SUMMARY
SUMMARY
SUMMARY
SCOPE
SCOPE
SCOPE
Designed the AI Playlist as a “safe practice ladder”: match structure, difficulty tiers, and progression intent.
Designed the AI Playlist as a “safe practice ladder”: match structure, difficulty tiers, and progression intent.
Designed the AI Playlist as a “safe practice ladder”: match structure, difficulty tiers, and progression intent.
PROBLEM
PROBLEM
PROBLEM
Lack of learning and players losing too many times made them feel bad and want to leave.
Lack of learning and players losing too many times made them feel bad and want to leave.
Lack of learning and players losing too many times made them feel bad and want to leave.
DESCRIPTION
DESCRIPTION
DESCRIPTION
Creating a playlist to accompany players in their player progression to practice at multiple levels and reducing the complexity of playing online by creating simulator matches with AI.
Creating a playlist to accompany players in their player progression to practice at multiple levels and reducing the complexity of playing online by creating simulator matches with AI.
Creating a playlist to accompany players in their player progression to practice at multiple levels and reducing the complexity of playing online by creating simulator matches with AI.
BACKGROUND
BACKGROUND
BACKGROUND
The competitive side is tough on new players, and there was no in between that and the Tutorials.
The competitive side is tough on new players, and there was no in between that and the Tutorials.
The competitive side is tough on new players, and there was no in between that and the Tutorials.
GOAL
GOAL
GOAL
Defined a bot difficulty model (human-like error / competence scaling).
Defined a bot difficulty model (human-like error / competence scaling).
Defined a bot difficulty model (human-like error / competence scaling).
SYSTEMS
SYSTEMS
SYSTEMS
UE tools, tuning tables, AI inaccuracy maps, settings tree, Figma/Miro for system mapping, Jira
UE tools, tuning tables, AI inaccuracy maps, settings tree, Figma/Miro for system mapping, Jira
UE tools, tuning tables, AI inaccuracy maps, settings tree, Figma/Miro for system mapping, Jira
OVERVIEW
OVERVIEW
OVERVIEW
KEY DESIGN DECISIONS
What “safe” means (reduced social pressure, controlled challenge, clear learning goal)
Difficulty scaling method (error rates, accuracy bands, recoil spread)
Team composition rules (bots as teammates vs humans; why it matters)
Matchmaking constraints and how they shaped design
KEY DESIGN DECISIONS
What “safe” means (reduced social pressure, controlled challenge, clear learning goal)
Difficulty scaling method (error rates, accuracy bands, recoil spread)
Team composition rules (bots as teammates vs humans; why it matters)
Matchmaking constraints and how they shaped design
KEY DESIGN DECISIONS
What “safe” means (reduced social pressure, controlled challenge, clear learning goal)
Difficulty scaling method (error rates, accuracy bands, recoil spread)
Team composition rules (bots as teammates vs humans; why it matters)
Matchmaking constraints and how they shaped design
INSIGHTS
INSIGHTS
INSIGHTS
This was a first version of our AI Playlist, we gathered a lot of data. We saw it didn’t perform very well because of the long queues in matchmaking and the lack of bots as teammates.
However, this first step was key to our future update (coming soon), because now we understand players expectations of what defines an AI, thanks to UXR and Data, in terms of human error and machine accuracy.
What’s more we defined clear next steps on how to grow it into a playlist that can adapt to multiple types of player progressions and safer zones for our newcomers.
This was a first version of our AI Playlist, we gathered a lot of data. We saw it didn’t perform very well because of the long queues in matchmaking and the lack of bots as teammates.
However, this first step was key to our future update (coming soon), because now we understand players expectations of what defines an AI, thanks to UXR and Data, in terms of human error and machine accuracy.
What’s more we defined clear next steps on how to grow it into a playlist that can adapt to multiple types of player progressions and safer zones for our newcomers.
This was a first version of our AI Playlist, we gathered a lot of data. We saw it didn’t perform very well because of the long queues in matchmaking and the lack of bots as teammates.
However, this first step was key to our future update (coming soon), because now we understand players expectations of what defines an AI, thanks to UXR and Data, in terms of human error and machine accuracy.
What’s more we defined clear next steps on how to grow it into a playlist that can adapt to multiple types of player progressions and safer zones for our newcomers.