A couple of months ago, I wrote about revising the core hacking gameplay of Darknet. The puzzle was too chaotic, and there wasn’t enough variety, so I made it turn-based and added special types of enemies. Both of those changes were good, but they weren’t enough, unfortunately.
The basic idea of the puzzle was this: The player had to spread “hacks” into the center of the puzzle, while security programs (guided by simple, deterministic AI) would try to destroy the hacks. The player could only inject hacks by targeting and converting one of the security programs, at which point the hacks would spread outward constantly. Here's what it looked like originally.
Even after my revisions, chaos was a problem. The turn-based interaction made it a lot more manageable, but in bigger puzzles, it was still impossible to predict how all of the enemies’ AI would react. Usually, all of the security would be pushed outward away from the goal, but that would leave the player with no good targets to hack. Sometimes, a lone security program would get stuck near the goal, providing the perfect target. Other times, a whole group of security programs would end up near the goal, making the puzzle impossible to solve. Because their combined behavior was so chaotic, every turn felt like a roll of the dice.
I needed some way to make the puzzle mechanics more predictable. Sometimes predictability is a bad thing, but in this case, there were so many moving parts involved that complexity would emerge even from very simple elements. The old security program AI was deterministic, but it was still too complex to predict except in rare, trivial situations. If players couldn’t predict the outcome of a choice, they couldn’t form theories about how to make progress, and any feeling of strategy would break down.
I spent a long time on this problem before I finally had a breakthrough. One of the special enemy types I had added was a stationary mine. If your hack spread to it, it would simply blow up, destroying itself and all your hacks within a certain radius.
Very simple. Very predictable!
So… what if there were only mines?
Your hacks are constantly multiplying, so that wouldn’t work if the mines could only affect a small radius. I made one change to accommodate this: Instead of having an instantaneous effect within a radius, the mines would spread a sort of counter-hack outward, 3-4 times faster than the hacks were spreading. When your hacks hit a mine, they’d very quickly be consumed by the security. Here's what it looks like now.
I didn’t realize immediately how well this system would work. It probably doesn’t sound very interesting at first, either! But in testing, I immediately encountered some interesting decisions. Right away, it felt strategic in a way that the previous system never quite did. Best of all, it still fits perfectly into the rest of Darknet’s existing mechanics; the new puzzles work with the old procedural-generation algorithms, and they can produce the wide range of difficulty that the game requires.
It’s pretty crazy to be revising the game’s core mechanic so much, especially when it's this late in development, but I feel very good about this change. However, it brings back another issue. In my last round of revision, I tried to solve the chaos problem and the variety problem. I may have solved the former, but reintroduced the latter: I’ve left myself with a stripped-down, minimalistic system yet again.
It’s not clear how quickly the new mechanic will get old. Hopefully there isn’t a variety problem at all, and if necessary, I’ve got some ideas for how to add some spice to the new puzzle mechanic. But the game’s development schedule is booked, and I’m honestly not sure where I’ll find the time.