[quote]Anything else come to mind?[/quote] Demo

For any% and minus world ending they are definitely identical (fceux/bizhawk at least), all the timings and mechanics are the same to the frame. In warpless or the SNES categories there could be slight lag differences: I don't see any differences to a real NES/SNES, but you can't test every single possible event.

EDIT: I should add that snes9x and snesgt run at 60 fps and higan runs at 60 fps by default too although it can be adjusted. The only good snes emu that runs at the correct framerate by default is bizhawk.

I don't think you need moderator approval, but have it anyway as far as I'm concerned.

Does anyone want more categories for this game? Imho I can't think of any one missing that would deserve to be a main category or even a misc. category as it stands now. Maybe move the misc. ones to main and use misc. for more arbitrary ones?

One I've seen requested often is Second Quest, although I personally see no point in this category as it is almost identical to any%. Warpless Second Quest may be different enough. Or Both Quests Any%/Warpless All-Stars as it is a continuous playthrough where you get to keep your power-ups and a slightly different ending, similiar to Lost Levels D-4 runs. But as it stands now, nobody has run these or expressed interest in running them.

Then darbian wants small only warpless. I'm not opposed to this as long as it stays misc.

I wouldn't be against adding glitchless or 1 warp and would personally run them, but I won't demand them, if others are against them.

Then the only other categories with some activity on SRL/PBTracker are All Coin Heavens, 500 Coins and partial categories like beat 4-4 warpless. I don't see a point in tracking these.

And to make it clear: I'm not pushing for any change, I set up the categories the way they are now. I'm asking what other runners think.

It's about fairness. There is anothing fair about this system where you can in theory "beat" a run on hardware that wasn't even released when the guy you beat did his run. There's a reason no relevant PC game community has this attitude: it would be blindingly obvious to even you that the guy who spent the most on his PC would have the lowest time every time. And you don't need arbitrary judge decisions for this either: You convert if possible, you set a list of allowed versions or make different categories if not. PC games do it, SM64 does it, the Mega Man community does it, we do it for Lost Levels too. The only reason not to do it is laziness and easy WR.

[quote]Or you could just get a PS3 like other runners from every other game and compete just as fairly instead of making a problem over nothing. Also PS3 is not an emulation, it is EXACTLY the same concept as running a PS1 game on a PS2, why haven't those communities done this either? Same concept as placing a GCN disk into a Nintendo Wii. Same concept as placing a Wii disk in a Wii U. Same concept as placing an Xbox game in an Xbox 360. What makes the PS3 special enough to warrant this? [/quote] All the communities who allow this are fucking degenerate already, no need to spread this shit further. Buying hardware is not optimizing a speedrun. Fuck off to reddit with this shit.

I just had a look at all the Bowser patterns from 4:57.01 to 4.57.39 (waiting 2 framerules on the title screen - equivalent to 4:57.36 to 4:57.74 waiting 1 framerule) and I found only 3 that kill you, if you jump close to his head:

• at 4:57.02
• at 4:57.04 very hard to avoid the hammers, but probably don't need to worry about these
• at 4:57.11 the hammer bro will throw a hammer and Bowser's fireball will be high, no other time in this range besides 4:57.02 will have this behavior So if you do see the hammer bro throwing a hammer and Bowser's fireball being at the highest possible position, prepare to do either of those jumps that are both very tricky: Yes, it's over his head again, but it's really really close.

Every other time (when you are in the 4:57.01 to 4:57.39 range) you will always be safe either running under him or jumping slightly over his head. It seems the current common method jumps too high and is an unnecessary risk. Maybe people who actually savestate practice 8-4 today will be able to confirm this?

I was really looking at something else that turned out to not be really relevant in this instance: The bubbles that come from Mario's mouth in the underwater section are on a global timer. By looking at where the bubbles are when the screen transitions to the final screen and also looking at the hammer bro and fireball, you have a lot of information and can pretty much predict exactly what Bowser will do, but now that information is pretty useless unless you are <4:57.06. Maybe for higher PBs or warpless looking at bubbles can become useful, if someone wants to look into it.

This one looks really hard honestly.

The WR's method can be done more easily, if you do a backwards jump to the spot you wait at and start holding right+down while in the air. This way you will keep facing left while waiting, but as soon as you jump and leave the ground the right input will be considered again. So you don't need to time pressing A and then right.

I think this method should be tested more: You either stomp or avoid the top bullet, when it would normally be in the way, and you make a full jump from the single block. You have a ~50% chance of the subpixel being correct and the trick being possible at all.

Here you need to time when you start running and when you do the full jump from the block and with the WR's method you need to time when to do the full jump on the block and when to do the full jump from the block. So it seems to be objectively better than the WR's method, because the difficulty is the same but it works with every top bullet RNG.

The streaming rule is good against cheaters who put in low effort. For example this run: Is it legitimate? We don't have anything but this single recording to base our decision on. Maybe it is real, but maybe he made this video in a few hours with savestates. When you see someone streaming you know that he actually does the attempts he claims to do. Maybe tomorrow darbian streams a cheated 4:57.2X because he got frustrated after 10000 attempts and we most likely would not find it out, but he still put in a lot of legitimate effort before and he would lose all respect in the community, if we found out. If a newcomer says he got the record after 15000 attempts offline, it's not possible to tell if it's real or if he only put in a few hours of savestating a record and he also loses nothing, if he is found out.

nicovideo, cavetube, stream down during run: The technical problems during livestreams that you describe have to be accepted imho, because you can't avoid them. But the submitted video has to be complete.

low spec PC, no stream tool: I see the problem here, but how usual is a PC/internet that can't upload ~300 kb/s? @PackSciences: What about Nestopia?

@JWILD [quote]It also brings up another question. Seeing as how emulators are allowed and how WiiVC is allowed, what about emulating Wii VC ? I prefer Dolphin over Nestopia for running SMB1 and if I get a PB on Dolphin's emulated Wii VC, I would be very upset if it didn't count.[/quote] It would just be listed under WiiVC[EMU] (and converted accordingly, if conversion happens). Emulation accuracy would be unlikely to be an issue, since it's just emulating software and stuff like the Wii lagging is unlikely to happen.

[quote](btw does Nestopia do this?)[/quote] Yes as far as I know.

[quote][...] Allowing Wii VC for certain cutoffs lets people who haven't played the game and want to try it start running, but if you want a WR or something you better run at the correct frame rate! IMHO[/quote] Well I don't really see the point of what you are proposing. It's only a few seconds anyway, so it would only really matter for optimized times, which you think should not be converted.

@nickj109 (replying to a 6 months old post lol) [quote]No. There are emulators that emulate at the correct frame rate (FCEUX, Bizhawk, I think higan does too?, etc.) and those emulators should be used. If they choose to use an emulator that runs at a slower frame rate, they should accept the time loss. Unlike the physical consoles, emulators are available to download on the internet and the choice of which one you use is free.[/quote] agwawaf's SMB2j times without loads were already converted from 60fps when I added them. Since I pretty much agree with you and the rules don't really say that, I've reverted it to the 60fps times.

[quote]There definitely should be cutoff points. I feel like exceptions could be made for people who have a reputation built up already and are trustworthy, but I like the guidelines you suggest. I think it should be sub 5:10 should be streamed while sub 5:30 requires video proof (However for this game on speedrun.com, it's set so you can't submit a time unless you provide a video, so idk) [/quote] [quote]I think there should be standards of proof for anything faster than 5:21.00 . I agree with i_o_l and nickj109 .[/quote] So I'll add it to the rules then since nobody disagrees.

NTSC NES is the "standard version", it's what every high ranked run uses and as far as I know the fastest. American, Japanese and the multi-carts are identical for speedrunning (except longer resets on the multi-carts). PAL NES (European) is significantly slower. FDS is almost identical to NTSC NES, but there's no reason to use it, if you don't do Minus World Ending, as it has a longer bootup time and more expensive hardware and no benefits. WiiVC and WiiUVC are identical to NES (A European Wii will have a PAL rom, but I guess that you can hack it and load a NTSC one on it?) except for the framerate, which is slower. I asked if VC times should be converted to be competitive a while ago, but there were no strong opinions either way. GBA and 3DSVC I don't really know.

SNES is its own category. Japanese is fastest. Again the Wii version of the SNES version is slower due to the framerate.

Since I see many people not know how to retime their runs or do ridiculous stuff like rewatching a 2 hour run in real-time with Livesplit running next to it, here's how I time recordings:

What you need:

• youtube-dl https://rg3.github.io/youtube-dl/download.html despite its name it can download from many places including Twitch
• VirtualDub https://sourceforge.net/projects/virtualdub/files/virtualdub-win/1.10.4.35491/VirtualDub-1.10.4.zip/download for the actual timing
• K-Lite Codec Pack http://www.codecguide.com/download_k-lite_codec_pack_basic.htm to decode the video
• Avisynth https://sourceforge.net/projects/avisynth2/ to load the video into VirtualDub
• FFmpeg (only for downloading from Twitch) http://ffmpeg.zeranoe.com/builds/ After downloading it go into its "bin" folder, copy "ffmpeg.exe" and drop it into the same folder as "youtube-dl.exe".

How to time the length of a recorded run: I will time TheTycon's Super Mario Land run as an example:

1. Go to the folder that holds the "youtube-dl.exe" you downloaded, hold Shift and right click it. Then select "open command prompt here".
2. Enter "youtube-dl " and paste the link to the vod and press the Return key
3. Open the editor and create a new file with the ending ".avs" that contains just this line: DirectShowSource("path to the video")
4. Drag the avs-file and drop it on VirtualDub.exe
5. Go to the starting frame and press the End and the Delete key to remove all previous frames.
6. Go to the ending frame and read the time And that's it: 12:36.857

This doesn't mean the time is 100% accurate. You have to take into account that many games are 60fps while recordings are often just 30fps. A lot of Youtube uploads have interlacing or other encoding issues that may mess with the timing. In these cases you can take a section of constant known length and compare it with the run or compare with what the onscreen timer says, if there is one, to see if the general timing fits. Twitch highlights are often more reliable because them being live ensures that at least the timing isn't screwed up afterwards.

time without loads You can go through the movie and use the selection keys (Pos1 and End or the right-most buttons at the bottom of the window) to select and delete all loading sections to read off the time without loads the same way.

when youtube-dl doesn't work If you want to download from some obscure site that isn't supported, the following method may or may not work:

1. Start Firefox press Ctrl+Shift+K to open the console.
2. Select the Network Monitor tab at the top of the console.
3. Go to the site with the video you want to download and let it start playing.
4. The Network Monitor should be filling up with files now. Sort by size and the file you search should be near the top of the list.
5. Right click the video file and download it.

[quote]nah apparently beating the game gives you 87% lol[/quote] There are 2 bonus missions.

In general there are 3 runs in 6 categories right now. One could argue that this is too many categories rather than too few.

What is the difference between All missions and 100%?

Not removing loads from PC runs is a fucking joke.

A lot of the stuff you say in the opening post is not as set in stone as you seem to think it is. Some communities decide to have all versions on a single board, some decide to have a seperate board/category for every release and some decide to hide some versions by default. Some communities decide to remove loading times and some decide to keep them. And the way you arrive at a community decision is by requesting the game first and only adding mods that agree with you.

You set the clock and start the game at a specific second and then you know the board layout. That is all.

Isn't it nonsense that someone can be the fastest and the second fastest at the same time?

It could maybe be added under Romhacks/Mods, but why not just play on emulator?

[quote]BizHawk is incredibly resource-intensive in my experience[/quote] That's not generally true.

1. Bizhawk has rewinding enabled by default. Disabling it will make it run much faster.
2. The speed depends on the console and core that you are using. SNES emulation for example is very slow, but GBA gets 400 fps, GB with the Gambatte core and NES with the QuickNES core get 1000 fps for me, when I disable throttling.