Face Off! A Face-Reading Criminal Interrogation Game


Independent Projects, Collaboration, Writing, Game Design, Storytelling


Clmtracker, JavaScript, Jade, Illustrator, Photoshop



Link to In-depth official page


Created in 48 hours for Global Game Jam 2015. You’ve just robbed a bank, and you’re in the hot seat. You must use your facial expressions to make the cop believe your far-fetched story. The game uses facial recognition software to register and display key features of your expression, and compares them to four set emotions: sadness, happiness, anger, and surprise. The player must accurately replicate one or more emotions per round in order to convince their interrogator that they are innocent. Too many failures, and the player loses the game.


Create a game in 48 hours with the thematic prompt of: what do I do now?


Beginning with concept ideation and voting, we moved onto writing a script and then collaboratively coding the game itself. We began playtesting and making system design changes (scoring, win/lose states, difficulty) by day 2. 


(+) The concept ideation phase got everyone's ideas on the table fast and was where our process was strongest, because we moved fast and built a lot of consensus and clarity about what we were making.

(+) The most important systems design moment was getting the chance to test the threshold values for what faces "passed" and "failed" the interrogation test, based on accuracy. We got the chance to playtest these enough to get our desired 60% success rate internally, and with people who'd played the game before, but not first-timers.

(-) We didn't iterate enough on the UI, though we did some. People didn't have an innate sense of feedback until they understood the controls. We suspect it was too many unlabeled meters to watch while also watching your face.

(-) We needed to playtest this earlier with an outside audience--possibly with a paper mockup of the UI.

(-) In terms of process, what would have been most helpful that we didn't do was definitely Usability Testing and Interviewing (seeing if at a base level newbie players could get the game to work and understood the feedback). We only playtested, to our detriment.

Role: Game designer, writer, programmer, artist



Face recognition and tracking "clmtracker" javascript library is available at https://github.com/auduno/clmtrackr courtesy of Audun Mathias Oygard.

Caroline Record - programming, product design

Sam Riordan - art, voiceover, audio editing

Abhinav Chandran - story, product design

Team Size: 4


Project Length: 48 hours