Log in

No account? Create an account

Previous Entry Share Next Entry
GDC 2010: Design in Detail: Changing the Time Between Shots for the Sniper Rifle from 0.5 to 0.7 Sec
Bar Harbor
Design in Detail: Changing the Time Between Shots for the Sniper Rifle from 0.5 to 0.7 Seconds for Halo 3
Speaker/s: Jaime Griesemer (Bungie Studios)
Day / Time / Location: Thursday 1:30- 2:30 Room 130, North Hall
Track / Format: Game Design / Lecture
Description: Halo's multiplayer is balanced across over 1000 objects, more than 40 of which are unique weapons including the Sniper Rifle, which has over 200 functional fields including one that determines the minimum time between shots. In Halo 3, that time changed from 0.5 seconds to 0.7 seconds, changing less than 0.00001th of the overall game data, an immeasurably tiny balance tweak that should not have been noticeable, let alone significant.

This session will address this design decision in exhaustive detail. Why 0.7 seconds? Why that particular field? What processes and design principles lead to that change? How was it proposed, tested, and evaluated? What were its effects on the game's balance? How were those effects evaluated to the effects of the infinite number of other changes that could have been made? What were the external considerations, such as community reaction or target demographics, that influenced it? And what can be learned from that change to improve our ability to make changes in the future?

[Alexx: I came in about halfway through this panel, due to planning to be at another panel which turned out to be useless.]

tool roles must be unique, and must all be filled
this is why almost all games have similar weapon sets

Sweet spot for complexity of role interactions:
rock, paper, scissors, lizard, spock is a bad game
so is rock, paper.

shouldn't have crappy tools, or dominant ones

practice iterative deletion
if you don't need it, cut it

good balance is between chaos and certainty
balance means that outcomes are probable, but NOT certain

Yomi - reading mind of your opponent
1. I know what I will do
2. I know what he will do
3. He knows what I will do in reaction to what I will do
STOP HERE! That's plenty complicated enough!

feedback loops un-balance

Varied roles means designers don't have to rely on single tool
Not every tool will be on every map

balance your core before your spice
Once you are done, cut half

sniper rifle has a specific cadence
not too slow: don't let fingers get bored
not too fast: see the results of every shot

causality exaggerated
time is artificially stretched out, reactions are amplified

play the final game in your head
see past bugs and missing content
vocalize your own sound effects!
but never be satisfied
criticize every little hitch in the weapon experience

play b-games with an open mind, find what they got unexpectedly right.
play a-games critically, find out all the things they got wrong

flow IS design
it's the thing that designers do that no other discipline does

change drastically when tuning! Small changes get lost in the noise
don't make changes <10%
change stuff in groups; tweaking a single variable in isolation will never get you to the perfect solution

find the core elements of the tool, which you cannot change, then crank those elements up to 11.
For sniper rifle example, range and deadliness. If it isn't deadly at long range, it isn't a sniper rifle. So make it *really* deadly, and *really* long range.

prerequisites for balance
need solid code -- game can't be full of bugs that prevent a valid game experience
Needs to run at at least 30fps, to reflect shipping conditions
get good fx, animations, sound in as early as possible

balance requires strengths - all tools must be strong
affordances - what the strengths are must be clear to the player
strengths are things you can get better at - possibility of improvement is an implicit motivator

make changes constantly during the balancing process
prevents inertia
prevents brain optimization, where you (and testers) get too good at the game in a particular state
dopamine is easy to mistaken for balance - are you happy because you achieved balance -- or just because you scored well that round? Being bad at the game can help prevent this!
Don't specialize in one weapon or play style, or you will overinvest

play a lot when the weapon WORKS, not when it's clearly broken. If you *know* it's broken, you're no longer learning anything useful about it; when it's working, you can still get new insights.

don't be reactive

if a particular sub-feature is not part of the tool's flow or strengths, cut it

don't tell playtest what you've changed lately

mixed strategies among playtesters is an indicator of good balance

fix things players can directly perceive, not mathematical averages. If sniper rifle is perceived as overpowered, lowering total ammo carried or clip size will lower the sniper's overall DPS -- but the victim still perceives the weapon exactly the same.

every fight is unfair

Example problem:
expert playtesters are using sniper rifle as a close quarters weapon
this is opposite to the intended role!
There are no viable counter-strategies; the sniper can kill people too fast for them to reach cover.
Can't touch strength knobs
finally have to touch flow knobs
0.5 - 0.7 seconds