Version differences

Version differences are a frustrating part of speedrunning. I've found them demotivating for even the best of speedgames. This post won't present any concrete solutions, but I hope it can shed some light on the problem and encourage healthier approaches.

In some cases, different version means different game. Two examples that come to mind are Pokemon Red and Super Mario Sunshine. The Japanese versions of these games have noticeable differences (text and in-game trades in Red, text and level design in SMS, to name a few). These differences affect both normal gameplay and speedrunning. Everyone has to learn part of a different language to understand what is going on, and speedrunners have to find different strats to account for the text, harder/easier levels, etc. Keeping these games on the same leaderboard makes no sense unless competition is scarce. Red has always separated English Red from Japanese Red, but most games, including SMS, allow all language versions, and most games run faster in a language other than English.

What about games where the only difference is the platform? That's trickier. The game is unchanged, but certain consoles might execute it faster than others. If this difference is only a few seconds, it might seem silly to ban the faster version just because it's less accessible. But that sets a dangerous precedent. What if a game becomes so optimized that a few seconds really matter? What if a new trick is found that makes an obscure console significantly faster, to the point where it's impossible to compete for WR without it? What if the fastest console is a modded one?

I often hear the argument that "you should play for fun, not WRs." This is a healthy mentality, but competition is also healthy. Competition drives speedrunning communities to find new optimizations, and speedrunners to improve their execution. Speedrunning leaderboards exist to promote competition, not just to display the best runs. Being forced to compete at a disadvantage, even a small one, is a huge deterrent.

Sometimes the problem isn't just competition; it's the nature of the optimization itself. At its best, speedrunning is about using your brain to solve complex problems. What's the fastest path from X to Y; how much risk is involved; how will taking this path affect other paths; how does a character's health and money route affect the run? But when multiple versions are in play, the important questions are often simplified (how much time does console X take to start up), duplicated (is the fastest path different on version X because of lag differences) or extremely tedious to solve (can we manipulate the RNG in a faster way on console X despite having no working emulator for it). If the end goal is just to go as fast as possible, it's easy to get caught up in these questions and spend less time optimizing the game itself.

Take Pokemon Red as an example. You can RNG manipulate many things in this game to save time. Manip works on all versions, but because the RNG differs from console to console, you can't apply manips across the board. Many versions require their own manip routing. Manip requires saving and resetting. Resetting is fastest on emulator, and we have almost perfect emulation at our disposal. Resetting is nearly as fast on several consoles, some of which play the game at an incorrect frame rate, some of which have no capture equipment (so potato cam or bust), and one of which is a modded Wii. Meanwhile, the community standard, the Game Boy Player, loses about 6 seconds to slow resets. It could lose more time if a new manip-heavy route comes to fruition. But that route is nearly impossible to test without emulation, and we don't have accurate emulation for GBP resets. So which route we use, and the extent to which route testing is even worth our time, depends on which platforms we allow.

What is the solution to this? In a perfect world, one of two things would happen. One, speedrunning would be popular enough that every version/console would have its own competitive leaderboard. Two, perfect emulation would exist for all consoles, and everyone would have the money to afford these consoles. Nice memes.

Realistically, difficult decisions have to be made. Do you single out the most popular version of a game and relegate others to side categories? Is "popular" even the qualifier we're looking for? Do we allow all versions and accept the consequences detailed above? Do we add time penalties for running on faster versions (e.g., 2 seconds for each hard reset on emulator)? All of these things have been considered, and as far as I know, no one has found a one-size-fits-all solution.

Ultimately, my concerns can be summed up by comparing two questions:
1. What is the fastest way to beat this game?
2. What hardware runs this game the fastest?

Which question should a speedrunner seek to answer?