Canada
SupporterThat's certainly a way to do it. We only require to see the entire run up until the stage the submission is for, just to be sure there's a proper reset, and a reasonable effort to time the run between the stage's start fanfare and the last hit on your last opponent.
We also accept stage grind attempts, so long as the run also includes the last reset up to that point.
First frame is the first frame of movement after "DUN DU-DUN, DU-DU-DU-DUN, DU-DUUUUUUUN!".
Last frame is when your opponent's last energy bar disappears on the top HUD.
As counter-intuitive to speedrunning this game as that sort of cheat would be (ie: dying = slow), I'd say it would still be prohibited.
No worries, I redid the count on my end.
This tool is great for YT videos: https://slashinfty.github.io/yt-frame-timer/
Yes, we have to verify them all, but there is a reason why we need the entire footage: we need to see that it was done from a reset.
With a cut out video, there is nothing indicating that the run wouldn't have been started from a save state.
Mesen was previously banned due to what was believed to be an undetectable feature (ie: run-ahead lag reduction) that could yield an unfair advantage from a speedrunning perspective. A fairly simple detection method for run-ahead was brought to my attention, and I find that it's easy enough to warrant allowing Mesen back.
So from now on, Mesen is no longer banned. However, showing the input display will be MANDATORY. Mesen runs without the input display being shown will not be accepted.
The IL leaderboards are in, people! Enjoy!
General rules for ILs are:
- Video must include last reset, and the run up until the intended stage is cleared.
- Milliseconds must be included in timing.
- Time starts on the first frame when movement is allowed on the intended stage.
- Time ends on the frame on which your opponents' last health mark disappears from the top HUD.
The question here is simple: would you be interested in having individual level boards for each level? I personally believe there would be some merit to that, the same way there is merit in Punch-Out, for instance, to have IL boards for each boxer.
So yay, or nay?
Elo doing the very bare minimum. Why am I not surprised...?
How do you demonstrate your sum of best, then? You'd need each video of each of your best stage attempts separately in the submission, because you can be sure that splits.io data will most definitely not be sufficient evidence, nor would it be viewed as conclusive evidence. Anyone can run LiveSplit without necessarily running the game. And editing the videos together could open the door to potential cheating.
That would make about... 8 videos for any%, and 32 videos for warpless. I personally think none of the mods would want to put up with that, first because they've already got their hands full with submissions as it is, and second because it would hardly be worth it if the demand represents merely a slim minority of the community.
Imagine if a runner posts a warpless SOB run with timing inconsistencies because they understood the timing regulations poorly. The mod reviewing this would then have to go and retime each and every of the 32 videos in there? No one in their right mind would agree to that sort of thankless work. And then what if the mod rejects the run? The runner will either complain about it, or resubmit it, probably with errors still. Which would yield the same result, and ultimately frustrations on all sides.
I personally vote no, because to me, it just seems like an excuse to get a new shiny trophy. And in my opinion, if you're speedrunning solely for trophies, you're doing it wrong.
Not to mention that MiSTer's cores are as close as can be to the real thing in terms of cycle accuracy. Meaning: identical.
But to be fair, I've been met with this sort of resistance from communities that didn't know about it and wouldn't bother looking into it because "it wasn't worth it", or because "it's just emulation by another name" and that by this virtue alone, it made MiSTer/SiDi FPGA platforms "worse than software emulators". Ergo, they were speaking from willful ignorance.
But then again, I can understand. How can you trust "untested and unproven" platforms such as these? I mean, it's not like big communities like Z2, Mega Man, ALTTP, or even one of the pioneer speedrunning communities in the form of the Super Metroid community are allowing it, right?
Oh wait... they actually do!
Going forward, rog_pf will be permabanned from submitting runs to the Double Dragon leaderboards unless they accept to be proofcalled based on our standards.
I have been informed by several moderators of other arcade runs they plays that their runs were very suspicious. First of, their twitch account is fairly new, and then they claimed several world records in several well-known arcade games. Which would raise eyebrows, but by itself would be plausible, while still being improbable. Second, it appears that their YouTube channel is just as old as their twitch channel, with several runs being uploaded in a short amount of time. Typically, when something like this happens, we are dealing with a cheater.
Notwithstanding the above, several mods reported that Rog's runs are TAS-level perfect, which I tend to agree with. In their DD run, they seems to be able to consistently pull the punch-uppercut strat on Abobos, which requires amazing timing, and which is bound to fail quite a few times in a run. However, all of their twitch streams consist of him leaving attract mode on for a while, talk, then "run" the game perfectly and end the stream there. Which is definitely NOT how speedrunners play. Normally, you'd see quite a few failed runs, and more importantly, you wouldn't see the stream end 18 minutes in because the runner would get a WR. The fact that this patterns occurs exactly as described on all the games they ran not only indicates a pattern, but a modus operandi.
I am highly suspicious of these runs being rough TASes being played back. Done with just enough "skill" to beat current WRs while not looking all TAS-like. However, given how ridiculously lucky and/or precise one has to be for the Abobo punch setups, I am convinced that their runs are TASes being played back.
Furthermore, if I can be blunt, the reactions seen on the videos and their acting skills are... subpar at best? Plus, the presses and controller movements don't match what we see. Nor do the button press sounds.
Finally, when I asked him for further info, he would not respond. I bet they will now that this decision is enacted, but the terms will remain the same, and will not be negotiable.
As such, after discussing with the other mods, rog_pf is banned from submitting runs until they are WILLING and CAPABLE of demonstrating they possess the skill level required to match the performance they submit on a livestream to which one of our moderators WILL assist. This proofcall WILL include Rog having to perform arbitrary actions during his proofcall runs in order to demonstrate that this is not a TAS being played back. And the execution will have to be on par with what his runs show. Should he choose to accept this proofcall, we will of course grant him a reasonable number of runs as we all know that several circumstances could make it so that we don't get it right the first time. While we will be understanding on that front, we will also be strict on anything we find suspicious.
We will give rog_pf 7 days to decide whether or not they want to go through with this. Should they not respond or refuse to go through with this, the ban will remain, and no further chance for appeal will be granted.
In summary, unless rog_pf agrees to prove his skill level in accordance to our standards, he will remain permabanned from submitting to the Double Dragon leaderboards. Given his other run-ins with other arcade game moderators, we are confident that his submissions cannot be trusted, and that said submissions are a detriment to the integrity of this game's leaderboard.
Please note that due to its frame run-ahead feature, runs using RetroArch will no longer be accepted. Furthermore, all previous runs using RetroArch will be rejected, as this feature cannot be detected.
Thank you.
The original post of this thread is far more meta than anyone could give it credit for. It's a question, an attempt, and a guide, all at once!
Looks like the site's borked again.
Clicking on any of my notifications redirects me to https://www.speedrun.com/run1.
FYI: MESEN is banned on a LOT of NES games due to its run-ahead feature.
That idea is so flawed in so many ways. No matter how you flesh it out, it would be easily cheeseable. Plus, not all speed games are created equal, so you can't measure "the best speedrunner" simply by sheer volume of games/WRs held. Finally, all it'll create in the end is more drama over a metric that, quite frankly, will only serve as an ego boost to the very few that would care about such a title.
I mean, we as a community have always prided ourselves over the fact that we're far more cooperative than competitive. This sort of idea would end up destroying this facet of our community... and for what? For a few people that need a self-esteem boost? Nah, I'll pass.
