.
This tweak to the rotation increases DPS about 0.75% to [1273k DPS] (http://www.askmrrobot.com/wow/simulator/report/b89987f66ec14ee8844800ee145cfb29).
To do this optimization in-game, you will likely need some sort of addon, like weak auras, that you can program to track the number of Aimed Shots that can be cast in the remaining vulnerable window. As you gain Lock and Load procs, haste buffs, etc., you will need to modify how many arcane shots you cast before you cast aimed shots.
The total DPS change from our very basic rotation to an optimized rotation is 5% DPS. Simply pooling focus to 65 before using Marked Shot and always using one Arcane Shot after Marked Shot puts us within 1.5% of the optimal theoretical DPS that could be done.
A look at T20 - Simulate each rotation & report DPS
Next we will look at how the 2T20 set bonus affects the rotation. Below is a simulation where I re-ran all of our test rotations we’ve done so far. I also included the default rotation, which includes optimization of the 2T20 set bonus. We will look at different ways to optimize around the set bonus next.
[Simulation Results] (http://www.askmrrobot.com/wow/simulator/report/2af633a3ecfe44219e9a4ff47391e5b4)
Basic: 1233k DPS
Pool for Marked Shot: 1257k DPS (+1.75%)
Pool + one Arcane Shot: 1311k DPS (+4.5%)
Cooldowns: 1329k DPS (+1.25%)
Default: 1367k DPS (+2.75%)
Here we see a 10.75% difference in DPS from the very basic rotation to the optimized rotation. Once again, simply pooling 65 focus before Marked Shot and using one Arcane Shot after gets us very close to optimal, within 2.75%. The optimization for pushing Aimed Shots towards the end of the vulnerable window also serves to optimize use of the 2T20 bonus because you naturally end up casting Aimed Shots clustered close together, which ends up buffing the maximum number of Aimed Shots with the 2T20 bonus.
Testing out T20 Optimizations
5.) The Tier 20 2 piece bonus gives a crit damage buff if we cast two Aimed Shots in a row. What happens if we create a rotation that specifically focuses on 2T20 uptime? These conditions for using Aimed Shot should maximize the buff uptime - Cast Aimed Shot if:
- We just cast Aimed Shot and need to cast it again to apply the 2T20 buff.
- We have the time and focus to cast two in a row when the 2T20 buff is down and vulnerable is up.
- We already have vulnerable and the 2T20 buff up.
Otherwise, the rotation will fall through to other actions and wait until you can put the buff back up again:
We observe a slight DPS loss of about 1.5% to [1309k DPS] (http://www.askmrrobot.com/wow/simulator/report/86cbd0652ebf4c068114a01daa0bed08).
The uptime on the 2T20 buff increased, but, overall damage didn’t increase because we ended up casting a couple less Aimed Shots and a couple more Marked Shots. This rotation is trying too hard to cast Aimed Shots in pairs, resulting in less Aimed Shots, which we know is usually going to be sub-optimal. The default rotation, as I mentioned above, uses a formula to decide if we can fit in Arcane Shots before we start casting Aimed Shot.
Many guides out there offer rules of thumb to govern your play instead or a dynamic formula. Here is a compilation of all the rules from a popular guide:
- Cast 1x Arcane Shot followed by 2x Aimed Shot after saving 70 focus and casting Marked Shot
- Cast 2x Arcane Shot followed by 2x Aimed Shot after saving 50 focus and casting Windburst
- Cast 2x Aimed Shot after saving 105 focus if you cannot cast Marked Shot or Windburst
- If you have 19%+ haste and the 4T20 bonus, cast 2x Arcane Shot followed by 2x Aimed Shot after saving 50 focus for Marked Shot
- If you have 110+ focus when you cast Marked Shot, cast 3x Aimed Shot immediately after.
- Use Lock and Load procs ASAP with vulnerable up.
- Use an opener that does Murder of Crows, Windburst, Trueshot, 2x Aimed, 2x Arcane, Aimed until vulnerable ends, Arcane until 7 seconds left on Trueshot, Marked, 2x Aimed, 2x Arcane, Aimed until vulnerable ends.
- During Trueshot, use a rotation similar to the opener, but don’t cap focus.
If you are taking a detailed look at this, you will notice that the guides usually tell you to pre-cast Windburst and then pop trueshot just before it ends and then use A Murder of Crows as an opener. Simulations show that you will do just as much damage if you use A Murder of Crows right at the pull and follow it up with Trueshot/Windburst. Use whichever you prefer in-game. I use the second version in my simulations - it has no impact on this rotation analysis.
The resulting sim gets us about 0.75% more damage up to [1340k DPS] (http://www.askmrrobot.com/wow/simulator/report/c99466b100764d96bff8a52932d1ad5e).
These rules of thumb work well, but are a bit too rigid to achieve 100% optimal DPS. We end up missing out on some possible 2T20 buffed Aimed Shots. Also, we miss out on a possible Windburst use over the course of a longer fight. The reason is because this set of rules is treating Windburst similarly to Marked Shot: as an ability whose main purpose it to keep up Vulnerable. Windburst does more damage/execute time than a buffed Aimed Shot. Delaying Windburst in order to pool up some focus is going to be sub-optimal compared to using it ASAP once you can’t cast anymore Aimed Shots in the current vulnerable window. Even with haste as low as 11%, if you are standing 35+ yards from the target, you can cast Windburst, 3x Arcane, and 2x Aimed before vulnerable falls off. This allows you to use Windburst at lower focus and delay it less without losing Aimed Shots, and also makes those 2 Aimed Shots a bit stronger.
4T20 simulation results + a look at legendaries
The Tier 20 4 piece bonus is not going to have a significant impact on the rotation. We can run another batch comparing our rotation tweaks up until this point to check:
[Rotation Test for 4T20] (http://www.askmrrobot.com/wow/simulator/report/355352f42de2484c89109e4fed1ebacb)
Here is a graphic with the 4pc T20 equipped:
One of the most sought after legendary items is Soul of the Huntmaster. This will add the Lock and Load talent to our setup. If we go back to our simple rotation that just uses one Arcane Shot after Marked Shot, we see [1483k DPS] (http://www.askmrrobot.com/wow/simulator/report/e5a612e1a74c4d268940911afe55fd40). One test to do here is to see if prioritizing using Lock and Load procs with vulnerable before using that Arcane Shot makes a difference:
DPS increases about 0.5% to [1488k DPS] (http://www.askmrrobot.com/wow/simulator/report/a85ef233ab8740499ac941ce10195f9c), since we get about one more Aimed Shot.
Interestingly, we now see that the “rules of thumb” rotation does [1475k DPS] (http://www.askmrrobot.com/wow/simulator/report/fab6d1bf59e94cb2bc0fd729ce2951fa), about 1% less damage than the simple rotation.
For comparison, the fully optimized rotation gets us up to [1512k DPS] (http://www.askmrrobot.com/wow/simulator/report/a22d15f1f166448ca9f1acaa942bb2f1), 1.5% better than the simple rotation and 2.5% better than the “rules of thumb” rotation.
The most important take-away here is that adding the Soul of the Huntmaster ring to the gear set actually makes the simple rotation perform better and get very close to optimal DPS.
1.5% theoretical DPS is of very little in-game significance. If you look at the number one Marksmanship parse on mythic Goroth - that player would still have a number 1 parse if they did 1.5% less damage. This goes back to one of my main points: the theorycraft community has become obsessed with very small theoretical DPS gains in simulations. You don’t need to even think about the 2T20 set bonus. You don’t need to use a special opener. You don’t need to remember a bunch of rules for how many Arcane Shots to cast at certain points. You don’t need to think about “cycles” in your rotation. Just save up 65 focus, cast Marked Shot, one Arcane Shot, and then use as many Aimed Shots as you can before Vulnerable runs out. Using that simple logic and then focusing on perfect fight/movement mechanics, you can get world-class parses as a Marksmanship hunter. It does not need to be a complex rotation.
I added in the second most sought-after legendary item for single target fights, the MKII Gyroscopic Stabilizer and did our batch test again:
[Training Dummy] (http://www.askmrrobot.com/wow/simulator/report/8dbdb3836a6742349580aa7ec199f932)
No changes in relative values for the different rotations, as expected.
Other talent builds
I also want to take a brief look at some other talents that affect the rotation significantly.
The main question I want to explore here is if you should use Piercing Shot on cooldown, or just use it when you happen to have 100+ focus in the normal course of your rotation. If using it on cooldown, we would presumably also want to try to pool some focus in anticipation of it being ready. Of course, we only use it with vulnerable on the target.
First I did a test on our base set of gear with no legendary items or set bonuses. I tried 3 different rotations:
1.) Use Piercing Shot when we happen to have 100+ Focus with vulnerable up, [1216k DPS] (http://www.askmrrobot.com/wow/simulator/report/ef4fcb06584e4c489e1feba76446d1e0).
2.) Use Piercing Shot on cooldown, don’t use Focus when we can use Piercing Shot soon, [1225k DPS] (http://www.askmrrobot.com/wow/simulator/report/71d43c2a7d87427ca00b42b6623b5a4d).
3.) Use Piercing Shot on cooldown, pool Focus, but don’t wait for 100 focus if we can cast, [1215k DPS] (http://www.askmrrobot.com/wow/simulator/report/3c467f37393e43b4ae68f3d0fed03ed9).
And here is the test again with our 4T20 set with Soul of the Huntmaster and MKII:
1.) [1500k DPS] (http://www.askmrrobot.com/wow/simulator/report/91faacbee617458d9539e16485acdb18)
2.) [1480k DPS] (http://www.askmrrobot.com/wow/simulator/report/93b377a49f864bee9c0e0f7e7f0d6495)
3.) [1475k DPS] (http://www.askmrrobot.com/wow/simulator/report/d97b9367196a460cab4df69bccb6a53c)
This data tells us that what gear you have can affect the optimal way to play with Piercing Shot on a single target. Piercing Shot is generally much better for AoE, but, this is an interesting test to do. In the simulator, I default to using Piercing Shot more often. This is mainly because if you take Piercing Shot, you probably want to… use it.
This talent only makes sense for AoE/cleave fights, but, you may have to do a single target rotation for some portion of the fight. When using Sidewinders for AoE, you will use it if you are at max charges, or if you have a Marking Targets proc. For Single Target, a good way to use Sidewinders is to think of it the same way you would think of Marked Shot, since it applies Vulnerable. Don’t worry about Marking Target procs against a single target, since we will be following it up with Aimed Shots. Essentially we end up casting far fewer Marked Shots.
[Sidewinders Single Target, 1091k DPS] (http://www.askmrrobot.com/wow/simulator/report/c21a259c988143a4a6de0a287c7b9975)
Alternatively, we could try to incorporate the Marked Shots back into the rotation by using Sidewinders when we have Marking Targets and then cast Marked Shot after Sidewinders. We refresh the vulnerable buff right away in order to get the damage from the Marked Shots back into the rotation and see the DPS increase slightly to [1098k DPS] (http://www.askmrrobot.com/wow/simulator/report/5db9fc5afc3f4106b55bfcf259156461).
If we do the same test with our 4T20 hunter with 2 legendary items we see [1444k DPS] (http://www.askmrrobot.com/wow/simulator/report/13606ebc25ff4eed945cfcd197b23903) vs [1410k DPS] (http://www.askmrrobot.com/wow/simulator/report/de9cbb2091e2492ca9c907750bdd1726). Once we have special effects that increase the damage of Aimed Shot substantially, those extra Marked Shots in the rotation don’t end up holding their weight like they do in the weaker geared setup.
This is why my general recommendation against a single target is to use Sidewinders when they are about to hit max charges, or if you can’t cast Aimed Shot before vulnerable drops off. Don’t worry about Marked Shot.
Steady Focus and Careful Aim are so vastly inferior to Lone Wolf that they aren’t worth talking about. But, if you do take Steady Focus, you will want to make sure to cast a second Arcane Shot to activate the buff when your focus is not already high.
Black Arrow is a weak talent for raiding, but it fits easily into the rotation as a low priority action that you would take instead of using an Arcane Shot.
Explosive Shot is best for AoE, but it fits into the single target rotation easily as well as something you would use when you cannot cast Aimed Shots, Windburst, or are too low on focus to optimally cast Marked Shot.
If you take Volley, toggle it on and leave it on for the whole fight, even against a single target.
Sentinel and Barrage are both not worth using against a single target.
The AoE rotation is very simple, so I’m not going to go much into detail on it as far as analysis goes. The ability priority becomes very simple:
1.) Piercing Shot when vulnerable is up and Power >= 100
2.) A Murder of Crows on target that will die in less than 60 seconds but live longer than 8 or so.
4.) Bursting Shot (with Magnetized Blasting Cap Launcher)
5.) Multi-Shot/Sidewinders if you have Marking Targets or Trueshot and Hunter’s Mark is not on the targets.
6.) Sentinel if Hunter’s Mark is not on the targets.
7.) Marked Shot
8.) Aimed Shot if vulnerable won’t drop before you finish casting and you aren’t pooling focus for Piercing Shot or Barrage.
Marksmanship hunters can be played at the highest levels with an incredibly simple rotation. The hardest thing you will need to do is know if an Aimed Shot cast will finish before Vulnerable drops off the enemy you are attacking. If you feel that you are executing fights perfectly and want to push for a small theoretical damage increase, you can try inserting multiple Arcane Shots before you start casting Aimed Shot - but the optimal number of Arcane Shots is dynamic depending on the current situation. I have shown in this article that more complicated rules of thumb are not going to yield a DPS increase in many situations, so, beware what you read out there on the internet.
The difference between theoretically optimal and a simple rotation is so small that you actually cannot do enough fights in-game to verify what rotation would work better. Marksmanship guides out there suffer from confirmation bias - a good player tried something, they get good parses, so it must be how we all have to play! Not necessarily. Simulators are the tools needed to help us quantify what is truly critical to optimal play.
Apendix: SimC vs AMR.
The theorycraft community is bound and determined to perpetuate an online war between the two available simulators for WoW: AMR and SimC. AMR has verified continuously over the course of Legion that the AMR simulator matches very closely to real combat logs. If you compare a simulation between AMR and SimC using the exact same setup, rotation, etc., the results will be nearly identical. Usually you will observe slightly higher average damage reported in SimC due to how it is implemented (about 1%). The AMR model is more realistic, the SimC model is more theoretical, but the effects are minor. Anyone trying too hard to convince you that you have to use one simulator vs. the other definitly has an agenda besides simulation accuracy. If all you want to do is single run simulations, it really doesn’t matter which simulator you use.