Conversely, my intuition tells me that if we remove the limitation of n being rational, in other words, we accept any real value of n > 1, then you can choose an n so that 21/n is rational.
However, I can't come up with an example of this. Either there is a really trivial example that just doesn't come to mind at this moment, or it's a bit difficult.
One thing I like about this TAS is that it manages to completely hide the fact that the control system of the game is absolutely horrendous. It almost makes it look like an actually playable game, of the likes of Super Mario 64.
21/n is irrational for any integer value of n > 1. I wonder if that's true also for any rational value of n > 1.
(I can't intuitively think of a counter-example at this moment. But of course intuition is no proof.)
Well, here's a topic I was thinking about some time ago, but didn't get around to write about: Family names.
Perhaps the only "family name" that there exists in the show is "Pie", with the entire family having that "surname" (except for Cloudy Quartz, who either didn't take her husband's "surname", or it's simply not customary in the MLP universe. My guess is that it's the latter.)
Then there are of course the "Apples", but that's not really a "surname" per se, as none of them have it. The (rather extensive) family is simply known as "the Apples". The name "Apple" does appear in the names of numerous members of the family, but not really as a "surname", but as a "first name" (or sorts). The only exception I could find is Caramel Apple.
(It gets a bit confusing with Apple Pie (or Perfect Pie), who's a member of the Apple family rather than the Pie family. The "Pie" part is probably just unrelated coincidence.)
Then there's Mr. and Mrs. Cake (Carrot Cake and Cup Cake). Rather than being eg. cousins, I'm assuming one took the "surname" of the other. (Since MLP is a heavily matriarchal society, I'm betting it was Mr. who took his wife's "surname". Could be the other way around, though.) Their offspring are also Cakes.
But other than them, I can't remember any more "family names" in the entire series. Unless you count "Twilight" as a "family name" (given that Twilight's mother is also named Twilight. But this could simply be a case of a child being given the same name as a parent, which isn't all that uncommon.) Many ponies also have single-word names, pretty much excluding the possibility of a family name or "surname".
There is a pattern I'm noticing here: Earth ponies seem to tend to have family names, while unicorns and pegasii don't. Maybe it's a thing.
Auto-hold is a good, universal solution in emulators. However, it may also be possible to remap your keys to support pressing more of them at the same time than your current key mapping allows.
It's not as simple as "this keyboard only supports at most three simultaneous buttons". It actually depends on which keys are being pressed. Certain combinations of keys will only be supported up to three or four simultaneous, while other combinations can go much higher, up to ten or even more. It depends on which keys we are talking about. (This is due to how keyboards are designed. They cut cost by designing them so that not all keys are supported simultaneously, but the way it works is that they are kind of in groups, and keys from different groups can be pressed simultaneously while keys within the same group cannot, or something like that.)
You could try to find out which combinations of keys are supported, and consider remapping your emulator keys to them. (Typically keyboards put an emphasis on the WASD keys and its surrounding area, because it's the most used in games. Keys in that area are more likely to support more simultaneous key presses than elsewhere. Of course it depends a lot on the keyboard.)
I think Nach makes a good point in that chaining consoles like this is too finicky and prone to something going wrong, with too little payoff to be worth it. Even if just one of the consoles in the chain does not sync properly, the entire thing will fail. If you have, for example 5 consoled chained like this, and it takes something like three minutes for eventually console #4 to be taken over, and console #5 fails to sync properly, what will you do? Start over, or end the demonstration there as only partially successful? What if it's console #3 or #2 that fails?
Something that Warepire wrote gave me a different idea, involving two consoles. I don't know if it's technically possible, but maybe: What if one console would be taken over by tasbot, and made into a tasbot itself, and would then proceed to TAS another console as normal?
In other words, for example, tasbot takes total control of a SNES console, which is connected to a NES console, and proceed to TAS a game as normal on the NES. Tasbot itself could be disconnected (if possible) on camera to demonstrate that it is indeed the SNES which is TASing the NES game.
This could be made so that even if the NES fails to sync at first, the "tasbot" program in the SNES is constantly running, so the NES can simply be restarted for a new attempt.
Obviously this should be clearly explained by the presenters and, if possible, shown on camera to the livestream.
As said, I have no idea if this is technically and physically possible, but could be an idea.
There is a lesser-known checking/repair program in newer Windows's named sfc (system file checker), which you can try to run (by running the command "sfc /scannow"). It might or might not fix your problem, but it doesn't hurt to try.
More information here: https://support.microsoft.com/en-us/kb/929833
The important part is that the favorable RNG could have theoretically happened in a single-segment run, if the player had happened to do the exact right things (eg. timings or whatever affects the RNG.)
The question was about whether segmented runs are ok, where the segment with the best RNG is chosen. Not about runs that use resetting.
In principle the saving/loading is not part of the run. It's done simply to start the segment over.
If I understood incorrectly how the RNG is optimized in M&M, then I'll change my answer.
I don't know the game, but the description looked like the bug is triggered purely with gameplay controls, so I don't see a problem.
Setting up a party seems to be an essential part of gameplay, so I don't see what the problem could be.
What is the purpose of segmented speedruns? To minimize suboptimalities by reducing runner mistakes as much as possible, and this is done by doing smaller segments of the game at a time, and choosing the best results, then appending them together as if it were a whole run. It approaches the same idea as with TASes: What would it look like if a skilled player played the game without mistakes, almost perfectly? Segmented runs could ostensibly have been done single-segment, if the player made absolutely no mistakes, and got really lucky in everything.
And getting lucky is part of it. Thus, I think, restarting a segment in order to get better RNG feels like a rather legit technique, especially if the RNG is affected by the gameplay (rather than, for instance, how many times the savefile has been loaded). It's not doing anything more than seeking that perfect playthrough, with everything going perfectly (or almost perfectly). Something that could theoretically happen in a single-segment run, if everything just goes perfectly.
I suppose I don't understand what you are asking.
I don't think it makes a difference how far in the game the glitch jumps, with regards to the question of whether I like it or not.
What is the meaning of the Star tier?
For very long it has (perhaps only de facto) had the meaning of "a small selection of runs that demonstrate particularly well what TASing is all about", or in other words, "if you are looking for cool TASes, check these first". In other words, cool TASes for newcomers.
If it does not mean that, then what does it mean?
But then we go back to the question of why eg. the developer console in HL2 is disallowed, even though it's fully within the game, usable from the game, using its own input.
What is and isn't allowed is, ultimately, somewhat arbitrary. The community decides "thing X supported by the game itself is allowed, while thing Y supported by the game itself is not."
The speedrunning community draws the line at the developer console. I would prefer the line to be drawn a bit more restrictively. My rationale is that a speedrun should complete the game by playing the game rather than using non-gameplay means (such as the developer console, deleting saves, and saving/loading for the purposes of glitching the game).
Let me put it this way: Would you think it acceptable if the speedrunner alt-tabbed to Windows and then used its file manager to delete the save file? If that's not acceptable, then why is it acceptable to do it from within the game? (And remember: "If it's supported by the game, it's ok" is not a very good argument. Remember the developer console, which is banned even though it is also supported by the game.)
I didn't check if it uses the reset button. If it does not, then it's done via gameplay yes. (But my problem with ACE is not whether it's triggered via gameplay or otherwise.)
If a glitch causes the execution of the program to jump to an unintended location, then it's a bit in that fuzzy area. In some particular situations I would consider it ok while in others it would be a bit bothersome. (After all, how do we define the execution of the program "jumping to an unintended location"? It can be hard to define in some situations.)
The ACE question and the "techniques used in speedruns" are completely different. ACE was only mentioned incidentally, unrelated to my original points. Don't confuse the two.
I'm not happy with, for example, deleting savefiles in HL2 speedruns because I don't consider it gameplay, but an external non-gameplay technique used to affect the game.
I'm not happy with considering ACE runs legit game completions, for rather different reasons (completely unrelated to gameplay).
These are two completely different and independent propositions. The latter is not related to the former. And the latter one is not the subject I wanted to discuss in the first place. The discussion veered towards it only because it was incidentally mentioned.
It's not navigating menus that's the problem. It's when you use none-gameplay elements in those menus.
You might just as well say "pressing keys is an indistinguishable part of playing the game (and therefore quicksave and quickload keys are the same thing as movement keys)".
It's not that I don't like it. As I said, they are marvelous demonstrations of technological knowledge and work, and can be very impressive.
The problem I have is when they are considered proper game completions (which, in my opinion, is one of the fundamental requirements for a speedrun).
This delves into the question of what constitutes game completion. It's not a trivial subject.
Not really. The level design was just an example.
Sometimes a wall can be bypassed by abusing actual programming errors or oversights (like bypassing that one gate in HL2 when getting out of the motorboat), and that's fine too.
It's when the glitch pertains to corrupting the game's data or program that it starts veering towards the uncomfortable side. I suppose that as long as it's done via gameplay only, it's still ok.
My original complaint is not about what kinds of glitches are used, but how they are triggered. Are they triggered via gameplay or via non-gameplay.
I don't know what you are talking about. Nowhere have I said that glitches shouldn't be used. I don't understand where you are getting that idea.
Once again: It's not about what is used, but how.
Well, as long as the glitch is triggered via gameplay, I suppose it's ok, even though it skips an enormous part of the game.
Sure, the distinction between "skip X levels ahead by triggering a bug in the game" and "skip X levels by ACE" can be fuzzy in some scenarios, I don't deny that. I suppose it depends on the details.
If the skip is done because of a bug in level design, for instance (eg. the level designers never expected the player to be able to jump high enough to get over a wall), then it definitely falls into the acceptable category in my books. If the skip is done by corrupting memory, though... Well, we enter the really fuzzy area. It's hard to give a clear opinion in that case. (I suppose that as long as it's done via gameplay, it's more ok than not.)
Optimally there would at least be a speedrunning category of that game that doesn't skip that much using that technique.
I don't think it makes much of a difference in principle whether it jumps to the middle of the game or the final boss.
No, I was talking about what should constitute a game completion. How do we define "speedrun".
The main point in my modding comparison was not that the game was modified, but that the game was not actually played. There was no "speedrun". It simply shows the end credits by bypassing the actual game code, and that's it. That doesn't feel like you completed the game by playing the game, but by other means.
It was indeed not my intention to talk about ACE. The conversation just strayed there somehow. (May I even say, derailed.)