Personal Post Mortems are brief self evaluations of projects I’ve worked on. They’re focused on my personal design goals, design successes/failures, and my takeaways.

 

My Role – Lead Design

B10_2_a

  • Designed character attacks, AI, bosses, levels, upgrades, and progression
  • Directed character animators and battle scripters
  • Implement character attack data, level geometry and objects
  • Develop, and communicate, and maintain project vision

Personal Design Goals

  • Revamp mechanics that had poor feedback
    • Hidden dart trap became a laser trap with more visual feedback
    • Smart bombs now give the player time to escape before exploding
    • Activation platform art much clearer and on brand
    • Upgrade menu better communicated how to unlock each upgrade
  • New approach to AI design
    • Maximize variety in AI mobility, mechanics, and combat role
  • Ensure upgrades never make other mechanics less interesting
  • Mercilessly cut features not in line with the core gameplay
  • Maintain goals from Omniverse 1

What went well

B10_2_c

  • New AI added much more gameplay variety
  • Laser traps, new smart bombs, and activation platforms are easily understood and less frustrating than prior mechanics
  • 4 added playable characters were well received by playtesters
  • Enemy special attack types better communicated and effects much more noticeable
  • Boss fights were much easier to understand and generally more exciting
  • New level and attack implementation tools saved a ton of time
  • Only 1 build to worry about, allowing for more focused polishing
  • Companion AI that set up enemies for the player to strike down

What didn’t go well

  • Blinding effect wasn’t as fun or engaging as intended
  • Experimenting with new ways to communicate attack design to animators
    • Too few new playable characters to see a return on investment for this project
    • Not enough time spent in each pipeline to learn if it can be useful
  • No time for significant updates to original character set
  • All bosses required much more work than expected
  • Some AI and hazard combinations are unfairly difficult

Lessons Learned

  • Blinding effects are difficult to get right, possibly not suited for 2D platformersB10_2_b
  • Filling all intended combat roles in Omniverse 1 made finding new roles difficult in the sequel
  • Tools programmers speed up pipelines, which decreases iteration time, which improves design
  • When testing a new pipeline, make sure the targeted tasks will provide an adequate sample size