I'm going to try to zero in on the core issue here. For the most part, I think we agree on the premises and we see many of the same issues. The difference is we wildly diverge on what we think are the appropriate solutions to those issues. To use a concrete example, you made the following statement:
both because you would no longer have undead (or indeed other racial distinctions, as a secondary consequence)
Among other things, I'm guessing one of your "racial distinctions" would include critical immunity for undead. You're using removing critical immunity from undead (and from everything, for that matter) as a way to indicate a problem in your mind, while I'd say "Bingo, nothing should have crit immunity!" What you view as a negative I view as a positive, and that's the issue which needs to be sorted out.
Generally speaking, I've seen two main methods of attempting to balance games. For lack of better terms offhand, I'll dub them "Homogenization" and "Rock/Paper/Scissors." The question becomes when to apply one method versus the other because each is appropriate in certain circumstances. As you might guess, in a game like NWN I'm suggesting the Homogenization route is better while you're advocating the R/P/S route. I suspect this is because you've never been in an environment where balance is essential and your second post would seem to confirm that suspicion. This will require some explanation.
HomogenizationAt the most basic level, the idea of Homogenization is that everything has the ability to do equally well in all cases (or as close to it as possible). The simplest way to achieve this is making everything exactly the same. If the only class in NWN was fighter and stats/feats/skills/weapon choice were pre-determined, everyone would be balanced. Differentiating between players would depend on *how* they play their character.
This is the case in many/most FPS games. It's also how RTS games started out (different factions were effectively mirrors of each other, potentially with different names for some units). RTS games have generally evolved beyond this, but the reason they are *able* to do so is something I'll discuss in the R/P/S section. Note that even RTSes have Homogenization at the meta level. To use Starcraft (II) as an example, Zerg, Terran, and Protoss are vastly different. But if you pick Terran, you generally have an equal chance to win whether you're facing Zerg, Protoss, or another Terran.
As seen above, Homogenization does not mean things have to be exactly the same. Rather, it means that each option has nearly equivalent capability at everything. It means a caster cleric and a mage can put out basically the same damage via spells, for example, but do so with different methods.
But how about some examples with actual data? Let's turn to my primary playground, World of Warcraft.
I've never played WoW, so I can't really comment on it specifically, other than to say that many of our players who have played WoW have remarked on the fairly limited set of class roles it offers. You seem to be mis-extrapolating from my remarks, however. If 'damage spam' is the only behavior that is rewarded, which I doubt, then WoW would be definitionally imbalanced. If, however, 'damage spam' is the only behavior for a GIVEN class that is rewarded, you could still have balanced play, if fairly tactically uninteresting play.
There are three roles available. Tanking involves getting the attention of monsters and absorbing damage (plus moving said monsters around and other things - the main gist is that tanks are generally the only ones who get hit by non-AoE or non-randomly targeted attacks), healing involves healing damage that enemies do to the group, and damage involves doing, well, damage to the enemies. There are also other factors that are fight specific (such as moving out of a flame strike, ringing a gong to deafen a blind dragon that hunts with sound, handling debuffs the boss gives out, etc), but those roles determine nearly everything and are by far the most important factor in terms of what you'll be doing during a fight.
Each class has 3 specializations available (4 for druids) that determine role. There are 5 tanking specializations, 6 healing specializations, and 23 damage specializations. Every class has at least one damage specialization and there are four classes that have nothing BUT damage specializations. For these damage specializations (which make up 60%ish of groups for the most part), assuming you execute fight mechanics correctly (fire is hot, ring the gong, use defensive abilities when appropriate, etc) your performance metric boils down to a number: DPS. DPS means damage per second.
The following link is to a site called World of Logs:
http://worldoflogs.com/In essence, people do a raid (group PvE content for 10+ players) and upload their combat log. The combat log breaks down the duration people were in combat versus bosses and figures out what they did during the boss fights and provides the data in an analyzable form. One of the main uses of the data is determining the DPS of each person. There's another website that analyzes the analysis:
http://raidbots.com/dpsbot/The numbers to the fight of the chart are percentile comparisons of specialization performance. I'm going to throw out the lowest six specializations for the moment (if you put your cursor over them, they are Demonology/Survival/Destruction/Unholy/Frost/Arms). The reason for this is that each of those specializations has a better DPS specialization within the same class. As a result, when people care about their performance, they do not use those specializations. Those scores could be higher than they currently are but the best people don't ever bother doing it because there's a better option (in fact, currently there are two damage specializations, Subtlety and Marksmanship, that don't even appear on the chart because there's not enough people using them to get any data: but the classes which have those specializations have two OTHER damage specializations).
As an analogy, imagine a fighter with feats for both Bastard Sword and Longsword (assuming default weapons). Given otherwise identical weapons, there is no reason to ever use a Longsword except to mess around, because it's definitely inferior, even if it's not by very much.
Looking at the current numbers at the time of this writing, the highest is 93.8 (Affliction) and lowest is 75.2 (Elemental). To do some rough math, the average of those is 84.5. The highest and lowest damage specializations are within roughly 11% of that average. Or to rephrase it, 15 different damage specializations are within 11% of the average (and the outliers are all lower). Even including all but the two not even on the chart, we get everything within 20% of the average. Note that I'm using multiplicative percentages because that's actually accurate (if someone takes 10% of incoming damage and someone else takes 5% of incoming damage, claiming they're 5% apart is misleading, to put it kindly).
To put this in perspective for a moment, if we assume that ((AC - 20) < AB < (AC - 2)) then a single point of AB is worth between 8.3% more hits and 40% more hits for a non-dual-wielder (more if we factor in Epic Dodge). If we average those numbers, we wind up with roughly 24%. But let's say we think that's on the high side, so we'll divide the number in half again, which gives us 12% (an AB of AC - 11 gives a 14% bonus, as a basis of comparison).
Let's rephrase that: gaining or losing 1 AB in NWN is more of a difference than the difference in damage output for the top 15 (out of 23) damage specializations in WoW. This is despite those 15 specializations having different abilities, playstyles, gearing needs, and more (if you don't believe me when I say that the specializations are very different, I can provide more details, but different specializations within one class in WoW are more different than entire classes in NWN). That's pretty homogenized for damage output.
Furthermore, while some specializations are slightly better on single target versus a few targets versus heavy AoE, Blizzard's goal is to crunch these differences to be pretty small. For example, a class that excels at attacking two targets at once might do 10-15% more damage when attacking two targets - which is the difference of 1 AB in NWN. Yet that difference is absolutely *massive* in WoW.
Here's a recent tweet from the Lead Systems Designer of WoW indicating what I'm talking about...]https://twitter.com/Ghostcrawler/status/282207519462543360]Here's a recent tweet from the Lead Systems Designer of WoW indicating what I'm talking about..."Yes, but during that time we've also seen the community shift from thinking a 10% dps delta is unacceptable to a 5% or 1%."
In other words, even with DPS differences generally being within 11%ish, people want it to be smaller and that gap messes things up at times. More on this later.
I am glossing over a lot of finer details regarding the WoW stuff because they're not really important and I doubt you want a treatise on WoW mechanics.
Rock/Paper/ScissorsRock/Paper/Scissors throws the idea of equality out the window. It doesn't matter if you like the idea of Rock more than Paper, you'll lose to Paper if the other person chooses it. It's not because you're worse at the game, it's because that's how it works. Paper counters Rock. Rock counters Scissors. Scissors counters Paper.
A game which clearly features this is Pokemon. Fire does 200% damage to Grass and takes 50% damage from Grass attacks. Grass does 200% damage to Water and takes 50% damage from Water attacks. Water does 200% damage to Fire and takes 50% damage from Fire attacks. Fire beats Grass beats Water beats Fire beats...it's a cycle.
Starcraft (II) is another game that features R/P/S at the unit level. The Terran siege tank, for example, is a powerful artillery unit which is devastating at long range and deals AoE damage. But it is helpless against air units. Every unit has something that can counter it, though the game features both hard counters (unit A utterly annihilates unit
and soft counters (unit A has a noticeable advantage over unit
. Air units versus siege tanks is a hard counter. A soft counter would be that Hydralisks (ranged Zerg unit) beat Battlecruisers (heavy Terran air unit). One Hydralisk won't defeat half a dozen Battlecruisers, but cost for cost the Hydralisk will easily win ($100 of Hydralisks will beat $200 of Battlecruisers or something).
When we look for this type of gameplay in NWN, we see it in a few ways. The most obvious are damage immunities/resistances. A fire elemental is resistant to fire but weak to cold. A skeleton is resistant to piercing and slashing. A outsider resists a flat amount of most damage types, meaning you need a large number to punch through. There are also miscellaneous immunities. Sneak immunity counters rogues, assassins, and blackguards. Critical immunity counters the above mentioned (by default, at least) along with crippling Weapon Masters and anyone with a character built around criticals. To do some rough math, a weapon master with a keen longsword deals 180% weapon damage on average (assuming every roll that threatens critical actually criticals), meaning he deals 56% of normal damage versus a crit immune foe. If the server has modified Overwhelming Critical and/or Devastating Critical, this becomes even more pronounced. You also have other random things like turning undead and spells that target undead specifically.
Why R/P/S Works in Pokemon and Starcraft (II)Okay, so why I suggesting that R/P/S works in Pokemon and Starcraft (II) but fails in NWN?
In Pokemon and Starcraft (II), you control multiple units and none of them represent you.
That's what it boils down to. Everything else winds up deriving from that statement. In Pokemon, you control a team of up to six Pokemon and in Starcraft (II) you control an entire army. In both of these cases, the composition of the army depends on what you choose to capture/produce.
In essence, the goal of the games is to figure out what the opponent has and use a composition to counter it. If your enemy has 6 Water Pokemon...bring 6 Grass Pokemon. There is zero reason to bring a Fire Pokemon to that battle. Likewise, if your enemy decides to use nothing but air units...you build anti-air units. The units you use can also shift mid battle for a game like Starcraft (II). If your opponent goes heavy air and gets crushed by your anti-air, he might then swap to units that counter your heavy air. Therefore you have to swap to units that counter his countering units.
But the units don't care. They're pixels on a screen. Whether you have all air units, all siege tanks, all infantry, or a mix matters to no one but yourself. You use the most effective composition you can figure out to crush your opponent. And you don't have to explain to the siege tanks why you aren't using them when your opponent is focusing on building air units.
The Problem of Difficulty and WoWWoW used to be more like NWN. Fire elementals were resistant to fire. Undead were immune to bleeds. Golems didn't care about poison. As previously mentioned, the pinnacle of WoW PvE is raiding involving groups of 10+ people fighting bosses. Specifically, each raid zone had a maximum group size (which was 10, 20, 25, or 40 at different parts of WoW's lifespan).
The initial content (such as Molten Core) wasn't tuned very well. Molten Core could be done with 20-25 people and was meant for 40. Blizzard has gotten far better at raid design since then and produces some tremendously difficult content. It took over 500 attempts for the best guild in the world to defeat a boss in a zone called Firelands, for example. Even guilds like my own, which only spend two nights a week raiding (as opposed to 6-7 nights a week in the top guilds) still took over 100 attempts. That was with us knowing the strategy, having better gear, and being in the top 0.5%ish of the WoW population skill-wise.
To beat the bosses on the hardest difficulty, you need a full raid with very skilled players. With a 10 man raid, for example, you're not going to win with 9 players, at least not initially. You might be able to get away with it months after you originally beat it with much better gear, perhaps, but that's pretty much it. The content is tuned to be on a razor's edge where every percent matters. All of the heroic raiding guilds in WoW have stories of wiping continuously at <5% on a boss while they struggled to find those last few percent. You didn't have room for people who couldn't pull their weight.
You can probably see where this is going.
Rogues, for example, depend on poisons for a large chunk of their damage. What do you think happened when groups went up against bosses that were immune to poison? They didn't take rogues. They benched them and took someone else on their raid team. Or even made the rogue swap to a different class. If the rogue lost 20% of his damage from the boss being immune to poison, the group was better off with the rogue swapping to a warrior or something because even with worse gear the warrior would do like 10% more damage than the rogue. In fact, the top guilds typically require their players to maintain several "raid ready" alternate characters that they can play if it is advantageous in a situation.
In short, the effect of the "racial distinction" of being immune to poison simply resulted in excluding people. Boss immune to bleeds or poisons? Tough luck if you're a warrior or rogue, hope you like cheering on your team from the sidelines or playing another character. Boss immune to Shadow (Negative) damage? Hope you're not a warlock or shadow priest! Boss immune to Fire damage? If you're a mage, you may be lucky and be able to swap to Arcane or Frost instead of Fire, but there's a decent chance you just won't be taken instead.
Let's take four players with classes A, B, C, and D respectively. If there are four bosses, and only A can harm the first, only B can harm the second, only C can harm the third, and only D can harm the fourth, you might argue that's balance since all are "useful." But what actually happens is that you bring four of class A to boss one, four of class B to boss two, four of class C to boss three, and four of class D to boss four.
You might think I'm joking or exaggerating. I'm not. There was a boss about a year ago called Spine of Deathwing. You were literally on a colossal dragon's back trying to pry up armor plates. However, you had a 20 second window to deal damage to the plates at a time. Guess what happened? Guilds brought people who were best at burst damage in a 20 second window. If the class you played didn't excel at that, enjoy the bench! You needed immense burst and there was little to no slack.
There were 10 classes in WoW at the time. In a 25 man group, you'd expect 2-3 per class. The first guild to kill the boss had something like 8 rogues and 7 mages because those classes had the best burst. Those two classes were supposed to make up like 20% of the group and instead made up 60%. Similar patterns (though not quite as extreme) occurred for the other top guilds. My guild is more casual, but even we benched our two Hunters (think Arcane Archer with a pet) because they did like 10% less damage than our other players in that 20 second window solely because of their class. And that 10% on two players was huge, the difference between successfully prying off a plate or failing with it at <5% health.
This is the effect of "racial distinctions" in a competitive environment where balance is essential. It excludes people for no reason beyond what class they happen to play.
The Death Magic ConundrumYou seemed irked at the idea of "damage spam" being viable in every scenario. You also suggested the idea of
Consider instead 15 firebrands vs 3 firebrands, 2 chain lightnings, 5 ice storms, a clarity, a heal pot, an elemental shield, a dimension door, and 2 more brands.
Clarity, heal potions, elemental shield, and dimension door are not primary spells you'd spam. You use them occasionally as the situation dictates. So that leaves us with Firebrands versus "3 firebrands, 2 chain lightnings, 5 ice storms, and 2 more brands." Which brings up the question of why you would suddenly want to do 2 chain lightnings and 5 ice storms if Firebands are the best spell to use at the start. Assuming you're attacking the same mob, unless he suddenly shifts his damage immunities or something there would be no reason to change spells.
In general, there seems to be three main "alternatives" for casters besides "damage spam."
1. Buffing/healing/dispelling allies
2. Debuffing/CCing/dispelling enemies
3. Death Magic
Note that I'm not including summons because I've yet to see anyone try to module where you literally summon a monster every 3 seconds. Summons are not a "primary" spell type as a result.
Let's look at Death Magic more closely to see some issues with it. The two primary issues are its binary nature and randomness.
Death Magic is off or on, alive or dead. The only things that matter are Death Magic Immunity and high saves. Let's compare this to other things.
For something like a stun or paralyze, you have immunity to those effects, high saves, and your effective hit points (effective hit points is your HP coupled with damage immunities and resistances, so a character with 400 HP and 50% immunity to all damage has 800 effective HP). In other words, even once stunned/paralyzed your enemy still has to wear down your hit points.
For damaging spells, you have high saves and your effective hit points. Having more HP/immunities/resistances increases your survivability along with better saves.
As you might notice, Death Magic is the odd one out here. It ignores your effective hit points. Doesn't matter if you have 100 HP and 0% immunity or 500 HP and 50% immunity.
But even that isn't so much an issue in and of itself, the problem is coupling that with its randomness.
Imagine a mage faces a pack of 10 mobs. He has access to two spells: AoE death magic and AoE damage magic. Using the damage spells, he will kill the group in 10 spells. Let's even pretend we think the mage should kill the group in 10 death magic spells (the fact that mobs dying means they're not attacking means killing a few off the bat can extremely significant and thus it could be argued it should take more than 10 death magic spells). But we don't know if he will. He could theoretically kill all of them in one spell. Or none could be dead after six spells.
We've effectively recreated the problem of levels 1-2! Where a single hit can kill many characters and it's a matter of luck who wins instead of being a matter of playing well.
If you make a boss fight where it is assumed that you'll spend X time killing an monster during the fight, then if said monster is vulnerable to death magic it could die in one spell or might take 2X (or more) time to kill. Obviously killing it in one spell means you can attack the other monsters far sooner which makes the fight much easier than intended. Or it could make the fight far harder than intended if they keep saving successfully.
There's a reason melee characters which do an average of 50 damage per hit tend to do something like 45-55 damage or 40-60, not 1-99 (and even that 40-60 number is heavily skewed towards the middle). Some randomness can keep things interesting, but if you make it too random you unbalance the environment. And Death Magic as typically seen in NWN is inherently unbalanced for these reasons.
ConclusionNow you may be thinking "Hang on, we have crit immune monsters and people still take Weapon Masters to those fights!" That may be true. But you have one of two options.
One, the fights are difficult enough that a person having their damage cut in half means they can't be taken. Or even losing 10-20% damage. This is the environment where balance actually matters.
Two, the fights are easy enough that you can take a person who has their damage cut in half. In this environment, balance doesn't really matter since there's loads of slack anyway. Or alternatively, weapon masters still are powerful enough without their criticals that they're clearly not balanced regardless!
Your point about the AB highlights this. If we assume ((AC - 16) < AB < (AC-1)) initially, then let's look at the effect of losing 4 AB.
At AC - 16, we get 25%/5%/5%/5%/25% = 0.65 hits per round. At AC - 1, we get 95%/75%/50%/25%/95% = 3.4 hits per round.
Now let's subtract your 4 AB.
At AC - 20 we get 5%/5%/5%/5%/5% = 0.25 hits per round. At AC - 5, we get 75%/50%/5%/75% = 2.3 hits per round.
So losing 4 AB is a loss of somewhere between 60%ish damage to 33%ish damage (or reversing it, a gain of
48% damage to 160% damage). Even going with the lower bound, that's a massive difference.
And in a balanced environment, even one or two AB would be the kiss of death you mentioned (again, see the Ghostcrawler tweet above). This is why I asked how many groups wiped because one person had 1 less AB than needed. That stuff happens in WoW all the time. It happens in any environment where tuning is tight and balance is essential. And if things aren't tightly tuned, then balance isn't all that important and you're free to tack on "racial distinctions" because they don't really matter anyway.
Modifié par MagicalMaster, 09 janvier 2013 - 04:41 .
