Chapter 1 Intramortem

A few days before returning to work on Death World after a couple weeks off, I told myself I would revamp the battle system - it wasn't fun to work with as it was, and that was affecting my motivation. I would get to work as soon as I released the first chapter with the current battle setup in place, because I had worked hard and could use a milestone. Better to publish and get feedback before diving down a new rabbit hole. In the course of preparing to publish, I replayed the chapter several times, found that the battles were enjoyable after all, and decided not to replace it after all. But by the time I published I had flip-flopped again, and my mind is still thrumming with ideas for new battle mechanics and dynamics.

With the demo released, my feeling is that the design of Death World is misaligned with its goals. I've had a specific goal in mind with Death World: to demonstrate the power of consistent patterns in enemy health and player damage to simplify the process of "learning" new enemies. That was one of the achievements of Lufia and the Fortress of Doom, and that game has been my model in pursuing this goal. The most noticeable example is that Fortress of Doom gave enemies low health relative to player damage, and Death Worldreplicates this math; it's easier to notice patterns in enemy durability when they die in 1-2 hits than in 2-3 (or more). However, I have relied on Lufia's precedent perhaps too much and emulated its design in ways I shouldn't have, mimicking design choices that were valuable in FoD's context, but possibly distracting or even destructive in Death Worlds. The most significant of these unwise imitations I've identified so far is in the pacing of battles.

Lufia and the Fortress of Doom had fast battles; it was rare for a random encounter to last more than two rounds unless something went wrong, like multiple party members becoming confused. In FoD, I could parse a battle scenario and intuitively devise a prudent distribution of attacks very quickly, and with a little luck, my plan would play out as intended, and the battle would be over with minimal waste. The speed of battle is partly attributable to enemies' relatively low health, but it's also a function of strategic play. Fortress of Doom's combat is centered on its targeting system, which has a few quirks, but none more notorious than the fact that attacks are not retargeted if their intended target dies before the attack executes. Instead, the superfluous action is wasted, and you have to watch your character strike at the empty space where the enemy used to be. Good play in Fortress of Doom is about distributing attacks carefully in order to avoid wasting turns in this fashion; if you know the enemies well, you can make wise allocations and end battles faster and more efficiently. If the turn order system is on your side, you might even wipe out the enemy party before it has a chance to act.

Death World's mistake is in overstating the importance of fast battles and winning before enemies have a chance to move. In general, all the characters on the player's side have higher speed than all the enemies, and most battles are specifically designed to be winnable without taking damage. This is tricky to reckon with. In Fortress of Doom, good play masks difficulty. An efficient distribution of attacks in a turn requires you knowing how much health each enemy has, how much each attack will hurt them, and what order your characters will act in. This information is never made explicitly available. Instead, your ability to estimate it through intuition is developed over the course of the whole game. The first chapter of Death World is meant to build the player's intuitions about enemy health and speed in a similar way; if you internalize them, it's often possible to win battles before any enemy can act. One playtester remarked that the first chapter was easy to the point of being tedious. Does this mean that the playtester successfully internalized the lessons of the game as intended, obviating all challenge? Or does it mean that there wasn't any challenge, and nothing had to be learned at all? My present impression is that both cases have a similar solution: the battles need to push back a little bit harder. Good play can still look effortless, but I think it needs to be the result of making more, or more interesting, decisions; making more, or more meaningful, use of accumulated knowledge. The questions before me now are what kind of decisions and knowledge those should be, and whether it's necessary to change Chapter 1 to include more of them, or if they should only come into focus in Chapter 2.

So do I revamp the battle system? Daydreaming about systems is the most seductive part of RPG design, and I'm always tempted to dive into building something new. But making this first chapter has taught me that designing encounters for a given system is slow, difficult work, and as much as I thought the lessons I learned from studying Fortress of Doom would prepare me for it, the two dozen or so battles of Chapter 1 were kind were a big labor to develop. Adopting a new battle system would mean returning to square on a project that's already taken far longer than anticipated. At this juncture I think it might be best to proceed with the current battle paradigm, making minor adjustments as needed, but refraining from adding any new gimmicks. This may not save me from having to redesign all my encounters, but it will least save me the overhead of designing and testing all new skills, stat distributions, and character relationships. Moreover, there's something to gain in my own creative practice by refusing to endlessly redesign functional content in pursuit of perfection. I recognize in myself a need to practice accepting my imperfect decisions and committing to them, making the most of them, and taking the lessons learned from them on to new projects. So I think I'll hold off on any major redesigns for now. Instead, I'm going to focus on making sure that there are meaningful decisions to be made for every character each turn. As RPG Maker 2000's flexible targeting system means the decision of who to target with an attack will never carry as much weight as in Fortress of Doom, my first task in the next phase of Death World will be to find a different way to supply that weight in a way that supports the learnability experiment.

One last feeling: I lament that I haven't been able to watch anyone play the demo live yet. There are some things I have no clue how much I should worry about. Do players realize that Japheth's attacks consistently one-hit kill and distribute their attacks to reflect this, or do I need to make the benefit more obvious? Should I bring her speed closer to Auredo's so they occasionally act in reverse order? Actually, this is the only thing I have no clue how much to worry about, but the fact that I only have one such question suggests many more blind spots, further highlighting the need for conscientious playtesting. I always hear people stress the importance of it, but sometimes wisdom doesn't sink in until I experience firsthand the problem it's meant to solve. Let this be a reminder for me to prioritize playtesting, including actually watching people play the game, for the full release.