Ohio, USAIf I had to guess the slowdown by looking at the footage alone, it'd be a combination of running DOSBox with default cycles (auto, or fixed 3000 for most games) and the recording software somehow mucking up the performance of the emulator (your computer in general seems to be fine since the timer never skips a beat). Seems like you've switched to higher cycles in later runs (I recommend "max 100%") but if the problem comes up again feel free to reach out and I can attempt to diagnose the issue.
In all my attempts running this game, that might be the worst Shikadi Master luck I've ever seen.
Throwing this here in case someone is interested in a category that doesn't yet exist. I'd rather not create a new category until there's actually a (planned) run for it, never liked the look of leaderboards with tons of empty areas. (As of this writing it's still pretty blank, but I'm aiming to fill it.)
I have some current thoughts for categories (both full-game and missions) that I may or may not get around to:
- Any% (Missions): Complete an entire mission, as fast as possible. I could've started with this, but All Challenges is a better demonstration of the game.
- Currently there are no missions set for "Toxic River" and "Infernal Gate", as there isn't a meaningful All Challenges for them, but naturally those can be added.
- Changes to what items are allowed. The most obvious is "No Items", where the player has nothing equipped. I can also see items split between those found in missions ("Search-only/less"), purchased in the Item Shop ("Silver-only/less"), or purchased at the Curator Animarium ("Soul-only/less").
- All Challenges (Full-game): Start a new-save file and complete the game, while also successfully completing an All Challenges run on each applicable mission.
- 100% (Full-game): Start a new save-file and keep going until it's a 100% file. Timing could have the player return to the main menu to confirm this, but (currently) having the Item Shop and Curator Animarium closed for business confirms this.
- Might wanna look into Low% (Full-Game) but I think this would force the player to deliberately lose challenges, which includes the time challenge, and therefore have a lot of wasted time in it.
If any of the above generate interest, or you have something else in mind, this is the place to ask!
That sounds like an excellent cleanup to me: no need to keep around blank categories, especially if they're obscuring the ones that people are more likely to pick up. Worst case scenario, somebody actually decides to perform an A-C run, at which point you can just put it back.
The other runs are on Normal, but that's because performing the run on Easy was considered more or less impossible until the run you see here. My take on an any%-styled run is that "any"-thing goes, aside from explicit cheats, and a change in difficulty fits that. I can add a difficulty variable to make it clear which difficulty is applied to which run, if you think that's important, but I don't think a category split is warranted seeing as how the two (or three, given that Hard now becomes an option) runs would be nearly identical.
I've altered the rules text a little over the years (mainly additions like no-sound timing), but I don't recall the rules at any point stating that the run must be on Normal. If anything, it would've said something like "Runs are played on Normal difficulty unless specified otherwise", as I never intended to disallow Easy difficulty.
I'm currently looking into this as a category to perform on individual levels. A rough estimate of how playing a run would look:
- Prepare Keen next to the level such that an action key will enter that level.
- Display the status window (verifies things like score before starting the run)
- Start and finish the level. Timing is between first and last visible frames of the level (i.e., as it's loading and unloading).
- After finishing, display the status window again (verifies difference in score to determine collectibles).
It's possible to create custom level saves, so the preparation plan in #1 is to build a set of saves with the same starting statuses for each level, and then make them available to download and easily switch out as needed. Currently I'm planning those saves to have zero score and zero ammo, as-is the default starting ammo in the game, though I need to look into whether the latter is reasonable for each level. This also makes it possible to build saves that both lack a pogo and are equipped with a pogo, increasing the potential types of runs.
Not sure if anyone here will see this in a timely manner, but I'm doing my best to make this information visible so that those interested can provide feedback during the early stages of this potential category. As for full-game runs involving max collection... let's see if the individual levels bring interest first, then we can look into how all levels might go.
In my experience running a bunch of different DOSBox games, I first decide whether the game requires a fixed or max cycle setting. A lot of older games execute as fast as a CPU lets them, sometimes leading to an unplayable mess of a program: fixed cycles solves this by suppressing the pseudo-CPU of DOSBox. On the other hand, some games require max cycles to play adequately at all: later games and a lot of FPSes fall under this category. While M&M3 seems to go as fast as possible without a fixed setting, it's also not a game that requires real-time reactions and can be played adequately under speed-maxed conditions, so it's not as cut-and-dry.
Given that we have my run (fixed 20000) and LiLbUg's run (probably fixed 3000 or 8000, would be good to get confirmation) to compare and contrast, the most notable difference between runs at differing cycles is load time and its effect on execution. A slower-cycled run is able to leverage many of the area loads by buffering inputs, but it's also a slower run in real time because of the loads themselves. A faster-cycled run, on the other hand, has far fewer opportunities to buffer, but is also faster from point to point as there is very little waiting around. I prefer the latter having watched both, though I'm obviously biased towards my own run and previous choices.
Ultimately my preference for this game is a faster run and therefore a faster cycles setting, but I'm equally inclined to hear from LiLbUg given their similar attachment. A quick check of M&M3 running at max cycles suggests it would be even faster than my fixed 20000, hence the "steady but brisk pace" rationale I apparently decided on at the time (i.e., fast but still has some pausing in which to breathe).
DOSBOX: Don't change certain settings to gain an unfair advantage, cycles, for example.
This is entirely fair, but what is the "unchanged" setting for cycles? The GOG version comes with a .conf of "cycles=fixed 8000", and I suspect that's a common use case: anyone owning an original copy, however, would likely install DOSBox manually and therefore have a default .conf of "cycles=auto" (which should result in a fixed 3000). So there's already a disconnect between naive runners.
The cycle setting (others can matter though very few ever do for timing/playability) is always going to be arbitrary, whether technical or aesthetic. Most games don't throttle their speed appropriately, M&M3 included, so it often comes down to what runners find necessary and/or sufficient. My own run from a few years back set cycles to fixed 20000: according to the description, "The cycles are to keep the game at a steady but brisk pace, allowing area transitions to occur quickly." (To be clear, I'm not here to argue whether my choice is better, and have no intentions of forcing a different-cycled run into the leaderboard.) By contrast, my Swords of Xeen run set max cycles because of the continuous saving/loading, which crashes way more often without it (though at least in this case, the GOG .conf also coincides here).
Regardless of what is chosen, I recommend clearly stating settings necessary for a consistent leaderboard. The GOG standard is reasonable here, but I've used DOSBox with many games and can tell you GOG doesn't always come up with settings I consider good for even a casual experience. Their settings are no less arbitrary than others that could make the speedrun more enjoyable to play.
I suppose there's nothing inherently wrong with this option. The manual itself does acknowledge this option (see PDF page 9 of this http://www.abandonia.com/files/extras/26916_Extra%20document_Manual.pdf) and, while it explicitly states "it is not possible to win the game" with this mode they probably weren't expecting available glitches.
I'm out of practice so I haven't yet confirmed that the game completes without saves when using this command. If it does...ultimately this comes down to a decision of preference, and I prefer to side with the convention of segmented runs from 10+ years ago that ran every level. Might make sense to rename the category to All Levels? Either way there will be an explicit mention of this particular command so as to avoid confusion.
Got the whole Commander Keen series of games up and running. Thanks for granting me mod on the games!
