Skip navigation

Talent analysis of World of Warcraft Raiders

Author: Dylan Northrup (Docx on Icecrown)

Executive Summary

Looking at the top raiding guilds for the World of Warcraft, several class talents are identified as “must have” raiding talents by analyzing the talent choices of characters in top rated guilds which have completed the most challenging raiding achievements currently in the game. The raw data indicates ‘Meditation’ (Priest), ‘Dual Wield Specialization’ (Rogue) and ‘Armored to the Teeth’ (Warrior) are “must have” talents. Four Hunter talents are indicated in the category as well, but the sample size of hunters and variation in spec makes it difficult to declare them “must haves” with conviction.
Many talents are indicated as “not worth it” for raiding as they are not taken by characters whose primary spec is in the same tree as the talent.


Note: The audience for this discussion is intended to be one somewhat conversant with the terminology and slang of the World of Warcraft community. While some terms will be defined, there is no guarantee of this being the case.

Hypothesis

I have raided as a Priest (Shadow spec) and Druid (Balance spec) in both the Burning Crusade and Wrath of the Lich King expansions. While my main is currently my Boomkin, I still keep an eye on the discussions of Priests and changes that may or may not occur with them. I saw a post[1] from Greg Street who posts under the pseudonym Ghostcrawler on the official WoW forums. Ghostcrawler is the lead systems designer for World of Warcraft and a part of the team that makes game design decisions for WoW. The original poster in the forum thread indicated that the Priest’s Meditation talent is so vital to the class that Priests should not be forced to spend valuable talent points on acquiring it. During the thread, Ghostcrawler responds to one of the posters who said:

“A very good definition of a bad or poorly designed talent is one that everyone takes or that nobody takes. It’s either too good or too bad.”

Ghostcrawler’s response, in part, was:

“I don’t think Blizzard would agree with that. If nobody ever takes the talent, then yes, there is probably something wrong with it. But I don’t know that we would ever want to see say an Arms warrior build that didn’t take Mortal Strike. We don’t want to make attractive alternatives to Mortal Strike. We pretty much want you to take Mortal Strike. The same is true of Titan’s Grip for Fury and Toughness for Protection.
“You can come up with examples for classes too. Omen of Clarity is near mandatory for all druids. Ruin is something every PvE warlock wants.”

With this as an inspiration, I set out to determine what percentage of Priests spec for Meditation, what percentage of Druids spec for Omen of Clarity and what percentage of Warlocks spec for Ruin. In addition, I sought to determine if there were any other “must have” talents for other classes. Finally, I set out to find if there are talents in specific trees that are not taken by players who choose points predominantly in those trees to identify talents deemed “not worthy”.

Given my experience with the Priest and Druid class, I believed Meditation and Omen of Clarity would be highly chosen by those classes regardless of the player’s spec. With little experience with the Warlock class, I could not make a prediction about how highly Ruin would rank for Warlocks independent of spec.

Methodology

Gather the top names of the top 200 raiding guilds from a guild ranking site. Using those guild names, gather membership lists from the WoW Armory site. Examine the Achievement history for those members and, if they have completed “Heroic: The Spellweaver’s Downfall” (indicating they had defeated the Malygos raid encounter) and/or “Heroic: The Twilight Zone” (indicating they had defeated Sartharion in the hardest configuration), get their talent specialization and add it to the data set for analysis.

For each character identified as having “won the game”, their talent spec was categorized based on how many points they had spent in a specific talent three. For many talents, there are a variable number of points that can be spent in that talent (eg. a Priest can spend 1, 2 or 3 points in the Meditation talent with each additional point giving a proportionally greater benefit). For each number of points for each talent in each class, a total was kept of how many characters in that class chose that particular number of points in the talent. These totals were compared against the total number of players of the class to determine the percentage of players who spent a particular number of points for each talent.

This analysis was also done for each talent specialization to determine which talents in a particular tree were not chosen by characters who had chosen their talent specialization from that particular tree (eg. Balance talents not chosen by Druids who had primarily spent their talent points in the Balance tree).

For both of these criteria, I made a note of any talent that fell within the criteria at least 95% of the time. The intent for focusing on these particular is to determine talents that fit into one of two categories:

1. Talents that are chosen by 95% of players of a class regardless of spec.
2. Talents within a tree that are ignored by 95% of a class that are spec’d into that particular tree.

Results

Here are the totals of the individual classes identified through the search criteria along with the breakdown of numbers for the individual specializations within each class:

Death Knight – 135 Druid – 94
Blood – 32 characters Balance – 20 characters
Frost – 14 characters Feral – 42 characters
Unholy – 89 characters Restoration – 32 characters
Hunter – 61 Mage – 56
Beast Mastery – 1 characters Arcane – 16 characters
Marksmanship – 3 characters Fire – 33 characters
Survival – 57 characters Frost – 7 characters
Paladin – 273 Priest – 391
Holy – 143 characters Discipline – 92 characters
Protection – 50 characters Holy – 134 characters
Retribution – 80 characters Shadow – 165 characters
Rogue – 118 Shaman – 369
Assassination – 73 characters Elemental – 95 characters
Combat – 33 characters Enhancement – 84 characters
Subtlety – 12 characters Restoration – 190 characters
Warlock – 204 Warrior – 313
Affliction – 170 characters Arms – 6 characters
Demonology – 23 characters Fury – 140 characters
Destruction – 11 characters Protection – 167 characters

Here are the results of the talent selection for each class and for the individual specializations:

Death Knight: Blood – 32
100% no points allocated by Death Knight: Blood into Vendetta
Death Knight: Frost – 14
100% no points allocated by Death Knight: Frost into Acclimation
Druid: Restoration – 32
96% no points allocated by Druid: Restoration into Furor
Hunter – 61
98% adoption by Hunter of Careful Aim 3 points
98% adoption by Hunter of Lethal Shots 5 points
98% adoption by Hunter of Mortal Shots 5 points
96% adoption by Hunter of Survival Instincts 2 points
Hunter: Beast Mastery – 1
100% no points allocated by Hunter: Beast Mastery into Catlike Reflexes
100% no points allocated by Hunter: Beast Mastery into Improved Aspect into the Monkey
100% no points allocated by Hunter: Beast Mastery into Improved Mend Pet
100% no points allocated by Hunter: Beast Mastery into Pathfinding
100% no points allocated by Hunter: Beast Mastery into Thick Hide
Hunter: Marksmanship – 3
100% no points allocated by Hunter: Marksmanship into Concussive Barrage
100% no points allocated by Hunter: Marksmanship into Improved Concussive Shot
100% no points allocated by Hunter: Marksmanship into Piercing Shots
100% no points allocated by Hunter: Marksmanship into Rapid Recuperation
Hunter: Survival – 57
98% no points allocated by Hunter: Survival into Counterattack
96% no points allocated by Hunter: Survival into Deflection
96% no points allocated by Hunter: Survival into Savage Strikes
Mage: Arcane – 16
100% no points allocated by Mage: Arcane into Arcane Fortitude
Mage: Fire – 33
100% no points allocated by Mage: Fire into Blazing Speed
100% no points allocated by Mage: Fire into Burning Determination
100% no points allocated by Mage: Fire into Fiery Payback
100% no points allocated by Mage: Fire into Firestarter
100% no points allocated by Mage: Fire into Impact
100% no points allocated by Mage: Fire into Molten Shields
Mage: Frost – 7
100% no points allocated by Mage: Frost into Elemental Precision
100% no points allocated by Mage: Frost into Frost Channeling
100% no points allocated by Mage: Frost into Frozen Core
100% no points allocated by Mage: Frost into Improved Blizzard
100% no points allocated by Mage: Frost into Improved Cone of Cold
100% no points allocated by Mage: Frost into Improved Water Elemental
Paladin: Holy – 143
99% no points allocated by Paladin: Holy into Purifying Power
98% no points allocated by Paladin: Holy into Seals of the Pure
Paladin: Protection – 50
96% no points allocated by Paladin: Protection into Guardian’s Favor
96% no points allocated by Paladin: Protection into Stoicism
Paladin: Retribution – 80
96% no points allocated by Paladin: Retribution into Deflection
Priest – 391
97% adoption by Priest of Improved Inner Fire 3 points
98% adoption by Priest of Meditation 3 points
Priest: Holy – 134
99% no points allocated by Priest: Holy into Blessed Recovery
99% no points allocated by Priest: Holy into Blessed Resilience
99% no points allocated by Priest: Holy into Searing Light
98% no points allocated by Priest: Holy into Spell Warding
Priest: Shadow – 165
95% no points allocated by Priest: Shadow into Psychic Horror
Rogue – 118
96% adoption by Rogue of Dual Wield Specialization 5 points
Rogue: Assassination – 73
95% no points allocated by Rogue: Assassination into Deadened Nerves
97% no points allocated by Rogue: Assassination into Improved Expose Armor
100% no points allocated by Rogue: Assassination into Master Poisoner
100% no points allocated by Rogue: Assassination into Remorseless Attacks
Rogue: Combat – 33
100% no points allocated by Rogue: Combat into Deflection
100% no points allocated by Rogue: Combat into Improved Gouge
96% no points allocated by Rogue: Combat into Improved Kick
100% no points allocated by Rogue: Combat into Lightning Reflexes
100% no points allocated by Rogue: Combat into Mace Specialization
96% no points allocated by Rogue: Combat into Nerves of Steel
100% no points allocated by Rogue: Combat into Riposte
Rogue: Subtlety – 12
100% no points allocated by Rogue: Subtlety into Filthy Tricks
Shaman: Elemental – 95
100% no points allocated by Shaman: Elemental into Elemental Devastation
96% no points allocated by Shaman: Elemental into Reverberation
Shaman: Enhancement – 84
98% no points allocated by Shaman: Enhancement into Anticipation
97% no points allocated by Shaman: Enhancement into Earth’s Grasp
97% no points allocated by Shaman: Enhancement into Earthen Power
98% no points allocated by Shaman: Enhancement into Guardian Totems
100% no points allocated by Shaman: Enhancement into Toughness
Warlock: Affliction – 170
99% no points allocated by Warlock: Affliction into Frailty
95% no points allocated by Warlock: Affliction into Improved Felhunter
Warlock: Demonology – 23
100% no points allocated by Warlock: Demonology into Improved Enslave Demon
95% no points allocated by Warlock: Demonology into Improved Imp
100% no points allocated by Warlock: Demonology into Improved Succubus
Warlock: Destruction – 11
100% no points allocated by Warlock: Destruction into Improved Searing Pain
Warrior – 313
96% adoption by Warrior of Armored to the Teeth 3 points
Warrior: Arms – 6
100% no points allocated by Warrior: Arms into Sword Specialization
Warrior: Fury – 140
99% no points allocated by Warrior: Fury into Improved Demoralizing Shout
Warrior: Protection – 167
97% no points allocated by Warrior: Protection into Improved Bloodrage
98% no points allocated by Warrior: Protection into Improved Disarm
98% no points allocated by Warrior: Protection into Improved Disciplines
97% no points allocated by Warrior: Protection into Safeguard

Analysis

A detailed, class-by-class and talent-by-talent analysis is beyond the scope of my current time constraints[2]. However, of the three spells indicated in the initial Hypothesis, the adoption of Meditation with it’s full number of points is greater than the adoption of Omen of Clarity and far greather than that of Ruin.

Meditation: 98% adoption with 3 points (384 out of 391 Priests sampled)
Omen of Clarity: 93% adoption with 1 point (88 out of 94 Druids sampled)
Ruin: 78% adoption with 5 points (161 out of 204 Warlocks sampled)

Meditation has been said by the raiding community to be necessary and the raid progression of those who
have chosen Meditation appears to back up that analysis. Omen of Clarity is also held in wide regard, though the smaller number Druids in the sample size does not give as much weight to this assertion. Ruin is regarded well, but not a necessary talent based on this analysis.

For Raiding Warriors, Armored to the Teeth has a very high representation. A similar case exists with the Rogue talent Dual Wield Specialization. In both cases, the talents are deemed as “must have” talents due to their high adoption by successful raiding characters.

While the results for Hunters indicate there are four “must have” talents, Hunters are also the most skewed class in terms of representation with the vast majority spec’d primarily into the Survival tree. While there are three talents in the Marksmanship tree, this could be an artifact of a Survival build that overly benefits from 13 points in those specific talents in the Marksmanship tree.

Future Improvements of Analysis Methodology

Certain classes or class/spec combinations are under-represented in this analysis. Specific examples are Beast Mastery and Marksman Hunters, Subtlety Rogues, Destruction Warlocks and Arms Warriors. This can be attributed to those specific class/spec combinations having fundamental focii that are not geared toward end-game raiding. While a larger sample size may increase the data sets for the under-represented classes and specs, it should be acknowledged that the choice of criteria for a character having “won the game” being that of completing the end-game raiding scenarios will bias the analysis towards talents that reward Raiding, or PVE, and away from talents that may be beneficial to Arenas and Battlegrounds, or PVP. If this is desired, changing the criteria could provide that information.

Data gathering in general is limited by polling rates to Blizzard’s Armory site and the Armory’s availability in general. If the Armory is down, data collection cannot be done. If data polling is done at too fast a rate, the Armory automatically black lists the data gathering mechanism. Because of this, data is cached for a period of time and gathered only when it becomes too stale. For this analysis, a period of 7 days was identified as acceptable to produce the desired data while staying within Armory access parameters.

Armory data only updates when a character logs out of WoW. Because of this, talent specs may have changes between when the character got one of the indicative end-game achievements and when their data was polled. It is assumed the spec they use is that they used while Raiding and has not been changed for other purposes (Talent experimentation or PVP being the two primary possibilities). A way to improve the analysis would be to have a method to snapshot character information when they got their achievement (or while exiting the instance immediately after they got their achievement). This would also allow more granular analysis on an encounter-by-encounter basis to determine whether there are variations between encounters within the same instance on specific talent specializations that are more useful than others. Doing the analysis when the character zones out would likely be more straightforward to implement from a technology perspective. Both of these scenarios would require direct access to Blizzard’s reporting databases across all realms and is unlikely to occur for anyone not employed directly by Blizzard.

The external site[3] used to gather the top 200 sites ranks guilds based on specific achievements and when they were attained. Based on this, the ranking benefits guilds who attain achievements earlier rather than later. As a result, it would minimize “cookie cutter” specs as the guilds who overcome content earlier would be less likely to copy “proven” specs. If performing the data collection outside of Blizzard, using the Armory as a primary source with historical talent tracking would help identify the specs used when the raider attained the specific achievement. If performing the data collection with Blizzard’s assistance, the recording of talents when the achievement was attained would help avoid confusion if a player used the incoming dual specialization changes to change their specialization before logging off.

All data used in this analysis was done for US servers. This was an implementation choice. Data for EU servers could be added to the analysis and would simply require coding time and effort. Given the author’s external time constraints, this is not something likely to happen in the near term.

Conclusion

Offering suggestions for changing of talents is outside the scope of this paper. The author does not have information from Blizzard to know whether the development team believe near universal adoption of a talent is something they wish to avoid, if they are satisfied with or somewhere in between. The point of view articulated by Ghostcrawler seems to indicate the development team is content with some talents being universally valued.

While it may be the case that the dev team is content with near universal adoption of specific talents, it is very likely those talents are monitored by Blizzard and evaluated periodically to determine whether it would be appropriate to baseline the abilities by incorporating them into the class, or by modifying class spells to achieve the same effect.


[1] “Blizzard, it’s time. Baseline Meditation” at http://blue.mmo-champion.com/27/14990442254-blizzard-its-time-baseline-meditation.html accessed 3 March 2009

[2] Though, I’d be happy to discuss any open Operations or Engineering positions Blizzard has available with appropriate parties at Blizzard.

[3] http://www.wowprogress.com/