In response to a thread on the official forums stating Death Knights were blatantly overpowered from their creation, Greg "Ghostcrawler" Street listed five mistakes the World of Warcraft development team made when designing the class. He does remind us, however, that the team "didn't make DKs intentionally OP" and he doesn't view the introduction of Death Knights as a failure.

1. The team avoided adding some of the limitations of other classes since they learned a lot about class design over the years. "We gave the early DK an answer to almost everything."
2. DKs "became a little too versatile" to counter the attitude that players wouldn't have room for the class in their raids and Arena teams.
3. The team may have listened to the community "a little too much" by adding or buffing abilities when it wasn't needed.
4. All of the blue items given to starting DKs "helped lead to the perception of favoritism." There may have been a better way for the team to communicate its goals for the hero class.
5. The developers "just had bad luck" when the DK's cooldown-based tanking became highly effective in boss encounters and the plate/spell damage combo was highly lethal in early Arena seasons.

Check out Ghostcrawler's full post after the jump.

We didn't make DKs intentionally OP. While we want people to purchase Wrath of the Lich King, there are lot of logical flaws with making one class purposely overpowered (such as angering all the players who play the other nine classes).

We made a few mistakes though, which I don't mind being up front about.

1) We have learned a lot about making classes. With the DK, we avoided some of the limitations other classes have. In the absence of a lot of tradition, we gave the early DK an answer to almost everything. While we definitely don't want to keep some of the crud that has built up on the older classes, it turns out when you make a new class without any of that junk before you scrub away the older stuff, you get a class that is too powerful. Who would have thought? The DK talent trees in particular had a lot of polish to make sure there weren't too many just terrible talents (a few haven't withstood the last year as well as others though).

To use just two examples, DKs can tank or PvP with any tree (though not necessarily with the same talent allocation per tree), and DKs have runes, a resource that is limited over short periods of time instead of long periods of time, yet they also have runic power that is quite useful in those periods where they may not have runes available. In some ways, the DK benefited from having two resources. (To be fair, there are weaknesses inherent in the rune model as well which players who have never played a DK seriously tend to overlook.)

2) As we started to work on the DK, we talked to PvP and PvE players who were comfortable with the BC status quo. What I mean is that raid leaders told us they didn't have room for DKs and Arena teams told us they didn't see the need to replace anyone on their team (especially a melee class without MS). Partially to counter this attitude, we gave the DKs a lot of different tools to handle different situations. They became a little too versatile. (Later, part of the big buff / debuff overhaul was designed to fix the same situation.)

3) I think this was one of those cases where we listened to the community, especially the beta participants, a little too much. Players had totally legit concerns about say the DK being too kiteable or the DK tank not having an answer for physical damage or whatever. Since we didn't have a ton of history or personal experience to fall back on and couldn't analyze a lot of data from live servers, we probably gave players the benefit of the doubt too often and added or buffed abilities when we didn't need to. More hardcore PvP and raid testing on the beta or PTR might help as well, since a lot of those concerns were born of solo experiences or pure speculation.

4) I think the fact that we showered the DK in blue items and mounts (to help them level up to the Northrend content quickly) and all of the effort we put into the start zone helped lead to the perception of favoritism. It feels weird to suggest we should have made the starting experience less epic to just to tone it down. On the other hand, maybe there was a way to communicate our goals better. A lot of players took the "hero class" thing (which really only meant that we modeled them after the Warcraft III heroes) to mean they should be overpowered, which was never our intent.

5) I think we just had bad luck with at least two of the DK systems. Their plate + spell damage combo just happened to be quite lethal in the early Arena seasons where 2s were predominant and burst damage was too high. Their cooldown-based tanking just ended up being really effective against a lot of the boss encounters. We didn't set out for either of those situations to be too uber. If the DK had been a caster, I don't think casters would have dominated the earlier seasons. If health pools had grown more (as they did going into BC) and we had been harsher about mana regen, then perhaps tanking wouldn't have been so overshadowed by giant boss hits that could only be survived by chaining say IBF and Bone Shield.

Ultimately, I wouldn't view the introduction of the new class as a failure by any stretch. There are a lot of players who love their DK even when she isn't overpowered. Adding a new class is a challenge, which we knew going in, and the DK didn't disappoint in that regards. It has taken a lot of iteration to kind of catch the death knight up all of the years of iteration on the existing classes, and I'm sure they're not done yet. It's also a good reason not to introduce a new class every expansion. :)