97 billion builds per class @ Blizzcon

I realize there’s a lot of information spread around, I’m hoping to bring some of it to a single post and hopefully get our point across and reassure you that the changes we’re making are for the betterment of character customization options, and ultimately your long-term enjoyment of the game.

So, why did we get rid of skill points?
(Note: this is a supplementary min/max explanation. There are lots of other reasons which have been touched on in the past such as how players approach our game, supporting the idea of builds, observing how players behaved in internal testing, etc. This is just further explanation that I think will resonate with some of you.)

In Diablo III, we really want to improve the combat depth. Part of having combat depth involves having skills that are useful in different situations. In Diablo II players often used a single skill to deal with almost all situations: Blessed Hammer, Frozen Orb and Bone Spirit to name a few. Players invest 20 points into a single skill and use it as much as possible. The only reason a player would swap away from their primary spam skill is due to monster resistances/immunities. If a monster was immune to your primary spam skill, you’d either skip the encounter completely or fall back on a second skill. Neither of these answers provides the player with much combat depth.

To support combat depth, skills need to have different roles. Here is a very simple example:

• Magic Missile deals 15 damage to a single enemy
  
• Arcane Orb deals area of effect damage for 10 damage each

With these two skills we’re beginning to develop some combat depth for the player. Use Magic Missile when you’re facing one enemy, use Arcane Orb when you’re facing multiple enemies. But you may also want to use Magic Missile if one enemy is a “high priority target” in a group, and you want it to die quickly. In this simplified example players can still defeat a horde of enemies by casting Magic Missile multiple times, or they could defeat a single large enemy by casting Arcane Orb multiple times, but that wouldn’t be as efficient as a player who uses the right skill for the right situation.

Ok so that basic layout of combat depth out of the way!

With skill point spending your skills get better as you invest points into them. The problem is that this destroys combat depth. If after pumping a bunch of points into Magic Missile it now deals 70 damage to a single enemy, assuming my enemies have any reasonable health, then Magic Missile becomes a better choice than Arcane Orb even in group situations. If after pumping a bunch of points into Arcane Orb it now deals 45 damage, then it deals more damage than Magic Missile to single targets. Now rather than using the right skill for the right situation, I’m using the skill I’ve put all my points into. Skill point spending has eroded away combat depth.

Why did we go from 7 skill choices to 6?
(Note: again, this is a supplementary explanation. We’ve gone over some of the other reasons elsewhere, but this is specifically targeted at those of you here who feel strongly that 7 means there would be more build diversity than 6)

Diablo III emphasizes build customization. We feel that 6 skill choices actually creates more build diversity than 7.

Why? Well for any given set of options, the greatest number of combinations exists when the number of choices you can make is close to half the number of options you have. Some of you may remember a high school math problem like this: There are 12 differently colored marbles in a bag. How many different color combinations can you get by choosing X marbles? Well as it turns out the solution for various values of X are:

• 1 marble: 12 different color combinations
  
• 2 marbles: 66
  
• 3 marbles: 220
  
• 4 marbles: 495
  
• 5 marbles: 792
  
• 6 marbles: 924
  
• 7 marbles: 792
  
• 8 marbles: 495
  
• 9 marbles: 220
  
• 10 marbles: 66
  
• 11 marbles: 12
  
• 12 marbles: 1 (there’s only 1 way to choose 12 marbles from the 12 in the bag)

The greatest number of possible combinations happens when you are choosing 6 from a possible 12.

You may be asking what 12 has to do with anything as classes all have over 20 skills available to them...

This is true in theory, but in practice players tend to (and really should) pick up skills to fill different roles so they can be effective. Categories such as single target, area of effect, auto-targeting, debuff, defensive, group buff, escape, crowd control, 2-minute ubers, pet skills, etc. etc. Players generally take at most two (and often one) skill to fill any particular role. For example, the Wizard has Ice Armor, Storm Armor and Energy Armor, but I don’t think anyone is going to take all three (though maybe somebody will take that as a challenge and prove me wrong), most players will choose one Wizard Armor spell (note that this can change dramatically with some rune effects). If we look at each class, depending on how you count, you get anywhere from 8-12 different types of skills. So we err on the high side in our category estimate (12) and that means 6 is a pretty good number to maximize build variety.

It's important to note that we’re not just talking about you and your friend having Wizards with slightly different skills, we’re talking about you and your friend having 6 skills that are different in functionally significant ways.

Closing remark! When we pull math out like this I’m sure somebody will point out that if our only objective was to maximize build combinations we’d have allowed people to also choose 6, 7 or 8 passives rather than just 3. So I’ll counter by saying maximizing build combinations is not our only objective. We also want our system to have aesthetic flavor, to be simple to understand, and to have the passives in particular feel impactful. We have many different goals that we take into account when making any design decision. In the case of active skills, we felt the increase in variety was one of many good reasons to go from 7 to 6.

So how many skill combinations are there now?

Well taking into account 6 active skills, all the rune combinations, and 3 passives we currently expect each class to have roughly 2,285,814,795,264 different build combinations. That’s not taking into account skill types for ‘ideal’ builds, but that’s always been a big part of the fun of experimenting (and longevity for Diablo II) - finding a build that shouldn’t work, and making it.