Page 1 of 1

Some numbers on how people spend their talent points

Posted: Thu May 03, 2007 1:42 am
by Amuna
This was posted on the wow forums. It's a good read :)

The Armory is a beautiful thing. I have been curious lately about talent distributions and talent popularity -- what 31/41 point talents are popular, and which ones aren't? The data is out there -- thanks to the Armory. The only challenge was mining it.

So, I wrote a program that looks up talent specs on Armory, and I talled information about 31 and 41 point talents.

My program pulls names from allakhazam's database, then looks them up on Armory. I'm only interested in level 70 players. I have 6635 players in my study; you can see the breakdown here:

Code: Select all

        Druids:    673
        Hunters:  1019
        Mages:     933
        Paladins:  438
        Priests:   776
        Rogues:    837
        Shamans:   315
        Warlocks:  761
        Warriors:  883
        TOTAL:    6635

Some players exist on allakhazam but not Armory (perhaps they are inactive or deleted/transferred?) and some classes have higher representation since there are more profiles of that class on allakhazam.

The data I was interested in is about players who specialize in a talent tree -- that is, players who spend the majority (31 or more) of their points in one tree. I wanted to know what the trends are for which players specialize in a tree, but DO or DO NOT take the 31-point and 41-point talents.

# The first column indicates the percentage of that class that specialized in the tree -- for example, 59.9% of all druids in my study put 31 or more points into Feral Combat.

# The second column indicates the percentage of specialized players, that SKIPPED the 31 point talent. For example, 0.5% of all Feral druids (31 points or more) in my study skipped Leader of the Pack.

# The third column indicates the percentage of specilized players, that did not take the 41 point talent. For example, 2.7% of all Feral druids (31 points or more) in my study don't have Mangle. They may have gone 0/31/30, or perhaps they went 0/61/0 but skipped Mangle.

Code: Select all

        Class/Tree            % Specced     % Skipped 31  % Avoided 41
        Druid/Balance           13.4        6.7                24.4
        Druid/Feral Combat      59.9        0.5                 2.7
        Druid/Restoration       26.7        2.8                21.8
        Hunter/Beast Mastery    22.0        2.2                 9.8
        Hunter/Marksmanship     72.3        0.4                12.2
        Hunter/Survival          5.1       30.8                61.5
        Mage/Arcane             38.5        3.6                73.8
        Mage/Fire               35.9        3.9                18.8
        Mage/Frost              25.5        0.4                11.3
        Paladin/Holy            67.0        2.4                23.2
        Paladin/Protection      20.6        3.3                10.0
        Paladin/Retribution     12.1        5.7                 9.4
        Priest/Discipline       12.9       23.0                75.0
        Priest/Holy             53.7       76.7                83.7
        Priest/Shadow           32.7        0.0                 3.2
        Rogue/Assassination     23.7        1.0                21.2
        Rogue/Combat            56.2        1.7                18.1
        Rogue/Subtlety          17.7        4.7                33.1
        Shaman/Elemental        24.8        5.1                48.7
        Shaman/Enhancement      38.5        4.1                22.3
        Shaman/Restoration      36.6        2.6                 0.9
        Warlock/Affliction      46.0       10.0                27.5
        Warlock/Demonology      35.0        5.3                17.0
        Warlock/Destruction     18.2        8.0                35.5
        Warrior/Arms            28.7        0.4                60.3
        Warrior/Fury            13.1        4.3                23.5
        Warrior/Protection      58.1        2.5                 4.1

What conclusions can we draw from this?

What talent trees are the least popular?

Survival Hunters
Retribution Paladins
Discipline Priests
Fury Warriors
Balance Druids

What 31-point talents are unpopular?

Holy Priests (Lightwell)
Survival Hunters (Wyvern Sting)
Discipline Priests (Power Infusion)
Affliction Warlocks (Dark Pact)

Lightwell stands out the most. More than 3/4ths of Holy priests SKIPPED it. They were already heavily invested in Holy, so all they had to do was move one point from a lower-tier talent into the 31-point talent.

What 41-point talents are unpopular?

Holy Priests (Circle of Healing)
Discipline Priests (Pain Suppression)
Arcane Mages (Slow)
Survival Hunters (Readiness)
Arms Warriors (Endless Rage)
Elemental Shamans (Totem of Wrath)

These talents may be unpopular for a variety of reasons -- perhaps the talents themselves are bad. Or, perhaps there are other alternatives 21 points into another tree that are more attractive, which is often the case for Arms Warriors and Holy Priests. Or maybe the "filler" talents to get to the end of the tree are undesirable.

Let's look at the data in another way:

Percentage of players of (class) that took the 31 point talent in (tree)
Percentage of players of (class) that put at least 31 points in (tree), but didn't take the 31-point talent

Code: Select all


        Class/Tree            % Taken 31      % Skipped 31
        Druid/Balance	      12.5            0.9
        Druid/Feral Combat    59.6            0.3
        Druid/Restoration     26.0            0.7
        Hunter/Beast Mastery  21.5            0.5
        Hunter/Marksmanship   72.0            0.3
        Hunter/Survival        3.5            1.6
        Mage/Arcane           37.1            1.4
        Mage/Fire             34.5            1.4
        Mage/Frost            25.4            0.1
        Paladin/Holy          65.4            1.6
        Paladin/Protection    19.9            0.7
        Paladin/Retribution   11.4            0.7
        Priest/Discipline      9.9            3.0
        Priest/Holy           12.5           41.2
        Priest/Shadow         32.7            0.0
        Rogue/Assassination   23.5            0.2
        Rogue/Combat          55.3            1.0
        Rogue/Subtlety        16.9            0.8
        Shaman/Elemental      23.6            1.3
        Shaman/Enhancement    36.9            1.6
        Shaman/Restoration    35.7            1.0
        Warlock/Affliction    41.4            4.6
        Warlock/Demonology    33.1            1.8 
        Warlock/Destruction   16.8            1.5
        Warrior/Arms          28.6            0.1
        Warrior/Fury          12.5            0.6
        Warrior/Protection    56.7            1.5

We see that Lightwell stands out dramatically here. 41.2% of all priests in my study are Holy Priests Without Lightwell. This is more than an order of magnitude more than any other 31 point talent.

What does this tell us?

That the priest Holy tree is highly desired, but the 31 point talent is not.

Some food for thought.

Posted: Thu May 03, 2007 3:34 am
by beowolf
didn't we all know lightwell sucked? and bliz is laughing at you?

Posted: Thu May 03, 2007 10:22 am
by Cathas
interesting info

Posted: Sat May 05, 2007 12:42 pm
by Lyese
Someone with too much time on his hands and, a programmer to boot.

Posted: Sat May 05, 2007 1:45 pm
by Kwag
while it's unfortunate that the 31 point talent is crap, the fact that priests keep speccing holy suggests that the other talents are strong, whereas a lot of the trees with low %s of people even spending 31 in them would seem to be a bigger problem, as the community is seemingly saying, "this whole damn tree sucks."

Posted: Sat May 05, 2007 3:08 pm
by Amuna
EH the guy that posted this obviously had an agenda about lolwell but The numbers are relevant for everyone and I just enjoyed the observations you can make with them.

Posted: Sat May 05, 2007 4:01 pm
by Caffeinated
it also shows that only 5% of hunters go survival spec showing that that is a very weak talent tree

Posted: Sat May 05, 2007 5:29 pm
by Ulwar
Yeah, survival is great for PvP, but is really lousy for raiding. Even BM isn't great because so much is dependent on your pet surviving. So, if your pet get's one-shotted, or like in many groups doesn't get healed, you can get your DPS nerfed mid-fight. 2.1 changes the hunter's Mend Pet to a HoT instead of channeled, so pets can be kept alive easier without sacrificing DPS. I'd be far more inclined to send the pet in to attack and throw a HoT immediately. Guess I'm going to need more mp5 gear :)