Blog Posts from "WoW Arena Power Ratings" - Gameriot.com

Power Ratings: The Final Countdown

no.spam@gameriot.com (briyan) — Mon, 15 Sep 2008 16:37:59 -0400

Four months after my last post on the Arena Power Ratings, I am back with a minor update.Â For various reasons, these numbers have become less and less meaningful as the BC expansion wears out, so I waited until Season 4's dying breath to make this last post.Â Given that so many players are inactive now, the beta is out, and the 3.0 patch is coming to completely destroy the classes as we know them, the end is here for BC arena. Â Consider this a "last look" at how things were before the new Death Knight class enters the fray along with all sorts of crazy new talents.

Today's Arena Power Ratings

Rank	Class	Power
1	Warrior	109.47
2	Druid	107.59
3	Rogue	107.22
4	Priest	105.12
5	Warlock	97.57
6	Mage	97.43
7	Shaman	93.36
8	Paladin	84.18
9	Hunter	76.53

People's Choice Arena Ratings

Rank	Class	Percent
1	Druid	181.72%
2	Rogue	162.17%
3	Warrior	160.76%
4	Warlock	99.24%
5	Priest	95.52%
6	Mage	69.91%
7	Shaman	59.34%
8	Paladin	45.48%
9	Hunter	35.55%

Hunters and Paladins, with Shaman not too far behind, should be looking to WotLK for hope.Â The WoW developers have not exactly shown a strong response to arena balance issues for these classes in BC, but the new expansion gives them a chance for a do-over, as all class balance and mechanics are tweaked at once.

After serving as the bottom feeders for the entire expansion, a return to viability for any of these classes should show up quite dramatically on our WotLK ratings charts.

Graphical Timelines (relative to the "balance point")

Explanation of Arena Power Ratings (Math)

no.spam@gameriot.com (briyan) — Sat, 24 May 2008 08:05:36 -0400

I've had a lot of questions and criticism about my Arena Power Ratings, and while I've talked about shedding more light on the way I generate them, I haven't really delivered on that promise. Until now! I hope this information gives some of you guys a better idea of what I'm doing with these numbers when I make the pretty graphs and give you a speech about which class is doing poorly.

The Main Data Source

Just for the sake of disclosure, I don't mine any Armory Data myself. The Official World of Warcraft Armory is quite a beast, and several people already dedicate a good bit of time to digging through it with various software tools. Rather than try to cook up some of my own code and constantly tinker with the processing, failed responses, irregular reporting of player names, arena teams, ratings, and so on, I decided to rely on someone else's work. After checking out a few sites, I found that RealmHistory seems to give me the most useful breakdown when it comes to Arena information, so that's what I use.

I hope that clears things up for you guys wondering how I get these numbers. I am sure these details will generate even more complaints, but I think the knowledge is important!

Power Ratings: BC's Twilight

no.spam@gameriot.com (briyan) — Fri, 23 May 2008 10:09:49 -0400

I haven't updated my Power Ratings in several weeks now, due to a few factors. First, I think my initial plan for weekly updates was a little much, when the measurable changes are so small from week to week. Second, the Armory has been performing terribly over the past month or so and I want to make sure I feel confident with the numbers from the data-mining sites before I post an update. With that in mind, I will target a monthly update from this point forward, though I will continue to record data at least once a week.

Today's Arena Power Ratings

Rank	Class	Power
1	Warrior	110.71
2	Rogue	108.04
3	Priest	107.76
4	Druid	102.23
5	Mage	96.81
6	Paladin	94.26
7	Warlock	93.81
8	Shaman	92.68
9	Hunter	81.53

People's Choice Arena Ratings

Rank	Class	Percent
1	Rogue	154.11%
2	Warrior	147.13%
3	Druid	125.87%
4	Priest	120.76%
5	Warlock	85.65%
6	Mage	84.28%
7	Paladin	72.72%
8	Shaman	70.24%
9	Hunter	47.31%

BC's Twilight

We basically knew we'd reached this point already, but the start of the WotLK alpha phase officially ushers in the twilight days of The Burning Crusade. Though Season 4 has not started yet, we can assume there will be no more attempts to tweak balance at the 70 level cap, and this is "it" for the rest of the way. Developers will be working fulltime on the 71+ talents and skills, attempting to get balance going there with the Deathknight, the new Battleground, the outdoor PvP area, and the flying combat mechanics. For "live" WoW, that means we can expect today's trends to continue throughout Season 4.

Trending Up - Rogue, Mage, Shaman

All of these classes have moved steadily upward over the past several weeks, for various reasons. Rogue has continued to benefit from the Subtlety overhaul that cemented Shadowstep as "the one true build", and their unique ability to derive big gains from PvE gear in a DPS role has allowed them to reach a dominant position. A nerf to Rogue would not be surprising at this point; though Warriors have been this well-represented from the beginning of Arena (for 1.5 years), Blizzard seems to react quickly if any other classes reach similar levels.

Mage and Shaman seemed to get most of their jump out of the initial buff patch of 2.4, but they have continued to creep up the tiniest bit afterward. These are two classes who certainly stand to benefit as the Warlock population drops and the popularity of melees increases, so Rogue is really helping them out.

Holding Steady - Warrior, Priest, Druid, Paladin

These classes have received a handful of tweaks but nothing that's really changed their position over the last several weeks. Paladin is the weakest of the bunch, and will remain well below the others barring significant tweaks. Druid received some nerfs, but this seemed to alter their "per bracket" positioning moreso than the overall representation. A slight dip in 2v2 combined with a slight increase in 3v3 has kept them in place slightly above even.

Priest continues to look very strong in all three brackets while basically avoiding attention due to Druid healer strength in 2v2. Warrior, our perennial chart-topper, has finally seen Rogue catch up with him in the small bracket so he now shares his lead (Warrior is 1st in my Power Ratings, while Rogue is 1st in the People's Choice ratings).

Trending Down - Warlock, Hunter

The only two to show sustained dropoff in representation over the last couple months, these classes were not really the target of Blizzard nerfs as much as they've been trampled by Rogues on the way to the top. Hunters were actually given a slight buff a while back, but unfortunately the minor improvements were not properly targeted and failed to get them into a winning position.

A few major pitfalls for these two:

* Pets increasingly at risk as damage scales up over time
* Inability to escape melee zergs as melee becomes more popular
* Rise of Shadowstep Rogue shuts down kite opportunities

Overall, these guys are in rough shape, and the growing frustrations (with no end in sight) have likely led to some class abandonment during Season 3, further exaggerating the loss in numbers.

Graphical Timelines (relative to the "balance point")

How these ratings are calculated

Power Ratings: The Age of Melee

no.spam@gameriot.com (briyan) — Mon, 21 Apr 2008 01:45:10 -0400

We've now had almost three full weeks for the 2.4 changes to work their way into the system, and the results are clear: melee is still king in WoW and they will remain comfortably on top through the end of the BC expansion.

Today's Arena Power Ratings

Rank	Class	Power
1	Warrior	111.39
2	Rogue	108.56
3	Priest	106.16
4	Druid	101.57
5	Warlock	96.22
6	Mage	95.75
7	Paladin	93.81
8	Shaman	91.20
9	Hunter	83.69

People's Choice Arena Ratings

Rank	Class	Percent
1	Warrior	151.91%
2	Rogue	148.10%
3	Druid	121.97%
4	Priest	116.18%
5	Warlock	91.76%
6	Mage	82.75%
7	Paladin	74.12%
8	Shaman	68.09%
9	Hunter	52.59%

The Age of Melee

We knew going into the 2.4 patch that melee were strong, and the Season 3 arena itemization made sure that they would dominate unless Blizzard stepped in with skills/talents changes. Luckily for melee players, Blizzard did not jeopardize the front-runner position for melee. This particular patch had the largest changes in store for Druid, Shaman, and Mage, and the past three weeks seem to have validated early speculation that these changes would have fairly minimal impact overall.

We have seen a slight drop for Druid. We have seen slight gains for Mage and Shaman. This would seem to fall in line with what Blizzard hoped to achieve. More telling, however, are the relationships between classes that largely serve the same function in the arena.

* Druid has drifted slowly down to a "more balanced" level, but we still see a huge gap between our "strong" healers (Priest, Druid) and our "weak" healers (Paladin, Shaman). It would seem that the nerfs to Druid have not really directly allowed other healers to become more viable, though Rogues seem to have gotten some benefit.

* Shaman has shown a definite upward trend since the patch, but remains last among healers and second-to-last among damage classes. We think there is still room to grow, especially for Shaman who may be gearing a new spec, but the class remains in a weak position overall.

* Mage shows continued (slow) growth over the entire month since the patch was released. While Mage did receive some minor buffs, we think the gains are basically coming in the wake of Warlock's loss in ground against the ever-rising melees. We actually show these two classes on the verge of switching places on the power ratings scale, where Warlock has held the higher position since the launch of Arena.

Who are the big losers? Warlocks and Hunters, it would seem. Both classes are at an all-time low since we started tracking, and they have gone straight downhill since the 2.4 patch was released. Given that the general trends of Season 3 should continue and strengthen in Season 4 (with no balance patches scheduled until the expansion), it really looks bad for both of these classes for the rest of BC.

And, for the first time, we have a class that looks like it may threaten Warrior for the top spot on our charts -- the Rogue! Fueled to new heights in the 2v2 bracket by Druid nerfs, the Rogue is gaining ground to his fellow pure melee class at a rapid pace, and is now clearly established as second best overall.

Graphical Timelines (relative to the "balance point")

Power Ratings: 2.4 Arrives

no.spam@gameriot.com (briyan) — Mon, 31 Mar 2008 01:17:03 -0400

The long-awaited 2.4 patch finally arrived on live servers this past Tuesday, bringing some subtle PvP balance changes that should have a visible impact on our power ratings. Arena-specific tweaks were mostly targeted at Druid, Mage, and Shaman, but we'll still need more time for the community to show if the scales tip toward any other classes.

Today's Arena Power Ratings

Rank	Class	Power
1	Warrior	112.02
2	Priest	106.24
3	Rogue	105.39
4	Druid	103.33
5	Warlock	100.29
6	Mage	93.24
7	Paladin	92.58
8	Shaman	89.17
9	Hunter	87.98

People's Choice Arena Ratings

Rank	Class	Percent
1	Warrior	155.35%
2	Rogue	133.31%
3	Druid	128.13%
4	Priest	115.97%
5	Warlock	106.12%
6	Mage	74.26%
7	Paladin	70.68%
8	Hunter	64.23%
9	Shaman	63.76%

Patch 2.4 Arrives

The changes in patch 2.4 will take time to show up, given that our data is mined from Armory and the balance of power and bracket popularity is slow to shift. The TTR is still around to distract players as well as new PvE objectives in the Sunwell, so we likely won't see the full effects for at least a month.

Unfortunately, the 2.4 patch also means we are entering the twilight of Arena Season 3, and this means the ladders will start going haywire soon as players try to position themselves for Gladiator titles. Friends will be powerleveled, wins will be traded, and the general integrity of the data will drop until Season 4 is released. We can hope this would impact all classes equally, but in reality it will be hard to measure, so I'm giving fair warning.

From the arena-specific changes, we feel Shaman have the most to gain this patch while Druid and Rogue have the most to lose. The changes to Druid in particular are rather far-reaching, in that their shorter Cyclone range puts them in a good deal more danger than before, and increases the relative strength of all their enemies. We expect Shaman to gain ground as healers in 2v2 (a weak bracket for them today) with the ability to Earth Shock the Cyclone and gain access to instant Ghost Wolf to chase Druids around.

It's hard to say just how much ground Rogue will lose in larger brackets with the fix to Cheat Death, but the psychological effect is already apparent in the competitive community. Focusing the Rogue is no longer seen as a complete taboo, and Mutilate buffs may shift a few bored rogues out of the deep Subtlety spec altogether. We should probably expect Warrior to profit the most from these changes to Rogue.

Mage changes take up a lot of space in the patch notes, but I am not confident that we will see much positive effect in the bracket representation based on the nature of the tweaks. I would love to be proven wrong, but I would probably chalk up any boosts to the rise of melee and fall of Warlock more so than the strength of Mage alone. Regardless, the jury is still out here.

Graphical Timelines (relative to the "balance point")

Power Ratings: A Second Metric

no.spam@gameriot.com (briyan) — Mon, 24 Mar 2008 03:03:01 -0400

We are showing fairly predictable results again this week on the Power Ratings scale, as no significant changes have been implemented or announced. Experimentation by top players on the TTR continues to slow the visibility of live server trends.

Today's Arena Power Ratings

Rank	Class	Power
1	Warrior	112.02
2	Priest	106.16
3	Rogue	105.11
4	Druid	103.57
5	Warlock	100.57
6	Mage	92.67
7	Paladin	92.58
8	Shaman	87.79
9	Hunter	87.32

A Second Metric: People's Choice

In response to requests for a simplified rating system, I have added a new metric this week, named the "People's Choice" arena ratings. This metric combines the brackets together weighted by the popularity of each bracket by number of high rated participants. This representation figure is then compared to the theoretical "balanced" representation (even for all classes), and we express the result as a percent of "expected performance," similar to Kalgan's now-famous chart on the WoW forums.

A value of 130%, for example, means there are 30% more of that class at high rating than we would expect if all classes were represented equally. I will continue to maintain this data (including the graph) along with my regular Power Ratings as time goes on, so we can reference this in the future.

What you should be noticing here is a slightly reduced emphasis on 5v5, and the result is a very poor showing for classes who are only well-represented in that bracket (Paladin and Shaman) and a depressed rating for Priests, who are ubiquitous in the 5v5 world. Given the various fudge factors involved with these calculations, we have Warrior as the clear #1, followed by Rogue and Druid basically tied for #2.

People's Choice Arena Ratings

Rank	Class	Percent
1	Warrior	155.31%
2	Rogue	131.81%
3	Druid	130.04%
4	Priest	116.16%
5	Warlock	107.85%
6	Mage	72.51%
7	Paladin	70.52%
8	Hunter	63.64%
9	Shaman	60.17%

Graphical Timelines (relative to the "balance" point)

Power Ratings: Holding Steady

no.spam@gameriot.com (briyan) — Mon, 17 Mar 2008 01:13:04 -0400

As expected, not much has changed this week in the Arena Power Ratings. The 2.4 data preload reminds us that the patch is creeping closer, but the live date is far enough down the road that we haven't seen any impact from anticipation.

Today's Arena Power Ratings

Rank	Class	Power
1	Warrior	111.62
2	Priest	107.01
3	Rogue	105.11
4	Druid	103.61
5	Warlock	101.08
6	Mage	92.67
7	Paladin	91.03
8	Hunter	88.44
9	Shaman	86.70

A value of 100 is considered balanced.

Holding Steady

While we expect to see some small shifts in the numbers due to the ever-changing FOTM compositions in the 5v5 bracket, the opening of Blizzard's Tournament Test Realm server has surely slowed the changes somewhat. With access to top-geared characters of every class, many top players are spending their serious arena time on TTR and leaving their "live" arena teams inactive, sold, or compromised with bench players and friends.

We don't think it's a huge effect, but the PTR will further delay the visible changes in class representation until the new patch arrives.

Graphical Timeline

Showing values relative to the balance point of 100.

2.4 and the Small Bracket Druid

no.spam@gameriot.com (briyan) — Wed, 12 Mar 2008 13:33:52 -0400

Today on the PTR we saw a range reduction from 30 to 20 yards for Druid's Cyclone spell. This is one of a few Druid-specific tweaks planned for the upcoming 2.4 patch, and it seems like a solid attempt to address the over-representation of Druid in the 2v2 bracket. Most players agree that Druid is too strong in today's 2v2, and I think this Cyclone nerf is a step in the right direction.

What are the important changes?

In the PTR so far, we have a few key changes to Resto Druid:

* 15% speed bonus removed from armor sets
* Lifebloom final heal slightly nerfed
* Natural Perfection slightly nerfed
* Cyclone range nerfed to 20 yards

Not negatives as much as "problems that aren't addressed by these changes". While Blizzard originally held the position that they would not perform class balancing based on the 2v2 bracket, we can see they have changed stance. With this in mind, we have to look at other classes in 2v2 when we consider changes to Druid.

Mage still struggles in 2v2, and Druid is a major contributing factor (along with Warlock). Mage/Rogue teams have built careers on killing Druid/Warriors, but this is a high-skill team that faces many other tough matchups to stay on top. Druids can easily outheal Mage damage, and are not really vulnerable to Root, Counterspell, or Polymorph. None of this changes with the 2.4 Druid nerfs. In fact, the expected reduction of "average" Druid/Warrior teams at high rating will cut back the main point source for Mage/Rogue.

Warrior, today's other dominant 2v2 class, will probably not be affected, and may even grow more prevalent. Druid is the best Warrior-controlling class, and fewer Druids means more Warriors with other healers. The melee scaling gap will continue to grow in Season 4. At what point will Blizzard try to address small-bracket Warrior dominance? I'm not sure either.

Perhaps the biggest negative is that Druid/Hunter will continue to gain even more ground in 2v2, as the Druid nerfs have the smallest effect on this team setup. Hunter is able to keep his Druid out of danger better than any other class, so the need for clutch survival skills from the Druid is reduced. This is really bad for everyone involved, because Druid/Hunter is without question the most annoying team you can possibly fight, and really sucks the joy out of 2v2 for many players. This will further increase the perception of Hunters being OP and prevent them from receiving the attention they deserve from the developers.

The final area for concern is alternate Druid specs. This may not be a major concern for Blizzard, but Balance and Feral are already struggling. While all these changes are targeted at Resto Druid, the weak specs are hurt as well. What can you do?

Power Ratings: Death of the Middle Class?

no.spam@gameriot.com (briyan) — Mon, 10 Mar 2008 00:57:44 -0400

It seems RealmHistory had some code problems over this past week, so we didn't get as many data points as usual. Regardless, the numbers were back up on Friday so we're all happy.

Today's Arena Power Rankings

Rank	Class	Power
1	Warrior	111.17
2	Rogue	104.93
3	Priest	104.86
4	Druid	102.87
5	Warlock	100.38
6	Paladin	95.18
7	Mage	95.08
8	Shaman	88.04
9	Hunter	87.50

A value of 100 is considered balanced.

Death of the Middle Class?

We started out a few weeks ago with a solid group of three classes near the overall balance mark (Rogue, Druid, Warlock). Since then, we have seen Rogue start to pick up steam, and now Druid is following the same upward path. If arena team trends continue as expected, we should soon see Warlock alone near the middle, and Priest/Rogue/Druid will be vying for second behind the One True King (Warrior).

Warrior, still the only class with an "above even" representation in every bracket at both 2000+ level and 2200+ level, shows no sign of weakening at this point, so we think his position is safe at the top of the heap.

Priest has started to drop a bit, likely due to the rise of Druid, but his overall versatility ensures that he won't drop too far. As the most important core piece of the 5v5 bracket, we expect Priest to be in a strong position overall even as the FOTM changes all around him.

We find Druid to be the general catalyst for change at this point; as the general arena population takes firm grasp of new larger-bracket comps involving Druid, we see gains for Rogue, reductions in Mages, and of course a dip in the populations of other healers. In the 2v2 bracket, we should see Druids holding steady, maybe even dipping slightly, but this should be more than offset by their growth in 5v5.

Graphical Timeline

Showing values relative to the balance point of 100.

Power Ratings: Rogues Creeping Upward

no.spam@gameriot.com (briyan) — Mon, 03 Mar 2008 00:29:15 -0500

This week's update to Briyan's Arena Power Ratings shows few changes -- not a big surprise considering we just started tracking one week ago.

Today's Arena Power Ratings

Rank	Class	Power
1	Warrior	110.11
2	Priest	108.09
3	Rogue	103.39
4	Warlock	101.21
5	Druid	100.57
6	Mage	96.46
7	Paladin	95.89
8	Shaman	89.76
9	Hunter	86.37

A value of 100 is considered balanced. Please reference the first post in this blog for more details on how I calculate these power ratings.

Rogues Creeping Upward

The only real class to watch over this past week has been Rogue. We have seen them reach the #1 class position in 3v3, and this dominance has helped Rogue create a little separation from Druid and Warlock in the middle of the Power Ratings pack. We expect to see this trend continue as the masses jump on the double melee bandwagon in 3v3, and more teams experiment with a Rogue in 5v5 after hearing of the BG9 tournament results.

Despite the ongoing buzz about upcoming patch 2.4, which is slated to include several significant PvP changes, we have not seen any real impact in the "live" arena representation. Given the number of bugs still on the PTR, it looks like we will have several more weeks before we can hope to see the patch finished.

Graphical Timeline

Now that we're into our second week, I will begin including a graphical timeline with each post, that gives you an idea of the change in Power Ratings over time.

Each class is represented by its standard Blizzard color, and I've graphed the Power Ratings relative to 100, the "balance" value.

Introducing Briyan's Arena Power Ratings

no.spam@gameriot.com (briyan) — Sun, 24 Feb 2008 17:52:17 -0500

The mass publication of World of Warcraft character data (via the Armory) has led to a revolution in data analysis among players. The area of greatest interest has been the general representation of each class in the competitive arena brackets, which is seen as a "measuring stick" for PvP power; these numbers have been the source of countless discussions since data mining websites started displaying them.

Addressing the need for a unified rating system

A few of today's major sources for arena stats include Vhairi's blog, SK-Gaming, and RealmHistory.net -- all of these provide a different look at the state of arenas, and it can be hard to get a good estimation of overall class power when looking at any particular page. We have three different-sized arena brackets, different cutoffs by team rating on each site, and some groupings (like the SK100) that are somewhat biased in what they report.

In creating my Arena Power Rankings, I have attempted to come up with a rating system that considers representation in all three arena brackets at the upper levels, and a handful of other effects (described below). The end result is a single number for each class that gives my estimate for the overall viability of each class in the WoW arena. As time goes on, the formulas will not change, so we can effectively measure the ebb and flow of class balance over time.

Components of the Arena Power Ratings

To compute my Arena Power Ratings, I start with the class representation figures provided on an ongoing basis by RealmHistory. These values give me a percentage for each class in both 2000+ and 2200+ in all three brackets. I then perform some combinations and adjustments to arrive at my final result:

* The Role Model Effect (extra weight for elite ratings)
* The Large Team Bias Effect (compensation for big-bracket success)
* The Unfriendly Bracket Effect (compensation for erratic performance)
* The Taint Effect (brackets weighted by perceived purity)

The end result is a single value for each class which is balanced around the number 100. A value of 100 shows a reasonably balanced representation across all brackets, and higher values signify overpoweredness. Classes under 100 are weaker than they should be.

The Role Model Effect

Even though the total number of players at 2200+ is very small, we feel that it represents a greater significance than its total proportion of the population. These elite rated players represent those who have pushed their class to its limit, and lower numbers represent a frustration at the high end that may not be evident if we simply look at the 2000+ numbers. To account for this, I combine the 2000 percentages and the 2200 percentages with a 3:1 ratio.

The Large Team Bias Effect

The competitive WoW community is somewhat biased against the 2v2 bracket for various reasons. Queue times are terrible, it rewards fewer points by rating, it frequently causes extremely long games, it is viewed as "less skillful", etc. -- the bottom line is that 2v2 is a generally unfavorable bracket. I argue that if a player is able to find success in the larger brackets (3v3 or 5v5), he is less likely to compete as fiercely (or at all) in the 2v2 bracket. To account for this, I take a weighted average of representation in the 3v3+5v5 brackets and compare it to representation in the 2v2 bracket. I adjust the 2v2 representation up or down based on the size of the discrepancy -- this slightly increases the 2v2 numbers for a class like Paladin and slightly decreases the 2v2 numbers for a class like Druid. Classes with similar representation in both large and small brackets, such as Warrior and Hunter, are largely unaffected.

The Unfriendly Bracket Effect

If a class shows large differences in representation between brackets, we view it as a sign of greater imbalance. As we are building an combined rating across all brackets, we feel a class with fairly even performance in all size brackets is stronger than a less consistent class. I address this by using a multiplicative approach to combining representation from all three brackets into a single number.

For example, if we were looking at values 10,20 vs. 15,15 ... both add up to 30, but if we multiply together, we get 200 vs. 225. In my ratings, this is scaled exponentially to temper its effect, but this is the general behavior of my bracket combination routine.

The Taint Effect

The data we are able to mine from the arena is somewhat tainted by the strange behavior of the Armory in addition to the erratic team-hopping behaviors of players. The smaller brackets are hit hardest by this effect, as they are home to the majority of point sellers and less serious teams where a player may jump around and help several of his friends play games each week. Due to this, I slightly scale down the overall contribution of the 2v2 and 3v3 brackets in my final number.

Today's Arena Power Ratings

I will be publishing these values each week, and maintaining a graph of the values over time for comparison purposes. Since I have just started this project, I have only one set of values.

Rank	Class	Power
1	Warrior	110.46
2	Priest	108.39
3	Rogue	101.19
4	Warlock	101.12
5	Druid	100.49
6	Paladin	96.53
7	Mage	95.36
8	Shaman	90.63
9	Hunter	86.75

Remember, all values are based around 100 = balanced.