I mentioned yesterday on Twitter that I would be taking some looks at Guardian DPS today. I did just that and here are my results.

Please note that all tests were conducted under the following conditions:

  • Using ungemmed, unenchanted 483 PvP gear.
  • Using Soul of the Forest and Heart of the Wild for max DPS.
  • Not using any cooldowns (Berserk, Enrage, etc.) since they are now actually defensive cooldowns.
  • Attacking the boss target dummy in Stormwind while fully armor debuffed.

0 Vengenace Bear

Here’s the 0 Vengeance Bear “control” parse. Please note that his parse was done while dumping all Rage to Maul. A rotation of Mangle > Lacerate > Thrash Upkeep > FFF was used. Considering my previous Guardian 0 Vengeance parses were in the neighbourhood of 25k, this seems much more reasonable for a tank. Damage is also much more evenly distributed across the top 3 abilities (Mangle, Lacerate, Melee).

Guardian Cat

Here’s the Guardian Cat “control” parse. Note that very few (if any) Ferocious Bites were actually used. Due to the terrible combo-point generation in this gear, a huge portion of the potential DPS is sacked in favour of maintaining Rip uptime. Once our Haste and Crit improves, it’s very likely that Guardian Cat DPS will see a sudden jump once we can reliably generate enough combo-points to throw in a Ferocious Bite every Rip cycle.

In any case you can clearly see from the parse that Guardian Cat DPS is currently about 20% higher than 0 Vengeance Bear. This means that once you dip below say 30% Vengeance on a tank swap, you’ll want to jump into Cat form for the highest DPS. Of course this will vary from encounter to encounter. Fights with very quick tank swaps (like Ultraxion) there won’t be any point since you’ll have to taunt right back in 7 seconds or less anyway. Additionally fights with large amounts of raid damage (Spine, Madness) will keep your Vengeance high enough when not actively being melee attacked.

In order to optimize your DPS as a Guardian, having a mod that keeps track of your Vengeance is going to be almost essential. If you can find one that will let you set warnings at certain values, that’s even better. I don’t know if such a thing exists though. If it does please tell me.

Heart of the Wild: Guardian -> Cat

Unsurprisingly nothing has really changed with Heart of the Wild. I ran 5 individual parses, all linked below. In each parse I edited the combat log such that the first entry was activating Heart of the Wild, and the last was it fading. This was done to skew the results in favour of HotW as much as possible.

Max DPS: 31,470
Min DPS: 27,807
Avg DPS: 29,212

As you can see this is well below an average Feral’s DPS. Tinderhoof ran some Feral parses and his average as around 41k (and I even think that is way low compared to other DPS classes). So at best HotW Guardian -> Cat is ~30% below a pure Feral. Based on what I know of the goals for Heart of the Wild, that is far short of the target.

I keep expecting Heart of the Wild to get fixed, but it keeps not happening.

Sorry guys, was super busy with the Guardian Round Table and didn’t have a chance to keep up with the Beta Class Analysis thread. I’ll link the podcast at the bottom of this post, but here’s what we’ve learned over the past week or so.

“I don’t know if it’s something new to level 86+ players but to me it seems like there is a new, unforeseen formula when calculating the Agility to Dodge variable. On my level 85 Worgen Druid on the beta, I can easily get the 243.58281085 Agility to Dodge conversion working. It does appear that Base Agility is now being affected by the Agility to Dodge (whereas on live it is not).”

All Agility (base and bonus) grants dodge at a rate of 951.158596 to 1% at level 90. However, dodge from bonus Agility and dodge rating is affected by diminishing returns. “Bonus” Agility refers to all Agility beyond what you have with no gear/buffs (for example, a level 90 troll Druid has 97 base Agility and all other Agility is Bonus).

Awesome. Basically it was the tooltip applying the diminishing returns on Agility and not on Dodge that was throwing everything off. With that, we should have better evaluations from the Guardian community.

Yeah, we will try to add dodge from Agi to the Agi and dodge tooltips. Currently they’re just included in the title of each of those tooltips, which are final values after summing all of the sources and applying DR. The “(before diminishing returns)” line in the dodge tooltip, refers just to the “dodge of X adds Y% dodge” line, not the total. It’s somewhat confusing right now. All of this goes for Strength / parry as well.

This is one of those things that won’t mean much unless you’re a Mathy type.

We were having trouble trying to figure out the diminishing returns on Dodge from both Dodge Rating and Agility because the tooltip for the Dodge stat in the character screen is behaving funny. It turns out the tooltip was automatically calculating the diminishing returns on Agility but not Dodge Rating. Thankfully GC took pity on us and told us how it really works. It’s helped us to make our models more accurate. It looks like GC will try and get to the tooltip for this as well as Parry fixed to make more sense.

I did some extensive mathematics on warrior tankiing, more precisely, Shield Block vs. Shield Barrier.

Short Version: There is currently nearly no reason for any warrior to use Shield Block. The amount of physical damage done by a boss melee swing to be high enough that the 30%/60% block outshines the absorb puts it out of reach of the vast majority of the playerbase. Obviously, this is even more true with non-blockable or magic damage, but even with the kind of attacks that can be blocked, Shield Barrier is just plain better for everything other than non-nerfed 25H bosses who use debuffs on the tank.

Thanks for the great analysis.

The scaling we chose for Shield Barrier was to keep it competitive with Shield Block with very high end gear (beyond even the gear available in the first tier). The problem is this is one of those situations where linear AP scaling can’t really fit the curve nicely. Shield Barrier (and by extension Frenzied Regen) will either be too good at low gear levels or too weak at high gear levels. We could use a more complex curve, such as squaring attack power as long as that doesn’t prove confusing. For sake of argument, imagine that Shield Barrier (and Frenzied Regen) is roughly half as strong as it is on beta when you’re in greens, but scales up in strength about 50% faster than it currently does.

While the original post was specifically about Protection Warriors and Shield Barrier vs Shield Block it also directly relates to Savage Defense vs Frenzied Regeneration. It’s also entirely 100% accurate. I’ve used similar terminology to describe the problems of the live version of Savage Defense vs a Block Tank. It’s basically the reason we had to keep receiving buffs all along this expansion. Basically the linear line was too far above the curve that represents damage reduction from Savage Defense and/or Shield Block. Turning it into a curve itself is definitely the best solution. However as you’ll see below the curve ends up being too shallow.

Now that FR has been reworked, I have some new numbers. Nothing has changed.

Your current build doesn’t have the changes to Frenzied Regeneration and Shield Barrier yet. If you’d like to get a head start on theorycrafting their value, the current formula for their heal/absorb is: 18.6 * (AP/1000)^2 – 0.047 * AP.

Another Guardian had posted some math that drew the same conclusions we already knew, and GC responded with the new FR formula. Fasc and Hinalover both posted math that came to the same conclusion: The formula is too shallow. Essentially there would be no reason to ever use FR, except in cases of magic damage recovery.

For reference you can find Fasc’s post here, and Hinalover’s post here.

Other than that, I’ll be doing some new parses (probably tomorrow) comparing various forms of Guardian damage including Heart of the Wild to see if there have been any relative improvements.

Also! We recorded the Guardian Roundtable last Friday with Fasc, Hinalover, Wenselaas and Buraan. You can find it here.

My CTC is…..

WTF EVER I’M A BEAR LOL

<3 Reesi.

There have been a couple Guardian oriented replies over the last two days. I wanted to take an opportunity to talk about them.

“If this is the prefered design why do Guardians have to spend rage for Maul? It has been stated a few times that Maul would never be used except to solo things (an only if the damage comes up).”

Maul is still a DPS button. It may not be one you use often, but there will be times when you don’t care about surviving (when the boss needs to die NOW, or even when you’re solo or running a scenario perhaps). We may eventually end up adding a proc or something (hopefully not exactly like Ultimatum), but we’re not sure it’s needed. At the moment Guardian damage is too high, so encouraging more Mauls isn’t super high priority. :).

?

Your logic failure betraaayyyyssssss you!

I understand that they may have more pressing matters to attend to than making something that is 100% functional simply more fun to play. That is a perfectly valid answer. However repeating the claim that Maul is fine as a DPS button that you never push (something we have actually proven is the case) doesn’t make any sense. What makes this even more confusing is that GC has said exactly the opposite thing about Heroic Strike earlier in the thread:

The idea is that we can’t ask Prot warriors to spend Rage on offensive attacks if we are also asking them to spend resources on defensive abilities — the defenses will just always win. The Ultimatum proc is there to let you use Heroic Strike and Cleave for free sometimes. When you aren’t tanking (like you are solo or off-tanking) or when survival isn’t a question (you’re running an easier dungeon perhaps) then you can Heroic Strike instead of Shield Block or Barrier. As you point out, you can use Heroic Strike and Devastate at the same moment since they aren’t on the same cooldown.

The two abilities fill exactly the same role for both class. They’re a Rage->Damage dump. His description of how Heroic Strike/Cleave would be used is actually bang on. That’s identical to how Guardians are going to use Maul. Only once we reach some insane number of RPS (rage-per-second) which isn’t actually practically obtainable do we start having excess Rage for Maul. So it only makes sense to give us some sort of mechanic to let us press it when we would rather spend our Rage on survival.

Wenselaas got an answer about Guardian DPS as well:

“Issue #1. Guardian dps seems to be EXTREMELY skewed towards mangle.

Issue #2. Guardian dps is pretty overpowered.”

You are probably right on both counts. We’ll look into it.

This means that most of the evaluations I’ve done thus far on Heart of the Wild and Guardian Cat DPS are likely wildly inaccurate. I’m not surprised. I’ve long maintained that Guardian DPS was too high so this is not news to me. Once the Mangle nerf comes in I’ll re-evaluate all of them and see where we stand.

Tinderhoof has long believed that Feral DPS is just terrible right now, so that probably isn’t helping things either.

One last bit of news. The Team Waffle Podcast: Guardian Round Table #2 is going to be broadcasting live this Friday the 13th of July at 7pm PST. If you have any questions feel free to go to the TWP website and post them for us to answer!

Nomnomnom.

TLDR:

  1. There are tens of thousands of Linux/Wine Diablo 3 players.
  2. Only 4 of them were banned.
  3. Whatever they were banned for is completely unrelated to Linux or Wine
  4. They were either cheaters or ran something else that turned up false positive by Warden.
  5. If they were innocent, then they are pretty much screwed without possible help.

I’ve been wondering what the fuss is all about since I’ve been playing Diablo 3 everyday on Linux using Wine (except for a 3 day vacation break in Phuket), and I have not yet been banned.

There’s been forum posts going around alleging that Blizzard has been banning Linux users using Wine to play Diablo 3. Since it’s posted in the Diablo 3 General forums and a Blue has responded to it, the thread has naturally became filled with trolls and retards. Amongst all the rubbish there are some accurate information which I will summarize here together with my experience and analysis as a Linux user that has been using Wine to play World of Warcraft and Diablo 3.

The forum post had links to various new sites that alleges that Blizzard has been banning Linux users, and pointed to the Diablo 3 Winehq appdb page as their source. Winehq appdb is a database of Windows application which have been tested against Wine. Users user appdb collaboratively to report if their Windows application work with wine, issues they faced, solutions to work around the issues. Developers also use appdb (and the associated issue tracker) to debug issues with Wine and to fix the issues for future versions of Wine.

The appdb page for Diablo has a thread where the Blizzard ban has been discussed, and 3 users have reported that they were banned. They are william, Marcus Meng and Mitch. I looked around for other cases of Linux users being banned and the only other one I’ve found was a PlayOnLinux user on this thread where fabioshot (and Mitch, presumably the same one from winehq appdb) claims to also have been banned.

PlayOnLinux is a software that includes Wine and automatically configures Wine to work with specific games that need tweaking to work well with Wine. For Diablo 3, this will presumably mean PlayOnLinux will include a version of Wine with the AcceptEx and the Direct3D modechange bugfixes as well as the setarch workaround the Warden issue with >4GB RAM and the taskset workaround for microstutters.

Now we need to look into context of the number of Linux users playing Diablo 3 on Wine. PlayOnLinux claims that there are at least 30000 users using PlayOnLinux to play Diablo 3. PlayOnLinux users are of course only a subset of Wine users use PlayOnLinux, so we can safely assume that the number of Linux users playing Diablo 3 on wine should number in the TENS OF THOUSANDS. This is only a small percentage of the millions of Diablo 3 players, but still a significant number of players.

Only 4 confirmed cases of Linux/Wine Diablo 3 players being banned out of tens of thousands of Linux/Wine Diablo 3 players should pretty much make it obvious that not only are Linux/Wine users are not being targetted for bans, there is no false positive issues where Linux/Wine gets falsely identified as a cheat. These 4 cases are definitely isolated cases unrelated to Linux/Wine, which can only mean 2 things:

1. These players were cheaters, and were either lying about using Linux/Wine or were running a cheat program on Linux/Wine
2. These players were running some other application/services/processes on Linux/Wine which has been falsely identified by Warden as a cheat

I’m not going to judge which of the above is true, it’s really Blizzard’s word against the players, and it could go either way. If the first one is true, then the players got what they deserve, and I wish all the worst for them for stirring up all these crap in the first place. Unfortunately, if the second is true, there is probably nothing that can be done to save these players.

Several years ago, World of Warcraft players using the Cedega variant of Wine to play the game in Linux all found themselves banned.  Blizzard initially claimed that they were cheats, but after being contacted by Transgaming (the company that made Cedega) and doing further investigation, they found that they were actually false positive, and reversed the ban and crediting the banned accounts with 20 days of game time.

In this case, there were thousands of banned users that were backed by a company. What chance would 4 isolated individuals have? It doesn’t help that he can’t really find out more about why these 4 were banned. Bashiok writes:

“It’s company policy not to discuss account actions with anyone but the account holder, or their legal guardian if applicable. It’s an issue between us and them. Trust me, it’d be much easier on me to just post exactly what they did, but we feel it’s important to honor the privacy of our customers, and that’s a policy I personally agree with.”

Now, I agree with this privacy policy, but the truth is the affected customers themselves do not have access to the reasons why they were banned beyond being accused of using a cheat. What would definitely help against false positive is Blizzard providing the customers with the name of the process that was detected and what cheat program Warden thinks that process was. This would allow innocent players to find correlation with other falsely banned players to identify the particular software that has been falsely identified as a cheat, possibly allowing the owner of the software to contact Blizzard (like Transgaming did) to help right the wrongs.

Of course, from Blizzard’s point of view, doing this will likely give the cheat writers an advantage against Warden that can and will be exploited. So really, if these players were innocent, I’m pretty confident that they are totally screwed.

I’ve decided to do this on a daily basis as new information comes out. It’s just easier on everyone (but mostly me) that way.

What happened today?

“How committed is the development team to having all tanks perform within an “acceptable variance” on all encounters? In Cataclysm we saw several “niches” become severe problems for tank balance on certain heroic encounters. For example:

  • Deathknights on Yor’Sahj
  • “Bearcatting” on Madness, Spine, Ultra, and Blackhorn
  • Deathknight and Druid compared to Paladin and Warrior for Impales on Madness
  • Warriors for Blood tanking on Spine
  • Paladins for most of heroic T12 due to Divine Guardian
  • Warrior/Druid/Deathknight vs Paladin on Al’Akir”

We don’t think any of those cross the line. If it were the same class showing up for all of those bullet points, that would be a problem. Our tanks all have strengths and weaknesses and unusual encounter mechanics may synergize or clash with them, but that’s more interesting than extreme homogeneity, which would be the alternative. We like the puzzle aspect of “solving” boss encounters according to the comp and strengths of individual groups. We think it has helped contribute to the fun of killing bosses having such extraordinary legs (meaning that we’re on tier 14 now and have made hundreds of dungeon bosses).

If memory serves the first Spine kill was Blood DK / Prot paladin and the second was a pair of druids. Now granted, world firsts sometimes have to resort to unusual strategies since they undergear the fights because they haven’t had weeks to farm up better gear. Your mileage may vary.

As a counter-example, paladins on Heroic Major Domo initially were able to solo-soak the scorpion cleave, letting guilds with paladin tanks keep him in scorpion form longer than those without. We thought that crossed the line and we changed the boss mechanics to disincentivize that particular strategy. I fully admit that these calls are subjective.

Um. What?

At Blizzcon 2011 another Feral Druid was in line ahead of me and asked specifically about Bearcatting. The response he got was at 28:50 of this video:

httpv://www.youtube.com/watch?v=XhRch8Smsaw

The answer is distinctly:

You are definitely going to be losing that DPS. The other tanks can’t do that.

That is directly contrary to GC’s answer above. Obviously people (especially designers) are allowed to change their minds. I do it all the time. I just wish there was some insight into the process behind the change.

Both Slash and Muspel had follow up posts:

“A lot of these problems are by and large, problems which a little bit of modification in the encounter’s design could be solved, if they were picked up early on. Death Knights on Yor’sahj makes a difference because the fight’s mechanics cater to them in every way. Simple changes could have allowed them to retain DK flavour and advantage, without making the difference anywhere near as large as it was.

If picked up by the encounter team, these simple adjustments could mitigate the potential for things to go out of control. This doesn’t apply just to tanks, but if encounter design had picked up, say, that Spine’s initial burst requirement was simply impossible for many classes and specs, that Anub was being broken by block being block, or that DK’s fed into this superclass for Yor’sahj, these situations would limit themselves to “nice to have” instead of “this feels broken”.

A lot of the problem for tank balance is that trends have tended to carry across entire tiers. Maybe this wasn’t intentional, but it’s like a person decided that a way to compensate DK’s for being weaker last tier was to cater many mechanics to them this tier. MoP could lend itself to more of the same, if it continues the same feeling where encounter designers do whatever, and class balance never seems to enter the picture even when the encounter’s very design breaks it, especially since Active Mitigation can work on these things now (hit ShoR, laugh at Impale)”

“It’s an issue for the same reason that fights like Spine of Deathwing are a problem for DPS balance.

It’s fine if there’s a “best” tank and a “worst” tank for an encounter. It’s less fine when the gap becomes too wide, as was the case on Yor’sajh.

By comparison, fights like Rhyolith, Shannox, Omnotron, Sinestra, and post-fix Domo still had a best and a worst choice, but the differences were minor enough that there wasn’t a strong incentive to switch.

In Dragon Soul, that hasn’t really been the case. DKs were insanely good on Yor’sajh, while warriors were at an incredibly disadvantage. Warriors were the best blood tanks, and DKs/Druids just plain couldn’t do it pre-nerf. Warriors needed tons of externals for fights like H Madness and H Hagara. DKs (and sometimes bears) would get wrecked by Blackhorn.”

We’ll have to see if there is more information forthcoming.

Nature’s Vigil is currently affecting Frenzied Regeneration. That doesn’t seem intentional.

I believe the intention is that the healing amount is increased, but that it does not deal damage based on that healing.

Muspel first pointed me in this direction, so I went to test Nature’s Vigil with Frenzied Regeneration. Turns out it does actually increase the amount healed by Frenzied Regeneration. This turns into an interesting choice analagous to what we had with T11 gearing – pick the 4pc bonus for burst or offset pieces for average survivability. Obviously we’ll have to take a look at how big the difference between Nature’s Vigil and Heart of the Wild is large enough to justify using NV every once in a while.

That’s all we’ve got for today.

Oh, and by the way Episode 34 of the Team Waffle Podcast is now available for you to listen to.

Shameless plug ^.^