So What Went Wrong?

Blizzard is never, ever going to tell us exactly why paladin development for WotLK unfolded as it did. Currently, Ghostcrawler is publicly blaming it on some bug with weapon switching. But I don't think that's the real story. There's too many other factors in play. So this is my speculation as to how Blizzard got into this position. Note that I have zero evidence that any of this is real, just what happened in Beta/Live, and my gut feeling as a somewhat experienced paladin.

For PvP, at the beginning of WotLK, I think Blizzard made the design decision that a Ret Paladin would be high burst, but very immobile. Essentially, it would be very hard for her to get inside melee range, and she would be pretty easily kited. However, if you let a Ret paladin catch up to you, she would be able to unload and burst you down.

The reason I think this is what happened is that none of the current Ret burst is surprising. It doesn't take a genius to see that Ret was going to try to run up to someone and hit all the buttons in a row. It's not like a Frost Mage's Shatter Combo, which revolves around getting off an instant Ice Lance while a Frostbolt is still in flight. That's an example of something which is tricky, and can be unexpected if you're not an expert with the class. There's no way that any person reasonably familiar with Retribution could fail to predict Hammer of Justice -> Judgement of Command -> Divine Storm -> Crusader Strike. It's not clever or something which takes mad skill, it's obvious.

Second, consider the initial Art of War talent on Beta. It gave CS a chance to double the damage of the next Judgement. The immediate combo with the old Judgement of Command, and the resulting *eight-fold* damage on a stun-crit comes to mind. As I commented in July, when I first got into the Beta:

Sometimes I really don't understand Blizzard. Back in March, they said they were worried about Paladin burst damage. Fast forward to the WotLK Beta, and we see the following two talents: [Righteous Vengeance, Art of War].

The existence of the original Art of War--and the crit damage increasing talents like Righteous Vengeance--says to me that either: no one at Blizzard plays a Retribution paladin (entirely possible); or paladins were deliberately being given higher burst in the initial WotLK design.

So in Beta, Blizzard toned down the worst excesses of burst, but left the design of "immobile + high burst" alone. Then they released it to Live, and immediately the vast majority of the populace made it clear that this was an unacceptable design decision. The outcry forced Blizzard to scrap the "immobile + high burst" concept, and just hotfix Retribution down to a nominal level.

So that's my theory for PvP. It is my explanation of why Blizzard is just going nuts with hotfixes. That's not a rational response to a few bugs or missing the damage target slightly. If that had been the case, Blizzard would have just shipped fixes in Patch 3.0.3. It's the reaction to realizing that the entire design concept was seriously flawed and needs to be scrapped.

For PvE, I think the original spreadsheet/model that Blizzard used to come up with initial damage numbers had a significant error. My guess is that the model didn't account for Seal procs from specials.

The reason I think this is the case is that Blizzard has consistently focused on Seal/Judgement damage when nerfing Ret's sustained damage. This is despite the fact that Seal/Judgement nerfs hurt the other specs, especially Holy, even harder. Blizzard went to a lot of trouble to make it easier for healers to solo, and I really don't understand why they targeted the paladin mechanic that is most important to the healing spec (and levelling characters). Last time around, Blizzard was willing to do things like play with cooldowns. If Ret's damage was too high, I would have expected Blizzard to target the Ret-specific abilities like Crusader Strike and Divine Storm. Decrease damage, or increase cooldowns. Change Ret damage increasing talents to do something else. But they didn't really touch CS/DS until the Live hotfixes.

On Beta, Blizzard kept adjusting Seal damage down, saying that Seal damage was too high. To me, that implies that the damage from Crusader Strike and Divine Storm was meeting their expectations, but Seal damage was consistently higher than they predicted. That implies that their model wasn't predicting Seal damage correctly, and the most obvious explanation is that it didn't account for the procs from specials. That's the major change from TBC (other than the change to AP + SP scaling, which is much easier to check and much more likely to have been the first data to be compared).

This is all speculation. I don't know if any of this is real. But the paladin class is not a very complicated class. There's no feedback loop like warriors and Rage. There's no complex mechanics like Combo-Point generation. In fact, if you look at the other classes, Blizzard was mostly in the right neighbourhood most of the time, barring crazy bugs. They've been much more surgical with the other classes than paladins. To me, that says that the spreadsheet/models were pretty accurate. The biggest change I can remember is the warriors losing Heroic Leap because it was too buggy, and the top end of Fury being shuffled around. And that's in a completely different league than getting the damage numbers wrong.

Paladins are very straightforward and unsurprising, and I really would have expected Blizzard to nail the paladin numbers from the very start, both in PvP and PvE. That they didn't, in my opinion, points to deeper underlying causes than mere weapon-switching bugs on the Beta server.

Edit: Additional proof for the PvP side of things. Ghostcrawler is now posting on the forums that Blizzard is considering giving Retribution some form of utility like a snare or interrupt. This is a 180-degree change from Blizzard's previous stance on Retribution, and further indication that their design for Ret is now completely different than in Beta.