This is the final part of a dev diary for “Super Cult Tycoon 2 : Deluxe Edition” a game I’m making in a month with Part 2 for more info!
It’s the final stretch, we have to make this mess of loose mechanics into something worth your time. Daunting! We had until Sept 30th to finish. But will we?
DAY 9) I ended up hacking in a pathfinding system that will need to be extended later, but at least we now get a better idea of how to move the abduction van around, so improvements can be made. Essentially, I restricted the van’s movement to go either to the cult drop-off location, or anywhere within the town’s bounding box. Between the two is a single waypointed path. Serves our purposes for now, but will need to bite the bullet and put in full pathfinding soon.
DAY 10) Put in better FBI spawning. It now spawns humvees that travel along one of several paths to the plantation, and then spawn FBI agents. To make the presence of agents a bigger issue to worry about, there is now an ‘Investigation’ progress bar that advances when FBI agents find their way to your cults HQ (a barn). If the FBI completes their investigation, you lose. This acts as a natural game timer, but is far more dynamic and, if we can balance it right, leads to scrambles in the end game as you become overwhelmed by agents.
DAY 11) Robert improved the GUI, added a launch screen, and put in a swish notification/tutorial system. Along with some quick and dirty ‘focus on event’ code, it really helps bring mechanics together. We had to pick a game title for the logo, and ended up sticking with our working title plus a few bells a whistles, to get ‘Super Cult Tycoon 2 : Deluxe Edition.’
DAY 12) We still needed the abduction mechanic to be more active, and settled on the idea of FBI road checks, which hold up your van and increase FBI presence. Since you now needed to be able to choose different routes back to base, I had no choice but to put in a pathfinding system. But first, procrastination. I cleaned up bugs I’d been putting off for a while (like followers that would slowly rise up as they wandered around. Believers may call it ascension, I call it sloppy) and put in a couple minor aesthetic features.
DAY 13) I decided to use the pathnodes + A* implementation I had already written for Muckraker. The problem we had was that our forest has nothing to block line of sight, so my old LOS based method of finding node connections would create paths through trees. We didn’t want to manually place LOS blockers, and really didn’t want to manually draw paths. After talking about a whole bunch of complex ideas (like sampling the texture/detail map to avoid pathing through forest, or raising the height of the terrain in forests by a tiny amount) we realised we were overthinking it. I put in a simple distance check, so nodes that are close to each other will always be connected. We have no scenario where nearby nodes wouldn’t be connected, and no changes in height, so it fits our needs perfectly. After doing this, it becomes trivial to find a path from any valid location to any other valid location through our good friend A* (I may post my general & super-simple implementation on my blog for anyone interested)
DAY 14 ) Pathfinding implemented, spent a while ironing out some bugs. Also had to put in workarounds for a couple of cases. Each time you try to place a building it will check whether there is still a valid path to HQ, and if not you won’t be able to build. It will also check if any NPCs paths will be interrupted, and if so, repath. All major coding is now complete, and although we probably won’t be ‘feature-complete’ until release, we have the basic game loop down pat and will mostly be refining existing features to focus on the game’s strengths.
DAY 15) Playtesting, and then more playtesting. Honestly, I’ve played this bloody thing so many times I’m starting to worry it’s affecting my sanity (it’s the ice cream truck music’s fault) But each time we play, we see something that should be changed. A bug perhaps, a GUI niggle or very occasionally, a peek at the ‘core’ of the game and perhaps a way to expose it. We got rid of the multiple vans, instead you upgrade the one vehicle. Most buildings no longer need to be staffed. Put outline shaders on the van and agents, so they can be seen behind buildings…
DAY 16) CRUNCH MODE. Reality struck. There is no way we can get this polished in time. We decided to finish an ‘alpha’ build for the pageant on the 30th. This alpha is really just a demo of the systems, with gameplay not implemented fully. Importantly, it forces us to hammer down showstopping bugs that we had been putting up with. Robert added a wonderful intro/tutorial, as well as an ending scene. I added some feedback where needed(such as the ‘money rain’ when bribing) and cleaned up features. We vowed to have it finished by Sunday, but at 5am Monday (the 3rd) decided to post an updated build and wait a few more days weeks, since releasing a game that crashes half the time isn’t so professional.
DAY 17 – ???) Just polishing now, the most frustrating stage since it seems that all you’re doing is fixing your mistakes. I spent a long time fixing a series of related crashes, as well as writing the end so it can land you back at the front menu (to play again perhaps???) Tweaking balance as I go. This game is small enough that we don’t need any complex balancing rules or spreadsheets, and lets us experiment with changes rapidly. Really want to get this out soon, since a couple sites (Kotaku, PC Gamer) had already posted about our unfinished/crashy alpha build, which doesn’t make such a good impression. But at last…
Altercation. (Webplayer, Win, Intel Mac)
Feel free to ask any questions about the development, or point out my mistakes. We used Unity3D, Photoshop, Audacity and Maya.
NOTE : This is reposted at my personal blog, grapefruitgames.com