Data Movies Visualizing Player behaviour to aid the design and balancing of F2P games Thomas Hulvershorn Operations and Analytics Manager Outplay Entertainment Amazon Appstore Developer Summit, London, 4 th October 2016
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Data MoviesVisualizing Player behaviour to aid the design and balancing of F2P games
Thomas HulvershornOperations and Analytics Manager Outplay Entertainment
Amazon Appstore Developer Summit, London, 4th October 2016
• Start!• Select Boosts• Play!
• Win • Play On
• Continue• Shop
Replay
Boosts
No Boosts No Boosts
Win and move on
Lost
New
Casual Games Core Loop
BoostsStart Play
Shop
End
Replay
Boosts
No Boosts No Boosts
Win and move on
Lost
New
Broken Loop
BoostsStart Play
Shop
End
• Frustrating, Too hard• Annoying, Impossible• Boring, Too easy
Game Design influences Performance
• Two Factors– FUN– Difficulty
Less Difficult
RetentionMonetisation
Game Design influences Performance
• Two Factors– FUN– Difficulty
More Difficult
RetentionMonetisation
Broken Level
Retention Monetisation
Game Design influences Performance
• Two Factors– FUN– Difficulty
Influencing Difficulty• Initial moves • Variations of Colours • Type and ratio of Blockers• Goals• Restrictive Tiles• Mechanics• Size and Shape of Board• …
Two Chained Loops
Game Design influences Performance
• Two Factors– FUN– Difficulty
Fun• Impact
– Greater Tolerance– Higher Monetisation– Better Retention
• How to design “fun”?– Quality in everything (Visuals, Animations, Audio, FPS, …)– Celebrate Achievements– Time to complete a level– Appropriate Difficulty
Game Design influences Performance
• Two Factors– FUN– Difficulty
Challenge• Balance each Level!
• Balance the Flow between Levels!
56 chained Loops
705 Levels in 38 Areas(09/2016)
Game Design influences Performance
• Two Factors– FUN– Difficulty
Game Analytics• Measuring
– Pass Rates (Outcomes)– Boost usage– Time to completion– Moves left, Star Achievements– …
• >1 Millions Data events per day per game (already sampled)
• Data stored in AWS, accessed with SQl, analysed with ‘R’
How to make sense of this?How to make the right decisions?
“Classic” Quantitative Analysis
Experiments
Game Design influences Performance
• Two Factors– FUN– Difficulty
Alternatives• Quantitative Analysis
–Answers Questions objectively• Qualitative Analysis
–Raises Questions subjectively
Lights Camera Action
Data Movie Director
Movie Selection and Setup
Detail View One Player
800 Players
Events per Level
Detail Analysis, one Event
Game Design influences Performance
• Two Factors– FUN– Difficulty
Data Model• Pacers
– Event Counter– Day– Session
• Context– Player ID– Level Number
• Content– Events Type, e.g. Loss, Won
Movie Configurator
Game
Events(Olga)
PlayerCity
ETL
Load
Mov
ie G
ener
ator
Cohorts, Selected Events, Time Frame
Generate
LevelDataDLC
RawEvents
API(Helios)
Sess
ions
/ IA
P
Anal
ytics
CommunicationLevel EditorWorkbench
Provide Insights
Data Movies
On Demand/
Batch
Integration
Game Design influences Performance
• Two Factors– FUN– Difficulty
Summary• Balancing one Level is hard• Balancing of a thousand Levels is harder
• Getting it right is the essence of sustained success for a F2P Game
• Qualitative Analysis can support the effort by raising questions and providing context and insights
“I know that eventually I can beat every level”
Related Documents
