This game looks awesome. I love the developer. I love the style. I love the game play. I love the pedagree. But my first thought when I saw it, what I was looking for as I watched that trailer, was “is it procedurally generated? No? Meh.” I have been ruined by the likes of Minecraft, Spelunky, Cloudberry Kingdom, Tiny Plumbers, Dwarf Fortress, The Binding of Issac, Diablo, and years of Roguelikes.
Randomized levels or procedural generation has been the holy grail of game development for years that recently has finally started to see some mainstream success. Maybe it’s just that we’ve played and seen and designed enough video games that we’re starting to generalize the patterns. Sure Nethack has been doing it for years and Diablo stole the formula and refined it, but it wasn’t until Spelunky and Minecraft that it was finally applied to something outside the 2D top-down hack and slash and we could see and analyze why procedural generation was so great and how to do it right. Not every attempt has been successful in my opinion, but I’ve valued every one.
To simplify the discussion I’ll break down what I see as the pros and cons of procedural generation:
- Pro - Proc-Gen allows for smaller game play experiences. Since it’s implied that you’ll be playing the game a couple of times at least there’s no reason you need epic sized maps or even save states. 30 minutes per run at best is totally acceptable.
- Con - You still have to develop enough content for a 120 hour game. In fact you might want to develop even more than that since you don’t want your players getting bored after they’ve beat the game a few times. If there’s nothing new to find there’s no reason to play more. That’s been my problem with Cloudberry kingdom, a good game that just doesn’t have enough variety to keep me going, unless you consider more things to dodge on screen variety. This is also the problem I fear Tiny Plumbers will fall into. But both these games are still in development so we’ll see.
- Con - All this necessary content can be overwhelming if you’re not willing to get into a game. Since every run is different that widget you saw on your last play through may not show up for another 3 or 4 runs making experimentation tough. This can be negated by proper design, but it can still be frustrating for people who don’t want to play with the game wiki open. This was the final death of Spelunky for me. When altars were added to the game I couldn’t figure out their nuances and got tired of trying.
- Pro - Proc-Gen legitimizes perma-death. It’s exciting when your losses have serious consequences. You feel every lost heath point and celebrate every health buff. A good run is exhilarating and a bad run doesn’t need to be conquered because the next game is a clean slate.
- Pro - This in turn legitimizes getting good at a game mechanic, instead of rote memorization. It encourages creativity. However…
- Con - Perma-death kinda sucks sometimes. Having a good run ruined just inches from a health-up isn’t a good feeling and without a do-over mechanic that may be the death that has you swear off the game… again. I’ve quit Minecraft a dozen times and Spelunky at least as many. But this has never been the nail in the coffin for me and I’ve always come back.
- Pro - Stories can be interspersed to great atmospheric as well as emergent effect. Dwarf Fortress is an excellent example of this that I wish someone else would employ because I just can’t get into that game.
- Con - Puzzles are impossible to procedurally generate, or rather the problem of proc-gen puzzles can not be generally solved and has not be adequately explored for many puzzle types. Sure we can generate a maze but has anyone made a procedurally generated sokoban knockoff and even if you could how do you balance difficulty? Consequently proc-gen games don’t yet get the broad yet interwoven puzzles of manually generated games. Tackling these problems is hard and often not as rewarding as just programming a few dozen puzzle rooms to be seasoned throughout.
- Con - Often times difficulty is uneven and it’s tough to balance it. A new player may be put off by a few bad runs that were simply not laid out well. If there was a way to insure a few easy runs to start with that would be good but most of the time the experience is going to be unpredictable by it’s nature.
- Con - Proc-Gen doesn’t lend itself well to the “explore everything” game play method. In Binding of Issac and Spelunky you often don’t have the keys, ropes, or bombs necessary to visit every corner and try out everything. The right way to play these games is to pick and choose what you can and want to pursue and sometimes, often times, just move on. This can be frustrating to people who don’t like to miss things.
- Pro - None of these cons are unsolvable… theoretically.
Procedural generation has tons of unexplored territory and I can’t wait to see where it goes.