I'd be more interested in a study of different specs based on more succesful players in different areas of the game. For example, it would be interesting to see a breakdown of the most popular talent builds among all classes in these groups:

The top 20 5v5 teams of each battlegroup
The top 30 3v3 teams of each battlegroup
The top 50 2v2 teams of each battlegroup
The top 100 most progressed raid guilds

To be fair, I'd bet on weekends you might see more arena spec and less prot warriors. The warriors in my guild change spec maybe 3-4 times a week besides the MT and one of the more casual OTs.

Unfortunately, I don't keep stats on the actual talents a character has, but I do have over 49,000 level 70 warriors in my database which is (I think) a decent sample size, though it is still skewed somewhat -- they are either self-reporting, post to the WoW Warrior Board, or have a guildmate who self-reported and or posted to their class forum.

This also means level 70 alts are included which will also skew the data. Here is my data:

Total Count: 49380 Warriors

31 or more points Arms: 16663 (33.74%)
31 or more points Fury: 7547 (15.28%)
31 or more points Protection: 25074 (50.78%)

I also checked out mages since it's basically the same script:

Total Count: 51683 Mages

31 or more points Arcane: 17677 (34.20%)
31 or more points Fire: 13764 (26.63%)
31 or more points Frost: 11532 (22.31%)

Interestingly it looks like there is a non-insignificant number of tri-builds for mages.

I don't have any priests in my database yet.

Interesting thread, I think I might start tracking all the talents instead of just the talent distributions.

Edit: the mage numbers are off. Please see my next post a couple posts down.

This doesn't especially surprise me. I would have guessed that the majority of non-raider, non-pvp warriors would be protection specced. I am fairly certain that the vast majority of players post-TBC do not do any content bigger than karazhan, so ~60% protection warriors seems pretty reasonable, at least for non-PVP realms.

I don't find this to be true in a blanket sense. It's hard to tank some heroics and some raid content sure, but that's hardly the whole game. I've tanked Olm with my 35/5/21 warrior alt on half our Maulgar kills. I'd gain hardly any mitigation by speccing deeper into prot and my poor threat generation isn't an issue there so it works out fine. And of course you can do any normal 5man with 0 points in prot quite easily.

I think as suggested above what we're seeing here is selection bias.

We call them bad mages. 16.86% deserves a

What the hell builds are people coming up with that don't get a 31 point talent? I can only imagine something asinine like this. PoM Pyro + Ice Lance against a frozen target 4TW!

Oh good, that means only 0.09% of mages are known bad without evaluating their gear. That makes me breathe a little easier.

I used to think that CoH was terrible (I still think it's a bit too situational), but not once did I try it. Recently when we were first working on Morogrim, I decided to pick it up and give it a try. I hadn't respecced since I hit 70, and my cost was at 15g like it's always been pretty much. So I picked it up to give it a try, and I've been finding more and more uses for it.

It certainly has it's limitations, but on a fight like Morogrim, it's pretty amazing. It's biggest limitation I'd say is the radius on it, but in a fight where you have raid-wide damage, and can have the raid all bunched up, it's a very useful tool. Another fight that I found it really useful for is Kazzak. I don't even want to think about how easy healing the meteor soakers on Thane would have been had I used it back then for that.

With my gear and talents, the spell costs 405 mana, and heals for over 800 average when you factor crits in (808 exactly is what my SWstats showed after 1000+ uses). If you manage to get 3 targets that need healing, it's 2400 healing for 405 mana which isn't that bad by itself. When it gets stupid is on fights like Morogrim or Kazzak where you can easily hit all 5 people in the group (more with pets too), and hit that 10:1 heal to mana ratio.

Sure, it's not an amazing talent, but even so, I can justify spending a point in it for those moments when it really shines. Same deal with lightwell really; there's fights where it's just too nice to have, but I still wish it had a few more charges. Despite me actually finding a use for it these talents, I still don't think that's reason enough for them to be our 31/41 point talents in their current form. I'm more under the impression that our 31 and 41 point talents should have a bit more versatility than this.

If blizzard wanted people to choose CoH, they'd just make DS trainable and leave Imp DS there. So you choose between Imp DS and CoH. A much easier choice.

Unfortunately, Blizz came out and said specifically that this was not being considered, which seems really stupid to me, but hey, who am I?

http://forums.worldofwarcraft.com/th...eNo=2&sid=1#24

41% arcane? Wow.

It gives me chuckle that there are 0.4% of the warriors who are considered "Arm" warrior but they didn't take mortal strike.

Even so, I just get my DS from our priest who is specced for it. It's nice to be able to buff yourself with it all the time, but it's not needed. In raids, we don't benefit from having more than one priest with DS/Imp DS which is why I never picked it up because we already have someone who did, and not having spirit for my 5 man groups isn't all that big of a deal.

I still envy those that can go to like 50+ points in a given tree because it's just that good. That, and they don't feel like they're hurting themselves by not picking up those tier 3 talents from another tree, but I can still fill out my talents and feel good about my choices, and better yet, enjoy playing the class.

The arcane mage turnout isn't surprising at all. As already mentioned above there are a vast number of mages who aren't end-game raiders - I would believe that a significant amount of those mages are deeply rooted into Arcane and call themselves "pvp spec".

fyi - arcane is a viable raid spec. 40/18/3 using AB*X Fireball*2 cycles can lay down about the same dps as a 10/48/3 fire mage on bosses (and crushes fire on trash hehe) & it's a more viable pvp spec.

That is my exact spec (and rotation). X is usually equal to 2, btw.

You really shouldn't be surprised considering the level of intelligence of some of the players in WoW. Sure, 40/21/0 and 33/28/0 are good raiding specs but deep fire is a bit better for raiding and from what I've heard, deep frost the same for arena and BG's. That means you'd expect arcane to be 3rd by a little bit. What makes arcane attractive to many players is the "OMFG, huge numbers" of AP/PoM/pyroblast.

If you give survival hunters a 41-point talent named "I am gay like Legolas" with a 1-min cd that makes their next attack deal 10k damage while making them unable to move, attack or cast spells for the next minute, you'd see at least 30% of hunters be survival too.

I think it's pretty telling about the state of the hunter trees that we have both the most-frequently specced tree (marks) and the least-frequently specced tree (survival) out of the entire data set. It doesn't bother me personally, since I've always preferred to play my character as an archer archetype, but it's just one more shining example of the lack of actual planning that went into designing the hunter class.

The survival tree in particular has no focused direction or any real purpose at all as far as I can see. You should be able to point to a well-designed talent tree and say "That's for raw DPS" or "That's for raiding utility". It seems to me that with some of the less-popular trees, the developers just frankensteined together their leftover concepts after creating the other trees.

both the feral and prot prevalence is easily explained - most realms are short on
tanks in a big way, so speccing feral or protection means you can spend virtually every
logged in second tanking something, if that is what you want to do.
And being a tank with something to tank is very, very, very fun! and heroics rock.

As a former mage, I think the mage trees are so evenly distributed because at the end of the day they play the same. So you shoot red bolts that take 3 seconds instead of blue bolts that take 2.5 seconds - big deal. Take away the signature one point talents (Ice Block, Water Elemenal, PoM, AP, Dragon's Breath) and its all the same.

That, while seemingly suggesting that the class is balanced internally, is not neccessarily a good thing. The lack of meaningful options was always what frustrated me about my mage.

I had a minmaxxed 11/0/40 spec on an alt that included Arcane Fortitude and Arctic Winds but not Improved Frostbolt. The catch is that he was my second level 60 mage and was completely specced for farming crocs in ZG. (He did still have Ice Barrier, but the point is that not all odd specs are bad specs, as long as they're built to match the player's needs).

I was actually really disappointed when I actually read the post. All he cares about are 31 and 41 point talents, yet garbage talents aren't limited to 31 & 41 point slots. I would really like to see what percentage of Assassination spec'd Rogues actually take Deadened Nerves (-5% physical damage taken, 5 points) or Combat spec'd Rogues who take Blade Twisting (20% chance for specials to daze, 2 points). I'd wager that its less than 0.5%.

Correspondingly, it would be interesting to see if there are any specific talents that over 95% of builds for a particular class have.

So many mages I know LOVE their super-buttons. Is it really that much of a suprise? Same reason a lot of affliction locks don't take UA and go for ruin. Big numbers are fun, apparently.