An analysis: Number of builds for each class.

  • #21
    I find it very interesting how the Demon Hunter has just one more active skill than the Witch Doctor, but has 187 billion more combinations. It's not really something you'd think about normally.
  • #22
    Personally, the fact that there are 2000 or 2 viable builds, is less important to me than that I can make the build I want to. For example, I I first played LoD, (not sure exactly when It was) I wanted to play a Fissure Durid with some summons. I learned the hard way that this was not possible and I almost put the game down at that point.

    Now I have about 5 builds for the Doc (which looks to play more similar to the Druid than the Necro IMHO) some are silly (only dog skills) some are not, and they tell me that I can take at least the serious builds and with the right gear, and a some tweaking (which I can do as I play my first character: How Terrible!) I can play that character. and it will be mine!

    I don't care that I can change to be exactly like your character. I don't want to, your character is stupid (as in not interesting to me). If you change your character to be identical to mine, that's your perogative and I may feel proud that my awsome unique build was so cool you wanted one too or I can not care and let you go on your merry way.

    In any case I love that there are hundreds of billions of builds fo each class and thank you OP for all the hard work.
    If that made sense to you, Bravo! I think I even confused myself...
  • #23
    Quote from Nivius

    my math used the bassis of using any skill at any slot, ONCE, same whit passives.

    example:
    passive1+passive2+passive3
    is same as
    passive3+passive2+passive1

    in my math :]

    Quote from Drsniper

    pC3 is the no. of ways choosing 3 items from a list of p items where the order does not matter. So it takes into account the fact that

    passive1, passive 2, passive 3 is the same as passive3, passive 2, passive 1.


    lolwut? xD thats what i said?
    Game Designer - Micro Design
  • #24
    Hey Everyone,

    *** This post and it's calculations holds true ONLY if you play NON-elective mode or default skill mode ***

    *** The math below represents the lowest amount of skill combinations because in the initial calculation, I do not consider rune variations *** (however I did calculate that too)

    I’ve seen a lot of math done out there to provide clarity on skill combinations (viable or not) comparing diablo 2 to diablo 3 when playing in non-elective mode. What we need to consider is that, in the case of the d3 barbarian who has 22 active skills, is that they are divided up into level sets. So, again in non-elective play, first time around you have to pick 1 of 3 and then on the next set you have to pick 1 of 4 and so on. In total, assuming your character level is maxed out, you have to pick 6 active skills. In another words, you can’t pick “bash” on your right mouse button skills or assign it to your number pads in non-elective mode. Once you pick one in that set, the others are locked out. So in the case of the barbarian there is exactly 22 active skills and then you get to pick 3 out of 16 passive skills. The problem and formula for it is as follows:

    If you have to choose 1 item from six different categories, and then 3 out of

    16 separate items, how many combinations are there?

    Set 1 Choose 1 of 3
    Set 2 Choose 1 of 4
    Set 3 Choose 1 of 4
    Set 4 Choose 1 of 4
    Set 5 Choose 1 of 4
    Set 6 Choose 1 of 3

    Set 7 Choose 1 of 16
    Set 8 Choose 1 of 15
    Set 9 Choose 1 of 14

    [(3x4x4x4x4x3)(16!/13!)] = 7,741,440 total combinations.

    There is 7 million, 741 thousand, 440 total possible combinations (not billions). Now what’s great about this is that the way the “sets” are set up makes a vast majority of these combinations viable… of course depending on the gear. Now others will take this farther to say the 5 rune variations (or 6 rune effects if you choose to use “no rune”…which is pointless) greatly magnifies the variations. Here are the results of that effect:

    [(15x20x20x20x20x15)(16!/13!)] = 120,960,000,000… not including the “no rune effect” option. This provides us with 15,625 times more variations. In other words, for every 1 of the original 7,741,440 combinations in equation 1, there is 15,625 variations due to rune effects.

    [(18x24x24x24x24x18)(16!/13!)] = 361,184,624,640… which does include “no rune effect” option. Likewise with this equation, for every 1 of 7,741,000 combinations, there is 46,656 variations due to rune effects including “no rune”.

    Its hard to say how many of those are going to viable, but the way the Blizzard has these skills separated and setup, it seems initially like an intelligent system in place to naturally have these 6 options work well together along with the 3 passive skills. Now we’ll have to see how uber the gear is setup, but we can assume Blizzard created gear to enhance player builds and options.

    Now on Diablo 2


    Now if we look at Diablo 2 and focus on the Paladin. The benefit of diablo 2 is that we know pretty much everything there is to know. It’s widely accepted that the Paladin has generally the largest amount of viable builds… somewhere around 10. Looking at www.diabloii.net under the strategy guide section, you can good a good list of effective builds. The Paladin has 10, the Barb has 9, the Necro 5, the Amazon 5, and the Sorceress has about 5. Now I know people will call foul here so for the sake of argument, lets double that number. Lets just say that the number of viable- hell difficulty Paladins is 20. Still not satisfied, how about 50 builds for the Paladin? Keep in mind that there are 30 skills total with no variation to them. Plus most, if not all builds, require that at least 3 skills are maxed, taking 60 points with about 12 points in pre requisites (in which you don’t have a choice there) out of the total available of 110. Most builds level up at least 1 or 2 of a level 30 skills and very few players use lower level skills as a primary attacker (Go Skeletons!) but instead for synergies. I haven’t seen too many hell driven fire-bolt sorceresses out there! The best skills are generally at the latter parts of the skill tree. In diablo 3, spell damages are based on the weapons and therefore its like having all level 30 skills, just unlocking at different increments, and are all viable based on the items you’re carrying. Diablo 2 has the illusion of choice in which over 99% of them are “wrong” choices. I mean you could put 4 points in every skill you have and you won’t make it far into nightmare, or even beat normal I’m not sure (experiment anyone?). The reality is, each class only has a handful of viable skills that work in hell. That was the flaw with diablo 2 and is simply the truth. Because of the nature of stat point allocations and skill point allocations, its near impossible for me to figure out the best formula to figure out total combinations.

    Conclusion

    But let me conclude that its not the number of combinations (in which diablo 3 still has more) but the viability and playability of those combinations. Blizzard has definitely without a doubt, succeeded in marking each class have at least 22 viable skills that can be used in any difficulty level. Why? Because it’s based mostly on your items stats. Now multiply that by rune effects and passive skills and all the sudden the truth is pretty evident… is that you can have 100 players playing and creating characters on battlenet and every single one of them will do something completely different. You still don't believe the math!! Well let me tell you this... if only .0005% of the Barbarian’s D3 skill combinations are viable, we still have 3,870 combinations that are worth experimenting with. Because there are no “bad” spells or “wrong” choices. The players that will be rewarded will be the ones that can masterfully apply their skills in the right situations and play for the long haul to acquire the equipment that supports their build. Now the difference here is that no one is “locked” in a build because anyone can switch to another build by simply clicking onto different skills. However on the flip side of things, by the time we get to hell difficulty in Diablo 3, our gear, more than anything will define our build. But that’s a different story. Hope this clarifies what Blizzard is saying.
  • #25
    Quote from toad1701a

    In another words, you can’t pick “bash” on your right mouse button skills or assign it to your number pads


    You can assign any skill into any button, you can pick bash + cleave + frenzy if you wish. Elective mode in options.
  • #26
    I really didn't know that! Thanks for pointing that out. I did play the beta on open weekend but I had not discovered that option. I was the one missing out on something!
  • #27
    Quote from toad1701a

    I really didn't know that! Thanks for pointing that out. I did play the beta on open weekend but I had not discovered that option. I was the one missing out on something!


    Hehe, well now you do, so maybe edit your misinforming post and it's math above ^^?

    Edit: Your conclusions seems valid though. And just to poke at you a bit, as for lvl 1 skills, the Necromancers 2 lvl 1 skills for skellys kick ass as does his amplify dmg curse which is also lvl 1, buy overall, yes the lvl 1 skills suck for most classes (also barb passives 1 of them anyway, and claw passive for assassin if your are that kind of assassin) ^^
    Winter is coming...
  • #28
    Quote from Antirepublican

    Quote from Tinkertank

    I did a comparison in another thread where I determined that a Necromancer in D2 had roughly 15k possible specs which ended up as maybe 50 variations that could be considered viable (less then 10 that were probably considered optimal) in Hell.

    So if 15k Necromancer specs gave us 50 viable builds in D2, what will D3 with it's over 500 billion possible WD specs give us ^^!?

    IDK but I call BS about those 50 viable specs....


    Lets take Summoner:

    One with only summons (which works, I just wouldn't play it, but it does work aka viable).

    So you max out skelly passive, skellys, mages, revives, and a golem + extra points into summon resist.

    You switch out any of the above for 1 curse that you like, that is 6 new builds all still viable.
    You now switch out 1 more of the above for CE or another curse that you like, that is 6x6 = 36 new builds. That I bet are maybe not ALL viable (for example if you chose to remove both the skelly passive AND the skellys). But you see the number of VIABLE (most of which I would stay away from, but you can still kill Hell Baal with them, and probably Ubers with some patience) builds rack up pretty fast. So 50 is maybe a bit low, it's just and estimation that I made after having played D2 for a few years and napkin math. This is also why I would claim that the Necromancer has the most viable builds (not optimal) out of the D2 classes. He is just more forgiving in his skills choices cause he has such awesome skellys ^^

    Edit: Spelling.
    Winter is coming...
  • #29
    I like the analysis, but this is like plotting out chess combinations: it really craps on the bigger picture, or in this case, the more valid one.

    Just because you computed permutations doesn't mean that all permutations are actually playable, in the sense that that will provide a considerable amount of utility to the player. So, I propose the following:

    Each skill gets an inherent value that mimics that condition when that skill is being used as a "primary", in the sense that it is the focal point of the player. Some skills will score higher than others:

    For example, consider a wizard using NOTHING but energy armor, or making a build around energy armor. Not much utility in that. Yes we could argue this, but I think the majority of the players WON'T be doing this because they want to kill shit.

    Then each of these skills gets a rune RANGE next to it, i.e. how each rune affects the inherent utility of the skill.

    Then this data set gets amplified further by its possible synergies, also a range that represents how that skill would work with other skills. I can't think of a lot of standalone skills, but there are some skills that clearly work better with others, such as the signature spell passive with the wizard signature spells. Thus for each skill, we would have the following array:

    ["Skill Name", InherentUtility, RuneLow, RuneHigh, SynergyLow, SynergyHigh]

    Finally we have to consider equipment; some skills will have a considerable benefit over others in that they utilize the equipment of the player more (think of the skills that use multipliers of weapon damage). I propose that be a multiplier between .9 and 1.1 for the all the integers in the dataset above.

    From there we pump it into a supercomputer and have it run a script that optimizes the each skill in relation with 8 others, and sums up the total utility/power of each build/combination. Long story short that's a lot of fucking skills, as many permutations as listed above.

    I'm willing to bet that there will be a significant utility curve, and that it looks largely like a chi squared curve, and in addition to that, only a certain threshold or higher will actually be viable for completing Inferno with a group.
  • #30
    Quote from toad1701a

    I really didn't know that! Thanks for pointing that out. I did play the beta on open weekend but I had not discovered that option. I was the one missing out on something!


    No worries. Almost no one finds elective mode without somebody pointing it out to them. People have been complaining about this to Blizzard, but their response is that they "are ok with it being tribal knowledge" because they don't want to "confuse" people by putting it in the skill UI. :/ This is one of the few decisions made for D3 that absolutely baffles me.
    ...and if you disagree with me, you're probably <insert random ad hominem attack here>.
  • #31
    Quote from lexwraith

    I like the analysis, but this is like plotting out chess combinations: it really craps on the bigger picture, or in this case, the more valid one.

    Just because you computed permutations doesn't mean that all permutations are actually playable, in the sense that that will provide a considerable amount of utility to the player. So, I propose the following:

    Each skill gets an inherent value that mimics that condition when that skill is being used as a "primary", in the sense that it is the focal point of the player. Some skills will score higher than others:

    For example, consider a wizard using NOTHING but energy armor, or making a build around energy armor. Not much utility in that. Yes we could argue this, but I think the majority of the players WON'T be doing this because they want to kill shit.

    Then each of these skills gets a rune RANGE next to it, i.e. how each rune affects the inherent utility of the skill.

    Then this data set gets amplified further by its possible synergies, also a range that represents how that skill would work with other skills. I can't think of a lot of standalone skills, but there are some skills that clearly work better with others, such as the signature spell passive with the wizard signature spells. Thus for each skill, we would have the following array:

    ["Skill Name", InherentUtility, RuneLow, RuneHigh, SynergyLow, SynergyHigh]

    Finally we have to consider equipment; some skills will have a considerable benefit over others in that they utilize the equipment of the player more (think of the skills that use multipliers of weapon damage). I propose that be a multiplier between .9 and 1.1 for the all the integers in the dataset above.

    From there we pump it into a supercomputer and have it run a script that optimizes the each skill in relation with 8 others, and sums up the total utility/power of each build/combination. Long story short that's a lot of fucking skills, as many permutations as listed above.

    I'm willing to bet that there will be a significant utility curve, and that it looks largely like a chi squared curve, and in addition to that, only a certain threshold or higher will actually be viable for completing Inferno with a group.


    I like what you said, but I would maybe go a different path (if I have the energy to do the math), and take one class and categorize the skills (separating their rune variations when applicable) and decide how many of each type are valid to have in the same build. Such as 0-2 primary / build, 0-3 defensive (pure survivability), 0-3 utility (indirect dps or survivability), 0-3 secondary, 0-6 passives. When you have this data you could easily calculate a more accurate estimations on the amounts of valid builds for that class. This would still not exclude some combinations such as if you pick this and this defensive then you will not be viable unless you have this or if you also pick that and so forth, but it should be a pretty good estimation.
    Winter is coming...
  • #32
    Already on it.

    Monk has the least permutations, and I'm already rewriting how the objects will be handled and have their attributes defined.

    Yes, I agree with you. Let's go ahead and cut out the permutations that make no sense, like all defense skills and passives.

    Let's define utility as survivability, which will be composed of 50% output of damage and 50% prevention of death. Some skills like Near Death, who have 0 synergy whatsoever, is inherently super useful in Inferno.

    I'm doing this in Python, want to help?
  • #33
    Quote from lexwraith

    Already on it.

    Monk has the least permutations, and I'm already rewriting how the objects will be handled and have their attributes defined.

    Yes, I agree with you. Let's go ahead and cut out the permutations that make no sense, like all defense skills and passives.

    Let's define utility as survivability, which will be composed of 50% output of damage and 50% prevention of death. Some skills like Near Death, who have 0 synergy whatsoever, is inherently super useful in Inferno.

    I'm doing this in Python, want to help?


    Never used Python, but I will check back here to check out your results, looks like it could be a fun read ^^

    Math CAN be fun! Why count apples and oranges when you can count dps and cow kills / mana ^^?
    Winter is coming...
  • #34
    Quote from toad1701a

    I will also like to acknowledge rune variations is just that… a variation on the same skill. A rune effect doesn’t make it a new different skill altogether but only a variation on a theme.


    I have to disagree with this. It varies wildly from skill to skill. While Leap runes are all pretty cosmetic variations, for a skill like Shock Pulse, there are 4 very different feeling functions depending on the rune you choose.

    Further, a solid build is created by finding synergies between different abilities. In D3, those synergies come much more from the rune options than the skill categories themselves. Even in a skill like Leap, wherein none of the runes changes THAT much about how it works and feels, you can choose to knock enemies back, pull enemies in, slow them, stun them or give yourself more armor. Each fits a very different build concept, and its that runes variety, not the skills that will define the iconic builds we'll all flock to emulate as we works our way through inferno.

    You cannot white wash the runes when you talk about build variation.
    I grew up gaming without internet forums. The entire phenomenon of being upset with a game developer makes no sense to me. No sense. I cannot imagine spending my time and energy being upset about something I choose to do for recreation.
  • #35
    Edit - Double post accidental
    I grew up gaming without internet forums. The entire phenomenon of being upset with a game developer makes no sense to me. No sense. I cannot imagine spending my time and energy being upset about something I choose to do for recreation.
  • #36
    Quote from Tinkertank

    Quote from toad1701a

    I really didn't know that! Thanks for pointing that out. I did play the beta on open weekend but I had not discovered that option. I was the one missing out on something!


    Hehe, well now you do, so maybe edit your misinforming post and it's math above ^^?

    Edit: Your conclusions seems valid though. And just to poke at you a bit, as for lvl 1 skills, the Necromancers 2 lvl 1 skills for skellys kick ass as does his amplify dmg curse which is also lvl 1, buy overall, yes the lvl 1 skills suck for most classes (also barb passives 1 of them anyway, and claw passive for assassin if your are that kind of assassin) ^^

    *** Edit ***
    Well the math still holds true to non-elective play but now that I know that we can basically put any active skill in any of the 6 allocated slots, it does greatly magnify the possibilities. Truthfully, I can't understand too many people playing non-electively so I will definitely modify the post to reflect it for non-elective mode. Playing without the elective option on still holds potentially much more play and character variation. Only 18 days left...
  • #37
    Quote from Tuna9719

    Quote from asfastasican

    People had fun in Diablo 2 making characters and figuring out how to spend roughly 100 skill points by allocating them into roughly 30 skill slots (per class), both active and passive. Some builds were serious business and some builds were just goofy builds where people had fun figuring out how to make them viable.

    In Diablo 3, people will shift through different combinations of skills that have very little identity. They aren't builds, just combinations. When somebody says they are making a crit barb or an aoe barb or bash barb, you will have absolutely no way of knowing what the hell their build actually is. It's just another combination of skills that does the same crap as the next guy's.

    Do all of the math you want. Go ahead. Slobber over a big passionless number that's at the end of an exponent equation. Big whoopidee doo. It's all generic combinations with very little identity. More doesn't mean better. Every Blizzard fan should know that by heart. Nobody gives a crap if your rogue build is different, They only care if your Combat rogue or your sublety rogue rips into flesh better than the other guy... or if you have the skills or rotation down pat to do so. Who would give a crap if your rogue build is skill combination #3450?

    After 6+ years of development, we basically have the same situation where every build has the same purpose of killing crap, while having also having every build have less identity. We might as well just go back to Warcraft 2, where every one of our skills could be perfectly balanced, similar to have every unit in WC2 was a mirror matched unit and they did the exact same thing. Outside of making, let's say, an MF run Demon hunter or a PvM Demon Hunter, what else is there?

    Diablo 2 Player: I'm leveling a whirlwind barb.

    Diablo 3 Player: Cool. Well you see... my barb will have bash and leap and ignore pain and whirlwind and...

    Diablo 2 Playter: Um... so that's a whirlwind barb then, right?

    Diablo 3 Player: Well, yeah... but not really. There's more to it then that. You have to see my skills to understand how it all works Plus my runes are...

    Diablo 2 Player: Sounds complicated and boring. Too long, didn't listen. Brb, pindleskin run.

    Diablo 3 Player: Screw you, noobtrash!


    Gems, gear sets and followers will have to be more in-depth to compensate. If not, you'll basically just have a maximum of five characters with an extended stash and maybe mule characters for extra storage. Why? Well every character can switch to another characters "build" simply by switching gear, at absolutely no cost. Almost 12 years have past and all we have done was go from a game where we could have multiple builds of each class (example. having a hurricane druid AND a shapeshifter druid) to a game we just have one character of each class (ex. having one druid that can just switch to any build whenever he wants just by switching skills and gear.)

    Math does not make a game fun long-term. If it did, all of you guys and gals would be Actuaries in real life having a great time making a good living.


    I've been reading on these forums for a while now and I've become so fed up with this argument that I've decided to actually make a post. Correct me if I'm wrong but it appears that you believe having to re-level a toon to respec is more meaningful than just openning your spellbook and swapping out abilities. I can see how this makes sense to some people since each toon's skills are permanent, and therefore each individual has more of a unique identity. (ie. blizzard sorc and meteor sorc).
    The fact that you are ignoring though, is that in D2 it took less that 5 hours play time to level a toon to 80. On top of that, and I admit this is merely my opinion, those ~5 hours felt like a necessary but annoying chore that I had to complete before I could have some real fun. As a result, you are actually arguing that an annoying ~5 hours creates a meaningful identity that a toon in D3 will lack. I for one am glad that Blizzard does not share your opinion.
    What really gives a toon a unique identity is the gear. I'd be happy to expand on this but I feel like I've already said too much for my first post.
    /rant off


    And asfastasican is rebuffed yet again. Don't worry, he won't respond. He never does. He just comes to shit on a conversation by bringing up the D2 vs D3 argument (which this topic is not about) and then leaves. He does not come back.
  • #38
    Quote from Theungry

    Quote from toad1701a

    I will also like to acknowledge rune variations is just that… a variation on the same skill. A rune effect doesn’t make it a new different skill altogether but only a variation on a theme.


    I have to disagree with this. It varies wildly from skill to skill. While Leap runes are all pretty cosmetic variations, for a skill like Shock Pulse, there are 4 very different feeling functions depending on the rune you choose.

    Further, a solid build is created by finding synergies between different abilities. In D3, those synergies come much more from the rune options than the skill categories themselves. Even in a skill like Leap, wherein none of the runes changes THAT much about how it works and feels, you can choose to knock enemies back, pull enemies in, slow them, stun them or give yourself more armor. Each fits a very different build concept, and its that runes variety, not the skills that will define the iconic builds we'll all flock to emulate as we works our way through inferno.

    You cannot white wash the runes when you talk about build variation.


    *** Thanks for the reply. I agree with you with the synergy aspect to runes versus the skills themselves and that it really does make a huge difference. I guess what I was trying to say is that the rune changes the skill but not on a fundamental level when calculating the lowest possible build combination and comparing to D2. But you have a good point. if you consider that In D3, "electrocute" is a skill and then a rune modifies to chain lightning. I There are some runes that greatly change the skill. In D2, lightning was a skill by itself and then at lvl 18 you could pick chain lightning. In this scenario you're point seems valid in that a rune could reforge a whole new skill and because of it's effect, change the very nature of how a player plays.
  • #39
    Quote from toad1701a

    Quote from Theungry

    Quote from toad1701a

    I will also like to acknowledge rune variations is just that… a variation on the same skill. A rune effect doesn’t make it a new different skill altogether but only a variation on a theme.


    I have to disagree with this. It varies wildly from skill to skill. While Leap runes are all pretty cosmetic variations, for a skill like Shock Pulse, there are 4 very different feeling functions depending on the rune you choose.

    Further, a solid build is created by finding synergies between different abilities. In D3, those synergies come much more from the rune options than the skill categories themselves. Even in a skill like Leap, wherein none of the runes changes THAT much about how it works and feels, you can choose to knock enemies back, pull enemies in, slow them, stun them or give yourself more armor. Each fits a very different build concept, and its that runes variety, not the skills that will define the iconic builds we'll all flock to emulate as we works our way through inferno.

    You cannot white wash the runes when you talk about build variation.


    *** Thanks for the reply. I agree with you with the synergy aspect to runes versus the skills themselves and that it really does make a huge difference. I guess what I was trying to say is that the rune changes the skill but not on a fundamental level when calculating the lowest possible build combination and comparing to D2. But you have a good point. if you consider that In D3, "electrocute" is a skill and then a rune modifies to chain lightning. I There are some runes that greatly change the skill. In D2, lightning was a skill by itself and then at lvl 18 you could pick chain lightning. In this scenario you're point seems valid in that a rune could reforge a whole new skill and because of it's effect, change the very nature of how a player plays.


    There are even better examples:

    For example the Corpse Spiders spamable nuke for the WD turns into a summon instead with the Spider Queen rune. Two very different skills. One is a primary, the other is a passive (or maybe even a utility if the Queen can be targeted by mobs).

    The Plague of Toads can be changed to be a Toad of Hugeness which adds a 5sec CD on the skill that swallows an enemy basically CCing it instead of being used as a primary nuke.

    So just cause some skills don't change that much, it's definitely not true for some other skills.
    Winter is coming...
  • #40
    Quote from toad1701a

    *** Thanks for the reply. I agree with you with the synergy aspect to runes versus the skills themselves and that it really does make a huge difference. I guess what I was trying to say is that the rune changes the skill but not on a fundamental level when calculating the lowest possible build combination and comparing to D2. But you have a good point. if you consider that In D3, "electrocute" is a skill and then a rune modifies to chain lightning. I There are some runes that greatly change the skill. In D2, lightning was a skill by itself and then at lvl 18 you could pick chain lightning. In this scenario you're point seems valid in that a rune could reforge a whole new skill and because of it's effect, change the very nature of how a player plays.


    You are the most civil and open-minded forum user ever. You just won the internet.
    I grew up gaming without internet forums. The entire phenomenon of being upset with a game developer makes no sense to me. No sense. I cannot imagine spending my time and energy being upset about something I choose to do for recreation.
  • To post a comment, please or register a new account.
Posts Quoted:
Reply
Clear All Quotes