Visual Game Tuning

nebraskaslowSoftware and s/w Development

Oct 31, 2013 (3 years and 10 months ago)

66 views

Visual Game Tuning

Integrating Interactive Visualizations into
Game Development

V

G

T

Why
InfoVis

& Gaming?


Games are increasingly:



Complex


Hard to debug


Hard to test


Expensive to
develop

Why
InfoVis

& Gaming?


Information Visualization can:



Help reduce complexity


Give insight into processes


Assist the debugging process


Reduce cost

Think of a game as a
multidimensional dataset

Drift!


the racing game

Drift!


the racing game


1992 Mazda RX
-
7


236bhp, 156mph top speed, 0
-
50mph in 4.4s


How much does this matter in a video game?


It depends on the game!


What
does
matter?


The “feel”

Drift!


the racing game


How do we define “feel” ?


Perceived: Rigidity, Acceleration, Speed, etc…


Reality means less here


How do we find the “correct” feel?


QA Cycles

Traditional QA Cycle

Testing Approaches


Bug
-
tracking systems (JIRA,
BugZilla
, Mantis)


Playtests


Subject Matter Experts


Heuristic Analysis (more on this later)

Traditional QA Process


If the tuning process works, why change?


The
average AAA
Game:


Costs $28 million to develop


3 years to produce


Requires constant testing by the QA staff


Requires constant revision by the developers

Modified
QA Cycle

Visual Game Tuning (VGT)

Game

Publish
Game State

Find
Reationships

Visualize

Interact

Publish
Changes

Visual Game Tuning (VGT)

Visual Game Tuning (VGT)

VGT Collaboration

VGT
-
vis

client

VGT
-
game

client

Steering is
too stiff!

VGT
-
vis

client

1. Overview

2. Zoom & Filter

2. Zoom & Filter


Zoom on most recent timeframe


3. Detail

3. Detail

4.
Look for potential relationships


5. History

VGT
-
game

client

Steering is
too stiff!

“Like”

This is too hard! Try
reducing speeds?

“Dislike”

5. History


Using Indicators, we can assess user feedback

6
. Extract & Save Groups

7. Update

VGT
-
game

client

Much
Better!

VGT Architecture

Visualization

Game

“Tuning”

Middleware

Visualization
-

Protovis


Stanford Visualization
Group


Web
-
based (
Javascript
)


High Level Abstraction


Declarative Syntax

Game
-

Unity3D


Cross
-
platform


High
-
level language support (C#,
Javascript
)


Web
-
enabled


High performance

VGT Collaboration

Tuning Middleware

(ActionScript
/
RTMP)




vis
client update protocol



game

client update protocol



Time


1:29:46

Group Name


Player

Object Name


Car1

Field / Value Pair


suspensionRange


0.1

Field / Value Pair


maximumTurn


15



Field / Value Pair


minimumTurn


10

Time


1:29:46

Group
Info


Car:Car1:suspensionRange


0.1

Group Info


Car:Car2:suspensionRange


0.2



Group Info


Car:Car1:suspensionRange


0.1

VGT Future Work


More middleware needed


Database
l
ogging


A server to take care of deeper analysis


Allows for mobile lightweight clients (tablets)


Formal evaluation of cost savings


Does VGT put more burden on QA?



Questions?

Thank You!

Background Work

Heuristics


Quantitative analysis of player actions


Statistical approach to game design


Gives a way to compare different approaches


Heuristic Evaluation for Playability (HEP)


Desurvire
, H.,
Caplan
, M., &
Toth
, J. A. (2004).

Using
Heuristics to Evaluate the Playability of Games.


Born out of the need for a standardized way to
evaluate player feedback


Focuses on relevant gaming concerns


Heuristics

Heuristic
Evaluation for Playability (HEP)

Heuristic and
Descripion

Game

Play


15

Pace the game to apply pressure but not
frustrate the player. Vary the difficulty
level so that the player has greater
challenge as they develop mastery. Easy
to learn, hard to master.

Game Story


5

The Player has a sense of control over
their character and is able to use tactics
and strategies.

Mechanics
-

1

Game should react in a consistent,
challenging, and exciting way to the
player’s actions (e.g., appropriate music
with th攠action).

Usability
-

1

P牯癩d攠imm敤iat攠f敥dback fo爠us敲e
actions.

Interactive Visualizations


Scattering Points in Parallel Coordinates


Yuan
,
Guo
, Xiao, Zhou, &
Qu
,
2009


Interactive Visualizations


Interactive
Dimension Reduction through User
-
Defined Combinations
of Quality
Metrics


Johansson & Johansson, 2009


Interactive Visualizations


Lark
: Using Meta
-
Visualizations for Coordinating
Collaboration


Tobiasz, 2010