thegamer1185's Feed

thegamer1185
02-21-2019 at 03:55 PM
13 Comments
Rate this Entry

Speed runs.

OK everyone, so I've been "trying" to adjudicate timed runs at TG, and there is really no good of doing this using the video uploader. Is there any way TG could add a frame counter to the video uploader, or in a much easier fashion, find some program or timing device and make it the official timer used for speed runs at TG?

We have had enough time go by where the video uploader plays slightly different enough for the adjudicators that we are having to compromise with the submitter to get an acceptable time for submissions. If we want 100% accuracy in the data base, I think the time has come for something to officially be done about the timing issue. At the current stage, TG is inferior to all speed running communities that expect to the hundredth of a second in accuracy. TG we can only do to the second.

Not criticizing anything, just think we have seen enough back and forth submission discussion on times it's time to get this right. After all like many of us have just settled on at this point, the leader board is the only thing that really matters. So lets get it right. Thoughts?

Comments
  1. Barthax's Avatar

    There is no 100% accuracy currently. People are attempting to time to milliseconds when the evidence is too poor to be accurate to that measure. For milliseconds accuracy it is necessary to have 100FPS or better footage. It is a fallacy that anything in speed running is accurate below 0.1 and even that is often inaccurate due to the frame rates. 30 FPS is only potentially accurate to 0.033 which means it cannot meet an exact 0.1 resolution. Most camera footage won't be 30 FPS due to dropped frames or lesser FPS from the camera.

    Then add the rulings in and you've got vague start points and vague end points. Community discussion is the only recourse to the poor technology capabilities of today.

    Likesthegamer1185 liked this post
    Updated 02-21-2019 at 04:24 PM by Barthax
  2. thegamer1185's Avatar

    I guess saying we need 100% accuracy was just a vague statement. Speedrun.com, Mario Kart World Records, and other sites that are strickly speed running are far more pin point than TG. I do understand that there will always be some in accuracy, but TG is at best 1.99 seconds off because we can only view whole seconds. However, if TG does establish a program that allows us to go frame by frame, and is the ONLY allowed program, there can't be any arguments. I know there are programs that can already do this, but I've seen submission where one person uses one program, another uses another, and they can't agree. But if it is a single program, it gets rid of those issues.

    Timing could be much better. If it's up to the community, there can always be what I used vs what you used, etc. If TG makes a definitive decision on how to do it, or just picks a program to use, that gets rid of a lot of variables.

  3. Barthax's Avatar

    Quote Originally Posted by thegamer1185

    I guess saying we need 100% accuracy was just a vague statement. Speedrun.com, Mario Kart World Records, and other sites that are strickly speed running are far more pin point than TG. I do understand that there will always be some in accuracy, but TG is at best 1.99 seconds off because we can only view whole seconds. However, if TG does establish a program that allows us to go frame by frame, and is the ONLY allowed program, there can't be any arguments. I know there are programs that can already do this, but I've seen submission where one person uses one program, another uses another, and they can't agree. But if it is a single program, it gets rid of those issues.

    Timing could be much better. If it's up to the community, there can always be what I used vs what you used, etc. If TG makes a definitive decision on how to do it, or just picks a program to use, that gets rid of a lot of variables.


    Rose tinted specs, I'm afraid. Any tool is only as good as the evidence presented and lots of evidence being presented is generically crap currently - speedrunning sites included.


    Even if a tool was standardised upon, that then limits all future technologies to that tool and whatever failings that tool has (which it is bound to have).

  4. Snowflake's Avatar

    Quote Originally Posted by Barthax



    Rose tinted specs, I'm afraid. Any tool is only as good as the evidence presented and lots of evidence being presented is generically crap currently - speedrunning sites included.


    Even if a tool was standardised upon, that then limits all future technologies to that tool and whatever failings that tool has (which it is bound to have).


    oh god help me, but this argument against you here is gonna actually play into your argument why directfeed is better...


    what you're saying is true for analog as theres always more details. not so for digital. think turn based. if there are only x turns, you can record them all. all games are basically turn based where every frame is one turn. as long as you have every frame you'll have all you need.


    now, while that means i do believe theres a permanent answer its still ugly cause good look getting us all to move over to direct feed. i would agree with your statement that for anything short of direct feed we'll always have issues

  5. sdwyer138's Avatar

    Maybe that stuff should just live a SR, and TG should only track time games with in game timers?

    Likesthegamer1185, Blackflag82 liked this post
  6. Barthax's Avatar

    Quote Originally Posted by Snowflake



    oh god help me, but this argument against you here is gonna actually play into your argument why directfeed is better...


    what you're saying is true for analog as theres always more details. not so for digital. think turn based. if there are only x turns, you can record them all. all games are basically turn based where every frame is one turn. as long as you have every frame you'll have all you need.


    now, while that means i do believe theres a permanent answer its still ugly cause good look getting us all to move over to direct feed. i would agree with your statement that for anything short of direct feed we'll always have issues


    The subject matter is manually timed speed runs and in turn based, there's still screen transitions, disc read errors, human response, dropped frames (direct feed), camera transition issues (blinding white frames while the camera tries to focus on the end sequence which is the timer stop).


    ... but yeah, direct feed is better for this particular purpose given the current limited technology. As I mention, dropped frames still happen so is still inaccurate even within it's own boundary. Then it has to be uploaded to TG & is the re-encoding by TG _perfect_?

  7. Snowflake's Avatar

    dang, i guess i was being too optimist.

    i was actually abuot to go almost poetically about how amazing digital is for preservation, but then i got control of myself and i'll just say as sweet as digital is you're right i overestimated it

  8. thegamer1185's Avatar

    Quote Originally Posted by sdwyer138

    Maybe that stuff should just live a SR, and TG should only track time games with in game timers?

    Man, I was going to say this but since I'm so shy about my opinions, I just couldn't. Ha, made myself with that one. Completely agree. We have a hard distinguishing and leeching and point pressing anymore, figuring out speed running only adds to it. I do appreciate the systems we have, I just think things need a bit re worded, updated, improved.

    LikesBarthax liked this post
  9. Barthax's Avatar

    Quote Originally Posted by sdwyer138

    Maybe that stuff should just live a SR, and TG should only track time games with in game timers?


    Participants want this and TG should support it (and it does). TG has placed the adjudication in the hands of the community and it's just that the community should accept the shortcomings of:


    a) The technology displaying the content (possibly >100FPS for PCs, less than 30FPS for most consoles)

    b) The technology capturing the content (possibly 100FPS on a top-range PC capturing its own output but at best 60FPS on modern consoles with direct capture, but 30FPS or less for most camera)

    c) Humans stupidity (writing rules, reading rules).

    d) Human gullability.

    e) Humans are involved and argue.

    f) Other humans...

    g) ****, another crowd of humans with a different opinion from the last.

    h) More bloody humans who think they have the answer because they haven't done the basic maths that's it's not accurate.

    i) Another bunch of humans that don't read what all the other humans have agreed upon.

    j) Obnoxious twats.

    k) Know-it-all-arseholes (me)

    l) Trolls.


    LikesGarrett Holland liked this post
    Updated 02-21-2019 at 05:25 PM by Barthax
  10. Barthax's Avatar

    Inject between c) and d)


    c1) The conversion process provided by TG could alter the footage incorrectly.

    c2) The viewer of the converted TG footage could alter the viewing of the re-encode footage.

  11. Ragequit's Avatar

    Quote Originally Posted by sdwyer138

    Maybe that stuff should just live a SR, and TG should only track time games with in game timers?

    This is why i don't like the tekken tracks. I mean the game already has an internal timer, why do they have to make things more complicated by forcing you to measure time starting from the character select screen when the internal timer shows the real time playing the actual levels of the game? And this internal timer is what is used to generate high scores in the game anyway.
    Likesthegamer1185, Barthax, Snowflake liked this post
  12. thegamer1185's Avatar

    Agreed. If there is an in game timer, use the fucking thing.

    ThanksRagequit thanked this post
    LikesSnowflake liked this post
  13. Tompa's Avatar

    In game timers are sometimes inaccurate and they don't display the actual time the game takes to play through. Lag reduction strats can be a pretty major thing in certain speedrun games, if you use the IGT, this is completely ignored and it becomes more of an easy mode.

    A game like Donkey Kong Country only tracks hours and minutes. So a run that's 00:20:00 and 00:20.59 will both appear as 20 minutes. The timer in Metroid Other M will include the time in cutscenes, even if you skip them. For the level Riverside Race in Donkey Kong Country 3, you can actually beat the level quicker but get a worse time, as the timer doesn't increase on every frame, it's actually going backwards sometimes. 51.55, 51.56, 51.57 are three frames of the timer, but on the fourth frame it will display 51.50, 51,51 and up to 51.62. For this one, the IGT is still being used for runs of the specific level, just a general example of timers being inaccurate.

    Timers are good at times, but depending on the game, it's not to be trusted. Some games have combinations of the two, like Super Metroid or the Cups in a Mario Kart game. Both ways of timing have their ups and downs.

    LikesSnowflake liked this post
Join us