Wednesday, July 22, 2009

dynamic crime investigation

I wonder it's possible to randomize crimes in an open world for a crime investigation game, in the vein of the NCIS or CSI TV shows.

By open world, I don't mean a simulation where characters in the world are all living dynamic lives and crime springs from those activities. I mean a gameworld akin to those of GTA or Saints Row, with the game dynamically producing one crime in that city for the player to solve or close the book on, then producing another when the player is done.

For each crime, the player would be called to a specific location to find and analyze physical evidence like bodies, fingerprints, DNA, murder weapons, etc. Analyzing some evidence might prompt the player to go to another location to search and analyze. Eventually, the player would find the suspect and must chase down that suspect through buildings or the city streets. If the player loses the suspect, then a new chase is arranged. If, at any time, the player feels stuck or doesn't like the crime, he can close the case and move on. Ratio of successful vs unsuccessful cases would be among statistics shared online.

There are two reasons I'm not sure this sort of game is possible.

First, you'll notice I didn't mention any dialog.

If crimes are randomly or procedurally generated and not scripted, then it's unlikely questioning, interrogation, and such could be made fresh and interesting for each case. Investigators and civilians could blurt out humorous lines. But conversations can't be dynamically produced to mimic the lies, dodgy answers, and withheld information of conversations with witnesses and suspects.

Dynamic witnesses are possible if they simply blurt out basic details at the crime scene, though. They could even bear witness to verbal confrontations and scenarios in general ways, like "They were yelling at each other" or "They looked like they were having a romantic dinner". Anyway, could a crime investigation be fun without deep dialog if the crimes were countless and unpredictable in an open world?

Second, how can crimes be procedurally generated and feel both deep and fresh?

I figure this would be easier if the crimes actually happened. So the game invents a scenario, then plays out the scenario so that witnesses are possible and the placement of physical evidence is truly dynamic. The game invents these scenarios by creating relationships between random NPCs (family, business partners, crime partners, drugs, testimony, etc), searching its index for crimes possibly resulting from such connections, and then playing it out in a dynamic setting.

Honestly, I believe much of this is possible, but perhaps not enough to make a full and consistently fun game. What do you think? Is the crime investigation genre of games capable of procedurally generated or emergent content?


  1. How about a Spore type thing where you can make NPC's or Mobs and they are uploaded to a huge database that is pulled from when you zone in somewhere?

  2. It could be designed like Spore so players design the content and that content is spread throughout a single-player world. Players make the characters, as you say, and script the crimes; then other players try to solve those crimes. That certainly solves the dialog issue and challenge of providing countless missions. The main hurdle would be ensuring player-created missions are solvable (the game could run a test before allowing a mission to be posted online).

    And that also creates potential for co-op crime solving and even competitive investigation organizations. Good idea, Hudson.

  3. My question would be... why do they have to be solvable? Personally, I wouldn't want to play any CSI type game unless there was the possibility of unsolvable crimes and/or serial killers (i.e. - the first crime will be "unsolvable" because you simply can't get enough evidence to charge anyone, then later you get a similar case that allows you to get more evidence, and so on, until you manage to lock down this killer after his fifth victim but before he can claim his sixth).

  4. I recently came accross your blog and have been reading along. I thought I would leave my first comment. I dont know what to say except that I have enjoyed reading. Nice blog. I will keep visiting this blog very often.


  5. I'd like to see a well made game based around investigation but I have not gone anywhere near the CSI released games. The ones I have played that had promise were the Condemned titles and had hoped that Heavy Rain would offer something new.

    I'm not convinced that a 'investigation game' would draw me in enough without great dialogue/story. Loose ends, like unsolved cases that could be reopened at a later date, make this genre.

  6. Thanks, Dayana.

    Jason, I agree that serial killers and the chance to get "revenge" on a nemesis would be good additions. Just as Spore allows players to filter the player-made content they download, perhaps players could choose to allow or search out different types of investigation missions.... including crime chains. A simple checkbox could tailor the game to include unsolvable crimes or not.

    David, witty dialog is definitely a vital component of all investigation TV series. And, as I've said, that's the one gameplay element that can't be generated by a computer. Unfortunately, I expect a player-content driven game would have dialog that varied greatly in quality. But the bright side is that having so many writers with so many different styles and tastes could definitely keep things fresh and interesting.

  7. This comment has been removed by the author.


Note: Only a member of this blog may post a comment.