Did you use Fractal's input for the rest of this?
Edit: MUGG did use Fractal's input, so I listed both authors. Also, I made an MH TAS of Urban Champion in 40 presses. I shrunk the time to 00:49.98, but I can't upload until Microstorage is back up. I bet a bot would be very useful to shrink it further, and it's a pretty trivial process from what I can see, so I'll look into doing that.
Edit 2: Improved Urban Champion by me, NES Donkey Kong in 35 presses by nitrogenesis.
Joined: 3/9/2004
Posts: 4588
Location: In his lab studying psychology to find new ways to torture TASers and forumers
Thank you guys, especially bkDJ, Brandon, FractalFusion, MUGG, and nitrogenesis, for looking into and working on this crazy idea of mine.
I really like what I'm seeing.
Warning: Opinions expressed by Nach or others in this post do not necessarily reflect the views, opinions, or position of Nach himself on the matter(s) being discussed therein.
Sonic Advance's Neo Green Hill Zone as Knuckles in 31 presses.
Movie input frames: 11844
Average presses per frame: 0.0026
Frames with any buttons held: 6806 (57.5%)
There are instances where you have a one-frame window and need perfect pixel positioning to complete the stages as fast as possible (examples of them are avoiding the special stage spring on the first act, avoiding the spring hidden in the grass at the end of the second act or landing certain hits on the boss), and I think that is pretty cool.
I would appreciate feedback or improvement suggestions. Should I continue?
Link to video
Thanks to dada for the encode.
AzumaK wrote: I swear my 1 year old daughter's favorite TASVideo is your R4MI run :3
xxNKxx wrote: ok thanks handsome feos :D
Help improving TASVideos!
Warning: Opinions expressed by Nach or others in this post do not necessarily reflect the views, opinions, or position of Nach himself on the matter(s) being discussed therein.
The movie should be labeled "minimal (button) presses TAS" explicitly, and there must be a "total buttons pressed" counter in the encode. Otherwise it just looks like a bad speedrun. And maybe some additional counters for sections like bosses. Problem for the last idea would be that this will always be chosen somewhat arbitrary.
I found it interesting how different the boss fight looked from a normal playthrough.
If we say that "minimal presses" is a good goal, then it should be "minimal presses ingame" for this particular game, as I don't see any point in waiting for all those menus.
If we say that "minimal presses" is a good goal, then it should be "minimal presses ingame" for this particular game, as I don't see any point in waiting for all those menus.
Thats a good point and imo should apply to _any_ games, why force the audiance to watch cinematics and stuffs ? thats often not interresting (ok maybe not always but usually it is irevelant)
I like the "minimal presses in-game" interpretation.
Allegedly the focus is on entertainment value.
I'd love to see more Sonic runs done this way. I think there's a lot of potential for win.
What's the difference between this and "Minimum Holds"?
partyboy1a wrote:
The movie should be labeled "minimal (button) presses TAS" explicitly, and there must be a "total buttons pressed" counter in the encode. Otherwise it just looks like a bad speedrun. And maybe some additional counters for sections like bosses. Problem for the last idea would be that this will always be chosen somewhat arbitrary.
If these runs were to eventually be submitted, they would clearly be done so under a branch, so yes, they'd explicitly labeled as such. I already have created a Lua script to count both presses and holds while replaying a video (Though this doesn't seem to work in VBA for some reason), and with andymac's AVI-HeadsUp script, I should be able to work this into the encode. I will experiment this if / when I find a movie that is polished enough to warrant an encode.
partyboy1a wrote:
If we say that "minimal presses" is a good goal, then it should be "minimal presses ingame" for this particular game, as I don't see any point in waiting for all those menus.
The problem with exceptions like this is consistency. When you start adding conditions onto what parts count as presses, you get problems, and it'd be impossible to create scripts that score these runs properly with them. However, I wouldn't mind a "Press/Entertainment tradeoff" category just like we have with our speed runs. The submission text could mention where extra presses were taken in order to prevent alienating the audience, and the judge could decide whether this was a good decision or not.
OK, at some point, I hope to continue working on my Lua script and a PHP script that calculates the presses from a movie file. After thinking about this for a while, I think I have to admit "Minimum Holds" might not be the best term for 1 button for 5 frames = 5 presses. At first, I didn't want to use "Minimum Input" because input files would be the same size regardless of the buttons pressed in them for a given number of frames (At least with NES), but people would probably be able to distinguish between Minimum Input and Minimum Frames. Any thoughts on this?
It's difficult to come up with a short and clear name for the two concepts. "Minimum presses" could indeed relatively unambiguously refer to the minimum amount of button state changes from unpressed to pressed, but a term for the count of frames that a button is held down is more difficult. Perhaps there isn't anything better than "minimum holds".
A thought I had: What if minimal runs were scored as frames * presses (with the lowest being the best)?
I don't think they should be combined at all whatsoever; you would inevitably have to sacrifice presses for speed.
By the way, in an attempt to breathe life into this project, I created a TAS of Donkey Kong Jr. yesterday. Yay...
If we say that "minimal presses" is a good goal, then it should be "minimal presses ingame" for this particular game, as I don't see any point in waiting for all those menus.
The problem with exceptions like this is consistency. When you start adding conditions onto what parts count as presses, you get problems, and it'd be impossible to create scripts that score these runs properly with them. However, I wouldn't mind a "Press/Entertainment tradeoff" category just like we have with our speed runs. The submission text could mention where extra presses were taken in order to prevent alienating the audience, and the judge could decide whether this was a good decision or not.
I envision a sort of solution to this: code a sort of extra button into the emulators being used (I know there's a lot of different ones) that sets/unsets a flag for whether or not the button presses at a particular time in the TAS count for the press total i.e. whether they are ingame or not. The flag would be visible on the screen during playback just like the arrow key symbols and A and B button are displayed in the Sonic run posted earlier. If the runner used that flag to show what he thinks is "ingame" in the game he's running and the verifiers agreed, the script-generated count (which would be made to reflect whether the flag is up or not) would be correct again.
What do you think of this idea?
I can still see the game-specific rules thing is going to be a problem, but at least this would restore scripted counting of button presses if it works.
If we say that "minimal presses" is a good goal, then it should be "minimal presses ingame" for this particular game, as I don't see any point in waiting for all those menus.
The problem with exceptions like this is consistency. When you start adding conditions onto what parts count as presses, you get problems, and it'd be impossible to create scripts that score these runs properly with them. However, I wouldn't mind a "Press/Entertainment tradeoff" category just like we have with our speed runs. The submission text could mention where extra presses were taken in order to prevent alienating the audience, and the judge could decide whether this was a good decision or not.
I envision a sort of solution to this: code a sort of extra button into the emulators being used (I know there's a lot of different ones) that sets/unsets a flag for whether or not the button presses at a particular time in the TAS count for the press total i.e. whether they are ingame or not. The flag would be visible on the screen during playback just like the arrow key symbols and A and B button are displayed in the Sonic run posted earlier. If the runner used that flag to show what he thinks is "ingame" in the game he's running and the verifiers agreed, the script-generated count (which would be made to reflect whether the flag is up or not) would be correct again.
What do you think of this idea?
I can still see the game-specific rules thing is going to be a problem, but at least this would restore scripted counting of button presses if it works.
I think this is too convoluted. I still think the "Presses/Entertainment tradeoff" idea is the best solution possible, but I would discourage ever using that to begin with. The goal of these runs are not to be entertaining; it's to show how many buttons it takes to beat a game. If we justify sacrificing button presses for entertainment in the case of menus, there are other places where it could be justified as well. For example, in mugg's SML run, I could say that it's boring to watch him wait for the timer to go out, and then justify that he sacrifice that interesting button saving trick for the sake of entertainment. I think the lowest amount of presses should win, always.
Improved the SMB MP TAS. Super Mario Bros (JPN/USA PRG0) by Brandon Evans in 145 presses and 05:46.33. In addition to tightening up the speed significantly, I walked on a Bullet Bill in 8-2 to avoid jumping, and I luck manipulated the Hammer Bros. to be completely irrelevant in 8-3, saving two presses total. 8-3 can probably be optimized for time as the manipulation is fairly complex. Also, I had to delay the beginning of the 8-4 swimming section by 1 frame to make it possible to kill Bowser. I'm sure there are plenty of other potential speed improvements for those who know more about the game physics than I do.
Also, Kung Fu looks like a really good game for a Minimum Holds TAS. The strategy would be to tap left once, jump, and continue jumping whenever you hit the ground. You could also try to make it so that the minimum number of enemies spawn. When the enemies do spawn, you can time the initiation of the jump sequence so that you can always kick them while in there air, and the kick helps you stay in the air. None of this is particularly trivial, but it might be a fun challenge for anyone who'd like to try.
Never realized there was a forum topic for these sorts of TASes until recently. I have some TAS that I worked on back in 2014 and might as well post them somewhere, even if it is super late to when this topic had activity. The movie files were created using Gens v11. Rerecord count for minimum inputs is not accurate as I cleaned up some extra inputs in a separate file. The original had 9758 rerecords.
Sonic the Hedgehog 2 Rev01 in 320 pressesSonic the Hedgehog 2 Rev01 in 1886 inputs
