For a while, I've been retiming each segment of the best Warpless No OoB runs, to compare how long each runner spends on each split. Each time the lives screen appears when entering a new area, I count that as a split, and the run starts and ends on the same frame as it would do in normal retiming.

The best times have been pulled from a mix of my spliced Warpless run, my L4 WR, and Troyish's L6, to conveniently make a sum of best of 3:06 exactly. A green cell means that split is the fastest out of everyone else's full game PBs, and a deeper red shows how behind the runner's split is compared to everyone else's. For the most part, the greens are closer to the top, and there are deeper reds towards the bottom.

A couple anomalies here would be KesN - his level 1 is the 2nd best, and his level 2.2 and 4.1 are much faster than everyone around him, but because he loses over 5 seconds in level 3, he is only 10th (excluding Fro, who deleted his runs) while improving just his level 3 is enough to get him into top 6. EDIT: this isn't true anymore because idkjake just got a top 3 run a couple minutes ago, but I'll keep it here regardless.

Here are some pie charts (and analysis) of where each top 8 runner is losing time compared to a perfect 3:06.00:

1st: @SlushPuppy There isn't very much to say here, because I already do all of the strategies which are done in the SoB. Interestingly, my biggest time losses are in 2.3, due to bonking while trying to get to the flagpole, and levels 1 and 5, because of suboptimal acceleration boosts. This definitely shows that, at the top level, bonking will cost a huge amount of time, and acceleration boosts will make a big difference to your final time.

2nd: @Siver Siver loses a little bit of time in every split, but his main demons were level 1 and level 2. In level 1 he most likely lost time because of a bad acceleration boost (though solid accel boosts in 5 and 6 mostly make up for this). In level 2.2, his time losses are from slowing down to make the first section (going under the ? blocks) easier and a small bonk after getting past the ? orb. Then, in 2.3 he bonks on the staircase and the block under the flagpole, costing a whole 0.833 seconds. That level really isn't forgiving...

3rd: @idkjake I believe this runner does every trick except for acceleration boosts, explaining why almost half of his time is lost in L1, L5, and L6. The rest of his time losses are a bonk on the stairs of 2.3, and a couple small time losses all around, which will certainly go down over time as his movement improves even more.

4th: @IcySkid Level 2 is clearly Icy's strong suit - she's only about 0.5 seconds away from an SoB-level L2. Her main downfalls are a bonk near the end of 4.1 and a nonoptimal acceleration boost + a bonk in level 5. She's doing every advanced technique perfectly, so all she needs to improve further is better movement. This is about what I expected going into it - as she's a newer runner, she's put practice into the hard techniques, but hasn't (yet) gotten the movement that newer players have.

5th: @xXGaming_FroXx This run was never officially submitted (and the rest of his runs were deleted) but his PB can be found here: He does every advanced technique except for the fast L3 ending where you jump between the shell and the flagpole so you can touch the flagpole faster without dying, which explains a quarter of his time losses. The rest of his time losses are just down to unpolished movement, especially in level 1 where he is over a second behind the WR. This is more evidence that the new-school runners are better at the advanced techniques, but haven't had the time to develop their movement yet.

6th: @okot / @Tokonata Here is the oldest run on the list so far. The run was done back when we didn't know much about acceleration boosting, so despite doing well on levels 1, 5, and 6, he could have been a combined 2 seconds faster by perfecting those levels with acceleration boosts. His level 2 is really good, as he did every advanced technique in that level, but almost half of his time losses came from L3 and L4.1. His L3 didn't do the shell jump at the end of L3, and slowed down a lot in several places, especially at the start where he hit the ? block then kept going, so he had to turn left THEN turn right to start moving again. In L4.1, he does 2nd firebar skip much later than he could have. He also tended to slow down in the harder parts of that level, like after the first firebar skip where you have to jump right without hitting the hidden block. Some of the time loss likely would have came from not doing fast pipe too - it was discovered after this run that falling into the pipe instead of entering it while you're already on top of it was a little faster.

7th: @SirTroyish Since Troy is an Any% runner, it's no surprise that his acceleration boosts are really good (as you can tell from his level 6 - I've never seen a level 6 faster than his). Other than a bonk in 2.3 (which a lot of runners get messed up by constantly), his only notable time losses are in L3 and L4. In level 3, he loses most of his time from holding back too much (and also not having shell jump). I'd guess that he's less confident with level 3 because he hasn't done as much warpless compared to any%. His level 4 is mostly fine, there were just two small mistakes in execution: he messed up his lion jump and missed fast flag. I think all he would need to be at around 3:10 is work on his level 3 then get a lucky run where he nails everything.

8th: @Merlyn Most of these splits are really solid, even the tougher levels like 3 and 4.1, but there are 3 that stand out as having potential for time saves, and all 3 have a similar cause. L2.3 had a staircase bonk and missed fast flag, L4.2 had another staircase bonk and another missed fast flag, and the bulk of the time loss was in L5, where there was a slow flag (this is when you miss fast flag and instead of travelling faster than normal, you travel slower than normal.) His movement all around is great, so I think he's 100% got potential to be sub-3:10 very soon even though he's 8th.

BONUS: @f1 f1's run is very interesting, because it was done back in early 2016. This was before a lot of the major time saves were discovered. This means around 5-6 seconds are lost due to no fast flag alone, and it's the reason his L1 is notably slower than the top 8 runners. If you ignore time losses from not having fast flag, a lot of his splits are actually really good, with the only big losses being a lack of any of the advanced techniques used in L3 and L4 by people who have run the game more recently.

BONUS: @Gohgo Since this is a 3:52, the results are going to be very different here. Half of his time loss is in L3, and there's an obvious reason why: there's a death at the end of the level, and although he respawned at the last checkpoint, getting to that checkpoint cost him a lot of time too. Another quarter of his time losses are in level 4. He does the first firebar skip, and then after the checkpoint, he goes over the firebar (instead of going for lion jump, which is still much faster than this even if you miss the setup). Although his other levels could all be faster and he could cut down his time with some intermediate techniques like fast flag, the main thing to take away from this is that, for the average runner, cutting down on deaths and optimising your routing are by far the most effective way to improve.

I hope everyone liked this post! If there's any other cat mario stats you want to know about, let me know :)

This video is made as a demonstration of how incredibly far warpless can be pushed.

All splices happen at lives screens. All segments were taken from my own runs.

This could go faster in a lot of places, there's many slight optimisations that can be made regarding general movement throughout. However, there's two main places where time could be saved, both in level 2.2.

There is the "toko jump", where you hit the green ? circle and jump over, under, or between the circles of fire that spew out. This is mostly reliant on RNG and it is very rare that the fire will line up well enough for the jump to be possible at all, so I doubt that it will be seen in runs any time soon. There is also the "instant pipe entry glitch", as seen here: https://www.speedrun.com/cm/thread/ndekm. The problem is, doing it in one fluid motion without losing any time is very difficult compared to the amount of time that it saves, so this also might not be seen in runs for a while.

I'm not certain how much more time can be saved, but I feel like nowadays sub-3 isn't out of the question at all.

This run https://www.speedrun.com/cm/run/y6l0x06y by fatcat006lol was accepted into Warpless No OoB when at 2:16 he goes out of bounds? If the moderators don't have time to watch through videos before verifying, I'd gladly take the position of moderator to help with this.

My run was submitted 2 weeks ago and it still hasn't been accepted, and the moderator hasn't been online for over 2 months, so I'd assume that there are no active moderators. I'd gladly fill this role to make sure that there are no huge delays before runs are accepted in the future.

This may be a tied world record - although I hit the flagpole first, siver starts walking after hitting the flagpole first, and we both reached the level select screen at the same time. The most likely explanation is that we got the same time.

"But Slush, the cat appears on your side first, so you must have started earlier!" Keep watching the first second of gameplay and me and siver will alternate between being behind or ahead. This is due to the fact that my video was at 60fps while his is at 30fps, all on a 40fps game.

This is a flash game, played via the web, so it should be marked as a web game.

The leaderboard has a few active-ish players (more now that ryanorfeed is considering getting back into the game). So, maybe a discord server would be a good place for runners to gather and discuss the game together.

There will be a minor change to the timing rules for this game.

Previously, runs were timed such that the timer started on the screen beginning to fade after clicking a level, and ended on the last level on the frame the victory animation begins.

However, this has many problems - it's hard to find the exact frame the timer starts to fade due to video compression, and sometimes finding the last frame can be impossible as standing next to a bomb while it explodes makes the player invisible.

As a result, the new timing method will start on the first frame the first loading screen is visible will end on the first frame the Game Complete or Level Complete screen is visible and then a modifier of -31/30 (-1.033s) will be added to the result.

What does this mean for me? Not much at all - runners will continue to time the same way as always, this new rule just affects how verifiers will retime runs.

Does this mean I don't have to start IL runs from the level select screen? Yes - runners can now start from either the previous level or from the level select screen.

How are runs retimed? Moderators are using this software https://github.com/Slush0Puppy/retime/releases to retime runs precisely.

Will this affect existing runs? Yes - all existing, non-obsolete runs have been retimed with the new method. This hasn't affected any places on the leaderboards, as any differences will be by one or two frames at most.

For some reason, when lined up correctly, it's possible to enter a sideways pipe from above, skipping the pipe entrance animation. By corner clipping against the ceiling, the cat is pushed into the exact correct position, where he can fall down into the pipe. This likely won't be done in RTA as it seems to only save around a quarter of a second and it's very easy to fail, but it still brings us closer to a potential sub-3 warpless.

I've noticed that it's very time consuming to perfectly retime a speedrun: you have to download the video, open it in a video editing software, then manually calculate the difference between the first and last frame.

That's why I decided to make a program that does the hard work for you, so all you need to do is paste in information from the right click menu of YouTube's video player.

Here it is: https://github.com/Slush0Puppy/retime All of the instructions are included in the readme file.

Please leave any feedback so I can continue to improve this project.

In theory, someone could submit a run like this: According to the rules, this would be a 23.767 as the player touches the flagpole at 23.767 seconds, but since the player dies, the level isn't really complete at all.

Toko's runs don't start on level 1. They shouldn't be allowed. Check the discord, everyone's agreeing

lol

At around 00:30 of the no spacebar WR, the screen fades to blue and then back. What's happening here?

In this video I show that you can collect both the red coins star and blue coins star in one go. This frame-perfect trick saves just under 40 seconds, bringing the best possible time down to just over 7:30. EDIT: I just found a new timesave that saves about 12 seconds. You do 100+Koopa+King as normal, but you don't manipulate King Bob-Omb's position, instead you go straight for Koopa the Quick. The moment King Bob-Omb explodes, you click to start the race. The stars will be separated, but you can use a glitch similar to this in order to get both stars anyway. Sub-7:20 hype???

Here I collect 'To the Top of the Fortress' and 'Blast Away the Wall' in one go.

The way this works is that if the player jumps forwards just before hitting the star, Mario will keep jumping forwards during the star animation. You can use this to your advantage and collect another star close by. There are many similar ways to do this, such as with a long jump or double jump. This is very precise - however, since there is minimal setup required and very little penalty for messing up, it is recommended that future RTA runs at least attempt this trick.

This reduces the best possible time to sub-8:30.

https://discord.gg/fynwjdt Come here to discuss strategies, easily contact users/mods, or just chat about Nitrome games!

I believe Feed Me would be an interesting game for speedrunning - and I'm surprised there isn't already a leaderboard for it in the Nitrome series. Will it ever be added?

Example run: