by Brian Urbanek
Twenty points of energy equals twenty five points of damage, and
players should survive about twenty seconds of trading punches against
an enemy of approximately equal power.
That's about it.
Well, okay, there's a bit more to it than that.
The first step to balancing a game is generating your baselines.
"Baselines" are average, expected values of just about everything. We
have baselines for expected hit points, attributes, damage done,
resources gathered, and most every other stat, metric, or value you
Generating these baselines is half art, and half science. The "art"
part comes from figuring out what kind of game you want to make.
Different genres and play styles lend themselves towards different
target ideals of "good balance," and hence valid baselines. A game
that's supposed to simulate a gritty, street level police drama, where
a gunshot is extremely dangerous, requires very different assumptions
(and different math and systems) than a game that's about super heroic
Once you think you have a good grip on the fuzzy "art" part of your
baselines, comes the hard math of making sure they feed into each other
in a way that makes sense.
Putting this into concrete Champions terms, let's look at attributes,
life expectancy, and DPS. So, first, the "art" part. There's no hard,
math-based particular reason why Champions characters at creation have
stats that range from about eight to twenty, or why I set the "target"
peak value of a stat at level 40 to be 200. Basically, I thought the
numbers just looked nice. They're understandable, they allow sufficient
whole number resolution and sufficient room for growth and variance.
Really, we could have made the range from 1 to 10, or 73 to 9,436, and
had everything else in the game could still be made to scale the same
way, and the game could have had exactly the same "feel" that it does
right now. If it's not yet clear why that's true, keep reading, and
I'll put it all together by the end.
align="right">The next step was working out the rate at which a
player's expected stat value would grow. The simplest answer is what's
called a "linear progression". That would be where I take the expected
peak at level one (about 20), subtract that from the top value (200),
and divide the result by the number of levels over which you can earn
those points (39), yielding a stat max growth of about 4.6. Well,
that's a perfectly valid and satisfactory number, but it doesn't quite
hit the right flavor. The nature of an MMO, particularly a (super)
heroic MMO, is such that players want to be able to resoundingly defeat
enemies that are lower level than you, but enemies that are just a few
levels higher than you should be significant challenges, and you don't
really get that with a linear progression.
What we actually want is a system whereby the increase at each
individual level is slightly more than the increase from the level
before. That kind of behavior is exactly what is modeled by an
"inflationary system." An inflationary system is described as: <base
value> + <base increment> * <number of increment steps>
* (1 + <inflation rate>) ^ <number of increment steps>.
That may look kind of long and scary, but it's actually pretty easy to
work with in Excel. All I have to do is figure out how much inflation I
want (and that's an "art" question; I used 1%), and it's easy to plug
it in and work out that with a base increment of just slightly more
than three, I get exactly the kind of curve I want, and the target stat
baseline grows from a 3 point difference between levels 1 and 2, to
more than a 6 point difference between levels 39 and 40.
Of course, it's nowhere near enough to just declare that the target
value for a player's stats is 76 at level 16; other systems have to
support that to make that a reality. This, really, is why baselines
exist - to inform the systems that generate them. Now that I know how
much of a given stat I want a player to have, I have the first piece of
information I need to figure out how many stats players can get from
So, back to "art." Champions is a super-heroic game, and to me, that
meant that most of a player's potential and power should come from
their own power, their own choice, and their own superhuman abilities.
Of course, as a competing tension, I want gear to be meaningful and
exciting - just not quite as all powerful and influential as it is in
other games. In the end, I negotiated out with Antiproton (our items
and rewards lead) for items to get 42.5% of the player's total stat
budget. Super stats got 40%, talents get 15%, and "base" value is 2.5%.
Those are the proportions that "felt" right after a lot of discussion
and testing, but they're far from the only valid choices we could have
So, now we have a new baseline value; a player should be able to get up
to 85 points in a given stat from their gear at level 40. This is what
we call a derived baseline, or more properly a one step derived
baseline. In on other words, unlike the arbitrary choice of "200" for
stat peak at level 40, this value is not arbitrary; it's calculated,
and it can't be changed. I could go back and change the precedent
factors that generated it, such as changing the stat peak at a given
level, or changing the itemization budget, but if those values don't
change, I can't change this value without invalidating everything else
that derives from those precedents. (I'll show what I mean by that in a
The next step after generating the item value baseline is to put it to
work, and derive actual values for actual items. In order to do that,
we need to have an understanding of what kind of gear, how much gear,
and what the quality of that gear is at each level.
In Champions, we expect players to make multi-dimensional progress
through their gear. That is, they not only get higher level gear as
they level up, but they also get higher quality gear. This idea of
quality is represented by the color bands; white, yellow, green, blue,
purple. At level one, we expect players to be in level one whites. At
level 21, however, we expect players to have graduated completely out
of whites, and be in an approximately balanced blend of yellows and
greens. It's certainly possible to do better than that. Even a solo
player that did lots of mission content, did all the open missions (and
did well at them) could have more greens than yellows and maybe even a
blue or two. That's not, however, who we're balancing for. We're
balancing for the "average" player. Doing so means that we can feel
confident that an "average" player has a good, balanced experience, and
that an exceptional player gets to feel extra strong as a reward for
their hard work.
This table, the "desired build" table, was created by feel; we
hand-entered values and ratios to make a gear progression that we felt
would give a slow but steady improvement. The effect of that table,
however, is dramatic. Every drop table in the game as well as the
schedule of rewards for missions that our content designers use is
driven by this table, and the rates and values of drops are calculated
with brutal specificity to generate drop rates that actually match what
our "desired build" requests. Actually, the drop rates are very
slightly rigged in the player's favor, to allow players with just a
tiny bit of luck or determination to do better than the table mandates.
The point of baselines isn't to limit players or crush their potential;
it's to establish the minimum expected values that we eagerly hope
players will work to exceed.
Now that we've set up the basic expectations, next time I'll delve more
into combat balancing and how it all ties together.
To read the latest guides, news, and features you can visit our Champions Online Game Page.