Thursday (25 Mar) -The Mechanics of DPSing

Thursday (25 Mar) -The Mechanics of DPSing

Wild is in a rut. As a healer his gear is good but unbalanced, overstacked in some stats and lacking in others. As a DPS it's the mechanics of targeting that is holding him back.

This is a self-assessment on where Wild is with his healing and DPS. There was no actual in game play on Thursday. I'll try to make that up to ya'll over the weekend.

Yesterday I did an analysis of Wild's healing compared to the other druid healer in the raid (Mg). I surmised that Wild's lack of +haste on his gear was a major factor keeping Wild from reaching that next level of healing. However, I wasn't completely satisfied with that result, so I also dove under the hood and looked hard at the healing numbers put up by Wild and Mg, spell by spell. The numbers were a bit startling, and slightly disheartening in that Wild is even farther behind than I thought.

Heal over time (HoT) spells are the bread and butter spells of a druid healer. So I took a look at how much healing each HoT did, on average, over the course of the raid on Wednesday night. I also looked at how many "tics" each spell got off to further assess just how much affect the haste difference was making. Note that "tics" are the number of times a HoT dropped a heal on a target over the course of the time from cast to when the spell expired or was over-written.

Rejuv:
Wild: 1781 tics, averaging 2305 healing per tic
Mg: 2227 tics, 2590 healing per tic

Wild Growth (AoE healing up to 6 raiders)
Wild: 2728 tics, 897 hpt
Mg: 2694 tics, 968 hpt

Lifebloom (HoT part only):
Wild: 773 tics, 761 hpt
Mg: 1924 tics, 1645 hpt

Regrowth (HoT part only):
Wild: 175 tics, 1351 hpt
Mg: 283 tics, 2333 hpt

Some of the count/tic differences come from Mg doing mostly tank healing (which will increase the lifebloom counts) and Wild doing mostly raid healing (which should raise his wild growth counts). But while Mg did way outheal Wild with lifebloom as expected, Wild was only barely able to outcast Mg with wild growth. The rejuv numbers were also telling because Wild pretty much spammed rejuv continuously the whole night and still couldn't overtake Mg. Note, also, that on every single spell, on each tic of the HoT, Mg's spells healed for more than Wild's (largely due to Mg's higher spell power). Even if Wild had the same +haste as Mg, and got off as many casts, Wild would still fall behind Mg in total healing.

Wild's main hand/off hand combination severely hurts Wild's spell power numbers, and Wild has had to make the decision to go back to his staff for healing until I can get a main hand weapon that does not further gimp Wild. That nice i264 off hand is going to sit in the bank until then. I have a solution for that, but Wild is going to need a little guild help to get it done. Since Wild finds it hard to ask his guild for help, I'm saying that here so ya'll will bug me about it. :-)

Changing the subject to DPS, I'm trying an experiment with Wild to see if I can improve on his DPS in moonkin form. There are a lot of factors that go into DPS, from gear to type of boss encounter, to what other non-DPS things a player has to do, etc. Druid moonkin DPS has an additional factor in that a lot of our damage is subject to randomness, which causes that DPS number to fluctuate quite a bit, even on multiple tries against the same boss where most of the conditions are exactly the same. That said, I think Wild's DPS is too low for the level of gear he has, and I think it's because I'm too slow getting those spells to the target(s).

Moonkin follow a basic boss encounter attack pattern that starts with two instant spells, Moonfire and Insect Swarm. These two spells are DoTs (damage over time) that tic for 14-15 seconds, and must always be refreshed, ideally right at the point the tics stop from the last cast. With those two spells active, a moonkin then starts spamming either Wrath or Starfire until Eclipse procs. Eclipse is an odd thing. If Wild is spamming Wrath, Eclipse buffs Starfire for 15 seconds with additional spell power. Obviously, Wild immediately switches to Starfire to take advantage of the buff. If Eclipse procs while casting Starfire, then it buffs Wrath for 15 seconds with a better crit percentage. So, moonkin spams the one spell until Eclipse procs, then switches to the other spell until the next proc. It sounds simple and it pretty much is.

The randomness is that Eclipse only procs if a hit creates a critical hit, which is a random event itself. Also note that the increased crit percent of an Eclipse Wrath is just an "opportunity" for more damage, as the crit might not even occur. If Wild gets a lot of crits, proccing Eclipse, then his DPS will be much higher than if he is unlucky and gets fewer procs of Eclipse.

Wild can't do anything about the randomness of his major source of damage, but I'm hoping I can do something about how quickly I get those spells on the target, increasing my chances of getting an Eclipse proc and also raising the base level of damage I'm doing when it doesn't.

Oddly enough, the best place to do Wild's experiment is not in a raid, but in heroic dungeons, where Wild pretty much ignores both Eclipse and the spell Starfire. There are three reasons for this: bosses die fast in heroics; wrath does more damage than starfire in shorter fights where mana isn't an issue; and starfire takes too long to cast to be effective in shorter fights. So Wild spams wrath full time (in addition to his DoTs, of course). When Wild does take advantage of Eclipse, Wild is pretty quick about switching spells, so I'm not worried about that part, anyway.

What I am worried about is getting faster at targeting and getting spells off. Sinstar, and expert moonkin, would probably be appalled at how Wild operates as a moonkin.

Wild first sets the tank as his focus. What this does is it lets Wild see who the tank is targeting, and that then becomes Wild's target. I use the mouse to click on the tank's target, and when the fight starts Wild starts his spell rotation. Wild actually clicks on each spell with the mouse to cast it, a process that is slower than button casting and prone to error if the mouse gets out of line while I'm looking elsewhere or moving. Not only might I miss a spell cast, I have more than once accidentally untargeted the mob and had to target him again. It gets worse after that first mob goes down, because Wild waits for the tank to pick up the next target, and the process begins again. Sometimes clicking the target fails, and Wild loses spell casts while re-targeting. Some tanks stay on one target until dead, others tab through targets (when there are more than one). If a tank changes targets mid-stream, Wild might find himself pounding a mob the tank is not focusing on, and Wild will draw aggro from that mob, irritating the tank and forcing Wild to react, again costing him spell casts. On boss encounters where there are adds that must be switched to depending on the situation, targeting and downing them takes a lot more time than it should. Those situations are a little harder to deal with, and so Wild needs a different approach for that as well.

So. The first priority is for Wild to force himself to become a button pushing spell caster instead of a mouse clicking spell caster. I've tried that before, but since I was still mouse clicking the target first, it didn't help much, plus my heart just wasn't in it (it's that healer mentality). So the second thing I needed to do was eliminate as much of the mouse targeting as I could.

I set Wild up with a small 3x3 button grid, the same grid Wild uses for healing, but obviously with DPS spells instead of healing ones. I made a macro command with the line "/assist focus." What this does is automatically target the tank's target without Wild having to click on it. I originally thought that even if the tank changes his target, Wild's spells with hit whatever his current target is. But I asked that question on the wowhead forums, and had an excellent discussion of that macro in particular and about DPS targeting in general. The macro sets the target when the macro is run. If the tank, or Wild, changes targets, the macro would have to be run again to target the tank's target again. That actually works better for Wild.

Using the assist tank macro should address about 70% of Wild's problems with targeting. Once Wild forces himself to become proficient at bashing buttons to cast spells (so that it is second nature, like his healing), it will also speed up his casting.

The two situations where tank assist targeting doesn't help is when Wild has to switch to a target the tank isn't targeting, such as to kill an add, or in the case where a tank might mark a target to be killed first and which might not be his main target.

Sinstar is raising his hand and yelling at Wild, "TAB targeting!" Yep, DPSers use the Tab key to switch hostile targets. It's essential for DPSers who DoT up multiple targets as part of their routine. If/when Philly goes into a dungeon as a shadow priest, that's exactly what she should be doing. Wild complained that Tabbing was awkward because, as a left-handed mouse user, I couldn't use the Tab key effectively, since it was on the left side of the keyboard. I solved that by binding the tab command to a key on the right side of the keyboard.

But Wild is a fail at using tab targeting. Hitting the tab key targets the nearest enemy, and every time you hit tab again it targets the next nearest target, cycling through however many hostile mobs are in range. But for Wild this is fraught with peril. Sometimes nothing seems to happen (ie, tabbing didn't change the target), and other times it may target a mob we haven't even attacked yet. Want to make a group really mad at you? Target and a hit a mob that draws aggro from mobs we aren't ready for yet. And why does it fail to target the mob right in front of you more times than not? It's so frustrating it takes me less total time to use mouse targeting, as bad as that is.

How do DPSers know what the heck they're hitting when they tab target? Do you just get really quick at identifying the mob or something? I wish Wild could have a "raid screen" that showed all of the in range enemy mobs, similar to the raid screen Wild uses to target raiders to be healed. But there is no such thing. I checked, again on the wowhead forums, and confirmed that all DPSers were in the same boat. Blizzard does not want us to be able click target enemy mobs from a raid screen, and therefore it can't be done. The reason has more to do with pvp, though, in my opinion. Imagine being in Wintergrasp and being able to tell an entire raid to target and focus fire on a single enemy player from a raid screen. Devastating.

Sometimes I think I should set focus to one of the other DPSers in the group, and follow the lead of that one. Two problems with that. One, there is a longer delay in picking up the target, as in most situations the tank targets first, then DPS targets his target. Wild would have to wait through both those actions before he picked up the target from the DPSer. The second problem is that the DPSer I pick might be worse off than Wild. Although maybe that would be a stretch.

I got one other bit of advice from the forum folks. There is a setting that makes the nameplate of mobs visible. Where there are a lot of mobs this can get very confusing and crowd the screen, blocking the view of the fight. But it's a lot easier to find and click on a nameplate than it is to click on the body of a mob. I just might try that.

Any advice?

I'll let ya'll know how Wild's "70% solution" works out.