Thursday, February 26, 2009

Balancing DPS and Utility

Araxes, the Ratonga Warlock, is not very happy about the level of damage done by EQ2's utility classes. I obviously have absolutely zero idea how any of these classes perform in EQ2 raids since I'm churning along in the mid 20's (though the fact that I'm voluntarily playing utility classes instead of pure DPS, even though I don't intend to do any EQ2 raids, may lend some support to the claim). I can say that I've seen this complaint elsewhere, Araxes gets the link love because he's the first one I've seen to tie the same complaint into three of the game's four archetypes. (I'm guessing that none of the healing classes do enough damage to compete in raids.)

The reason why this particular post caught my eye is because it's the same exact discussion we're having in WoW, using slightly different terms. Blizzard used the Wrath expansion to rebalance and consolidate the various group buffs in the game, bringing the damage of hybrid and utility classes up to compensate, under the new "Bring the Player, Not the Class" mantra. What's so hard about this problem that two of the top PVE games on the market can't tackle it?

The Utility DPS problem
Let's have a hypothetical raid group with 10 DPS slots - doesn't matter what game. Tanking and healing are irrelevant, as these guys all require the same amount of healing, and don't have any buffs or debuffs that affect survivability. The only thing that anyone does is A) damage or B) buffs/debuffs that increase damage.

To start with, we have no buffing or debuffing class, which gives us:
10 damage dealers, each does 100 DPS, 1000 total raid DPS.

Now let's add in an overpowered utility class, who buffs everyone in the raid for 2 DPS.
9 damage dealers (100 DPS), 1 utility (100 DPS and 2x10 DPS from the buff), 1020 total raid DPS

Obviously, that situation is a problem, because there's no reason NOT to replace as many DPS as possible with utility classes. (Doubly so if the utility guy actually does MORE damage, which apparently is possible in EQ2 at the moment.)

Either you have to give EVERYONE an equal buff to compensate, or you have to nerf the utility guy's base DPS. Giving everyone buffs is pointless, as everyone would be a buff class. At that point, you would want to stack your raid with exactly one of each so that you can collect as many buffs as possible - this is what happened to WoW in the TBC era that caused the current revamp. You might as well just raise everyone's damage and call it a day. So now we're at the point where we started, looking at calibrating the DPS of utility classes.

Who is buffing whom?
To return to my example from above, perhaps you're looking at it and thinking of nerfing the utility guy by 20 DPS. Now we would have:
9 damage dealers (100 DPS), 1 utility (80 DPS + 2x10 from the buff) = 1000 total

That sounds like a balanced solution, but now the question becomes who exactly is buffing whom. If a DPS member of the original raid with 10 DPS dies, the raid loses their 100 DPS. If a DPS member of the raid with the utility guy dies, the raid loses 102 DPS, because the utility guy's contribution to the raid is dependent on there being exactly 10 people (including themselves) doing damage.

Then the math gets even more complicated. Say that my deliberately chosen numbers are now percentages, and the buff is now 2% instead of 2 DPS. Now say you bring along TWO utility players with 80 DPS and 2% buffs that stack. Suddenly, they're both under 100 DPS because 2% of 80 DPS is less than 2% of 100 DPS.

The point being, the overall contribution of the utility guy to the raid has to be GREATER than the contribution of bringing another generic DPS, or it won't be worth bothering. So now say that you're going to buff the buff to 5 DPS (back to non-percents). This leaves:
9 damage dealers (100 DPS), 1 utility (80 DPS and 10x5 from the buff) = 1030 total

Now the devs need to go back and rebalance the raid around 1030 DPS, a 3% increase for the raid, which will mean that groups will HAVE to get a utility player to beat the content, but at least there IS a point to being a utility character.

Magnifying the tie to group size
Unfortunately, now we've hardwired our class balance around group composition. That utility character will now overperform in a group that, for whatever reason, includes more than 10 DPS slots, and will underperform in a group that doesn't include at least 4 DPS slots. You can fix the overperforming angle by capping the number of players that can be affected by a single utility ability - for instance, WoW mages have a talent called focus magic that puts a spell crit buff on a single lucky recipient. (EQ2 characters have five "concentration" slots which limit how many buffs they can have active, but it seems that most buffs hit your whole group.) The downside is much harder to solve.

Say that you have to solo through 80 levels and then gear up in groups and raids half the size of our elite endgame raid - two factors that WoW and EQ2 share. Now that utility guy is struggling through the early game at 85% of the DPS of a DPS class, and then is an actual liability to any group with fewer than five DPS. These are the kind of issues that can make it difficult to find enough of a certain class you need to fill your group out with.

One way to address this would be to have the buff scale upwards in smaller groups. For example:
1-2 recipients: 15% buff (utility guy does 95% solo or in a duo, if there are two recipients the second guy does 115%, for a toal of 220%)
3-4 recipients: 10% (utility guy does 90%, DPS do 110% for a total of 310% in a 3-man group or 420% in a 4-man group)
5-10 recipients: 5% (the group is always slightly ahead under this plan, eventually scaling up towards the 3% boost)

Of course, now it actually matters that all your pure DPS classes actually do 100% DPS. In practice, some classes do more (with a risk of aggro or death), some classes do less, etc, and it now becomes advantageous to pick and choose who gets the buff - maybe 15% for your top guy is actually worth more than 3% for everyone. Or maybe you want to SKIP the top guy because he won't be able to benefit without pulling aggro.

Alternately, you can do as Blizzard has done and try to arrange it so that you need a small-medium number of buffs that are relatively widely available, and then balance the classes with less of an individual handicap under the thinking that the class might show up with a buff that the raid already has. However, the smaller the difference between pure DPS and utility classes, the greater the danger that the utility classes will overtake the DPS.

Either way, I think this is a harder problem to solve than just buffing the pure DPS class' numbers and calling it a day.

Aside on DPS meters
Personally, I think we're past the time when players should have to analyze log parses to determine their DPS. Games should come with DPS meters, and the meter should credit additional damage done as a result of a buff/debuff to the person who PROVIDED the buff/debuff (something that a player-run mod would not have easy access to).

That's not to say that meters don't cause negative effects like idiots jockeying for damage meter position, but all of these things are already happening anyway. Class balance or imbalance would be a lot more obvious if you could see exactly what was going on, and utility classes might be more inclined to accept lower DPS numbers if they actually got credit for doing their jobs in the first place. (The bigger issue is how to credit tanks and healers, and that's a much bigger topic than I can tackle in a paranthetical note to an addendum. ;))

2 comments:

  1. In the case of wow, and what most "pure" dps classes forget, especially in the age of aoe trash mobs, is that they have Crowd Control. For wow that is a great deal of utility that though currently ignored, won't be ignored forever.

    Hybrid classes bring the possibility of off tanking, off healing when needed, and those small saves are often ignored, not seen, or misunderstood.

    DPS feel that if they're not topping the meters, even though they had to keep one guy sheeped, for example, suddenly feel that they are inadequate.

    Ultimately the exact numbers for the tweaks are currently above my head, and Blizzard so far seems to be doing very well on it. Time will tell.

    ReplyDelete
  2. Thanks for the comment! Responded as a new entry because it became long.

    ReplyDelete

Comments on posts older than 14 days are moderated and will not appear until manually approved because the overwhelming majority of such comments are spam. Anonymous commenting has unfortunately been disabled due to the sheer volume of comments that are defeating Google's spam filter.