You don't need a leaderboard to do a speedrun. A few runners would do some silly runs every now and then to break up the regular category grind. We didn't need a leaderboard we just played and had fun. Really the only overlap in any of this was me and andrewg occasionally both messing around with 1-1 fireflower or 10 lives, and maybe a few different people had done coop runs. We made the meme leaderboards with a few of these categories in mind.

If you want to run a new category every week, that's fine, play the game however you want. But you don't need an SRC leaderboard to do it, and it doesn't really make sense to have one anyway because there won't be enough people submitting to hundreds of different categories to warrant a verified leaderboard. A verified leaderboard loses all meaning if it only has 1 submission in a category. I suggest you guys just make a google doc/spreadsheet and keep track of your times amongst yourselves if you want to compete in lots of new categories.

TL;DR: you don't need someone else or a leaderboard to tell you that you can run a category, just play the game however you want and have fun.

@Eddie, Sockfolder FPG saves 15 frames over normal top of the flagpole. A perfect FPG saves 19 frames. Fast accel can save 2. It's humanly possible for a human to get this framerule. But I agree as of now it seems unlikely in RTA. The main point of the setup in this thread though would be to save a run that has bad judges.

Things like these really need an explanation of the setup lol. It would actually be useful if you miss the star block. But from what I can tell it's just normal flagpole glitch setup but with a random slowdown before the last pipe..?

You can thank dysfunctional leaderboards that were created by people who don't really run the game. GS submission was a meme and I guess someone (probably Kaons) actually updated some of the slop.

I responded to your comment on my youtube video but I may as well post this here also:

This isn't really accurate. Perhaps you took your framecounts at the black screen transition instead of after it. There's something called the 21 frame rule which means all time saves come in 21 frame or 0.35 second increments because the game only loads the next level periodically on those intervals. This doesn't apply on 8-4 because no new level is loaded afterwards.

The accurate comparisons are as such:

TAS 1-1: ±0 frames (0.00) 1-2: -21 frames (0.35) 4-1: -21 frames (0.35) 4-2: -63 frames (1.05) 8-1: -84 frames (1.40) 8-2: -105 frames (1.75) 8-3: -126 frames (2.10) 8-4: -170 frames (2.84)

Recently I tried to find a flagpole glitch setup in 8-3 by walking at the beginning of the level and beginning to hold B on a specific frame. I found a setup, but it wasn't very good. The next day, Pallidus found a much more effective setup.

Begin by holding right and NOT B. Start holding B on the frame shown in this image:

You can know if you did it right based on the enemy patterns in the level. Specifically, the first set of hammer bros will have 1 in the middle position, and 1 on the bottom, and the bottom hammer bro will throw two hammers. Here is an example of the enemy patterns after arriving at 8-3 on an optimal bullet bill glitch run:

It's important to note that if you interact with enemies as you go through the level, later enemies will act slightly differently.

Optimal framerule with killing first 2 bros:

Optimal framerule without killing bro on second set:

DevilSquirrel did additional testing and found the results you get if you hit B on different frames:

4 frames early: DB0

3 frames early: E30

2 frames early: C50

1 frame early: CF0

perfect frame: D70

1 frame late: DF0

2 frames late: D60 (This frame works!)¤

3 frames late: CB0

4 frames late: D30

5 frames late: CA0

6 frames late: E50

Notice that pressing B 2 frames late is also a working frame! Here are what the hammer bro patterns look like if you happen to press B 2 frames late:

And if you kill both bros in the first set:

¤Unfortunately the patterns look identical for pressing B 2 frames late and pressing B 1 frame late.

All of this data is only accurate if you start 8-3 on the right frame rule and if you do the inputs on the stairs correctly. Note that with the main B frame it's possible to get D70 and D50 (shown in the videos), while only D60 is possible with the 2 frames late setup.

TODO: Add videos for unoptimal BBG patterns.

Hello could the K in my username please be changed to a capital K :)

Just for the sake of being accurate, you need to move mario forward 20 pixels. His X position on the screen is 132 (112 by default), so that's probably what you're thinking of. Good bumps give you 10 pixels (and on rare occasions 11), hence why "Fast 4-2" is possible.

In addition to the excellent answer by Nickj I'd just like to add the best time that someone has actually gotten with said strategies. I thought Andrewg had a 4:58.29 but I can't find it so I'll link his 4:58.34

Hmm I don't see that it has been edited. Your submissions must not be going through for some reason. If you want me to fix the date and add the comment I can do that. Just let me know what it is.

I've been doing a lot of hunting for new strats as of late and have found several things. Some things only save half a second or maybe a couple seconds, and I won't make a video and list each one of those here, but I will post significant time savers. I'll continue to add more as they're found.

Things found so far:

5-5 3rd star coin strat (saves 6 seconds):

7-ship blind (saves ~25 seconds but requires an additional star medal somewhere else = ~10 seconds):

Total time saved: 16 seconds

I went ahead and submitted the run for you. If you want to add a comment or anything feel free :)

I won't respond to all the questions as I don't have time right now, but I'll share some of the thoughts I have at this point:

1. re-timing runs on other consoles: imo this should not be done. I don't know of any cases where people play english instead of japanese and then simply convert their time to what it would've been if they had faster text. The time difference in that case is likely much larger than in this game, but I see it as a similar situation. I personally played on Wii VC and got a 5:01 or something at which point I figured the console must be slower so I switched to NES. When I got the warpless wr (19:34/19:36 idk how accurate my timing was back then) that was also on the Wii but if that was still my pb I wouldn't feel comfortable converting it. I'm not familiar with emulator frame rates, though, so I can't give much feedback on that. I am, however, fully in support of converting a run if you can determine the time based on bowser patterns and whatnot.

2. This doesn't correlate to any #2 from the post so I hope it's not confusing, but as far as that claimed 4:57.58 by Kenneth Moser it looks tool assisted to me. The way he jumps at times doesn't look like natural human play. The way he bounces on the koopas in 8-2 and the way he handles the hammer bros. in 8-3 is very strange as well. Based on their patterns he clearly didn't start on the same frame rule as is common practice. Also, putting "Goal: 4:58.10" seems like a device to make the run more believable. 4:58.1X seems like a very strange time to shoot for for anyone that skilled, and you don't generally go for every frame rule but fast 4-2 when your goal is 4:58.10. I'd call it sketchy even if he streamed attempts prior to this submission. Are these fair points?

3. As far as being required to stream attempts goes, I agree with mav in that I feel that I play better offline. I've done offline attempts at times and sometimes gotten close to pb'ing (in warps and warpless). They weren't world record pace but I still have wondered how the submissions would be viewed. Sometimes I've opted to stream on an alt account so I could have a live recording. I feel that if you want to record all of your attempts that is a fair option if you're willing to deal with the hassle it may be. We may simply need a case by case basis. For example, if mav didn't stream and uploaded a sub 5 I would totally believe him, but if it was a 4:57 that would be more questionable.

I quickly wrote this up so hopefully it comes across with sound logic and doesn't seem abrasive.

I'm assuming there are no patterns between 4:57.01 and 4:57.39 where the podoboo gets in the way? Mav told me that between 4:56.92 and 4:56.99 (exclusive) the podoboo is not possible to pass. Do you have any information on this?

This is a great idea. There is half of the game that is completely untapped. Any% is hardly even beating the game when you think about it. It would give Luigi some air time as well. We shouldn't put it under misc. categories; In my opinion it is an even more legitimate category than warpless. Then we would have any%, Special Ending, and 100%.

I think Special Ending is the best name here. It is similar to "true ending," but fits well because of the special worlds. It should be a very fun category. But hey- let's not just talk about it, I want to see some people running it! This game needs more action! Hopefully I can get a new cap card soon so I can walk the walk.