Message xeroxfiend on twitch. It's an open tournament, every Saturday at 1pm cst. Although I won't be able to join this coming Saturday

Here's a gif of how to do the 8-4 wall jump.

The link posted above is great for learning about wall jumps in general.

Category "Walkathon" https://www.speedrun.com/smbce/full_game#Walkathon

sounds similar, except you can NEVER press B :)

You have to do a (I think) frame perfect wall jump on the last room in 8-4 to make it past the pit... which is what makes the category especially hard... since the hardest trick that you'll most likely fail is at the VERY end of the run.

Still cool TAS though :)

aah keeps that part constant

Why down + right? What advantage does that have over just landing on the second stair backwards then jump +right without ever pressing down?

Jonny use the pellson 1.7 rom and go to framerule 0539 for 8-2 and do the fast jump to get optimal shot - if that's what you meant

Red - It's what the gif file used to be for bbg. I didn't know you could do the backwards jump without pressing left first... So I pressed left then jump then right like the 4-2 jumps. Made it very inconsistent and sometimes fell in the pit

Have you ever done a large sample size of attempts to see your percentages? I've gotten anywhere from 16% (using an old bad method) to 27% over a sample of 100 attempts - curious what you get since you've clearly done lots of attempts and research

Forgot I was using Luigi and the Pellson rom - O.o ~~~ Oh well, I don't want to re-do it.

Luigi Love

Added a few things - I have learned a LOT since originally making this post, and some of the gifs reflect that - like the contentious 8-4 2 frame backwards wall jump, that is apparently not practical :)

First of all, I compiled all of the any% maps into one image file from Ian Albert's page (credit to Ian Albert from http://ian-albert.com/games/super_mario_bros_maps/)

You can go there to get the rest of the maps for warpless as well.

I deleted my previous "backwards" BBG that pressed left initially to instigate the backwards jump, which is unnecessary, and makes the jump more inconsistent, AND you can fall in the pit this way (thanks to RoyLGamer for showing me the new way to jump)

EVERYONE has an opinion on BBG setup, and I find that a lot of the opinions are different from each other. There are lots of ways to do BBG (just watch StuckInAPlate's BBG setup for evidence of this). So ultimately, it seems to me, that a method that is easy to do consistently the same is good to learn - which is what this is (at least for me)

Adding a new setup for BBG that I find to be the most consistent. I get this a little over 20% of the time after coinflip.

BBG with slowdowns

BBG 50% speed no slowdown

This setup is four parts:

1. Jump into the stair backwards (although you could probably jump forwards and then just turn to the "backwards" position)
2. Wait until the bullet is halfway past the second pillar (each time I've successfully gotten it, it's been 1 pixel past the halfway point, but halfway is still a good visual cue)
3. Jump without holding right, then hold right when mario's feet are one full block up and do a FULL jump
4. Run for roughly 3 frames on top of the middle pillar, then do another full jump

I have found some discrepancy on what works from this setup. For instance, I have run from anything between 2 and 4 frames on the top of the pillar and still had BBG work, I have also pressed right anywhere between flush with the top of the first block, and almost flush with the second block. However, following these 4 steps as best as possible has gotten me the highest percent success rate with BBG so far.

So it sounds like the primary issue is probably the speed at which the pipe was jumped over earlier in the level

But I've landed almost at the BACK of the Bill and kept the framerule, and sometimes I'm at the nose of the bill and lose it. (at least I think that's true - I've only recently been checking the rule I end up on in 8-3)

You BBG veterans will likely be able to answer this one. As a BBG newbie - it's confusing me.

I'm doing what appears to be the same setup, but sometimes I'm getting rule 335 on the pellson rom (the good framerule from the optimal shot) and sometimes rule 336 (the "bad" framerule from the optimal shot)

Any idea as to why this is? Is it just a luck thing? or am I somehow doing something that's affecting this outcome. Thanks :)

The included video is just one example of the two rules that I could find that were close to each other so you didn't have to watch a long video

The only thing I can figure - is that when going frame by frame, I took one additonal frame to jump in the second (bad) bbg... so I jumped when BB was slightly PAST the halfway point, and I think I ran on the pillar for an extra frame as well. That's most likely to answer to this question I guess... is that you have to be careful about when exactly you jump relative to BB's position past the pillar and how many frames you take on the top of the pillar before jumping...

Most people just split manually - I myself have a keyboard on the ground and I program the split hotbutton to be spacebar. So I use my foot to "split" every time I need to split. It works pretty well. Several other people buy actual foot pedals, but spacebar on a keyboard works just fine =) Welcome to the mario community.

The short answer to your question is, yes, the record in Speedrun is the World Record.

over time - speedrun.com has become the main place people go to submit their game speedruns.

Darbian has the current world record, as of October 2017, with a time of 4:56.528

SpeedDemosArchive still shows AndrewG as having the smb1 world record with a time of 4:58 (no milliseconds) from 2011 http://speeddemosarchive.com/Mario1.html

It looks like Twin Galaxies shows AndrewG as the world record holder with a time of 5:07 (no milliseconds) from 2015. http://www.twingalaxies.com/super-mario-bros/nintendo-entertainment-system

Twin Galaxies does not allow what they consider to be "glitches", so that's why the time is slower. "Glitches" have become most of the challenge and are, of course, allowed and encouraged on an any% run on speedrun.com

It's astonishing that neither of those sites use milliseconds. You absolutely need to use milliseconds on a 5 minute run like this to differentiate the HUGE # of runs that will fall under the same minute range. Even using milliseconds there are numerous runs that literally finish the run in the SAME millisecond.

There are almost 600 entries in Speedrun.com for SMB1 Any% alone. If you show obsoleted runs, it's probably more like 1500 runs. All were verified by an actual person watching a video (if the time is below a certain time threshold)

I've been doing offline blind practicing - I have the whole game sequenced... except one section in 8-2... 8-2 is the absolute worst... I even have that whole level done, except ENTERING the canon section... the only possible solution I can get I only get 30% of the time or so blindfolded... which is not acceptable for a run. Anyway here's what I have so far... all were IL done with the screen off... the ones that say "notes" - I was referencing notes... but was not looking at the screen. I will probably start blind runs of the game on Tuesday. Also 8-1 is pretty inconsistent, and it's also super long. The hardest thing about the blindfolded runs is the precise jumps where you only have a few frames to react... because mario does NOT react the same way every time you press a button. Sometimes he goes farther or not as far, sometimes he falls into a pit while walking, sometimes he doesn't. There is a lot of RNG even in Mario's movements which makes a blindfolded run difficult.

I started completely from scratch while sequencing the game - but there are lots of similarities in strategies between my runs and the only completed run.

I am using the music from the game and a metronome at 120 bpm to help with precise timing

Blindfolded 1-1 (339)

Blindfolded 1-2 (241)

Blindfolded 4-1 (with notes) (329)

Blindfolded 4-2 (246)

Blindfolded 8-1 (with notes) (54)

8-2 not done yet

Blindfolded 8-3 (with notes) (130)

Blindfolded 8-4 (with notes) (173)

for the actual runs I will be using an actual blindfold - and have a camera on my face and hands to decrease skepticism... for the IL runs I don't really care if anyone believes it was blind or not (it was) :)

Was a rough race this week... LOTS of game overs - We had the pleasure of having Roopert and Barnowl join us.

Race 1 1st - Barnowl - 16:35 2nd - Roopert - 17:52 3rd - Darpey - 19:12 RoyLGamer - DNF GTAce99 - DNF Xeroxfiend - DNF

Race 2 1st - Barnowl - 17:25 2nd - GTAce99 - 18:52 3rd - Xeroxfiend - 20:02 Darpey - DNF RoyLGamer - DNF

Today's the day - not running it, but EXCITED to watch :) GL to all who are competing

I got a run that looks like it should be a high 4:58 - but at least when comparing the videos side by side, the axe is removed at 4:59.174 again, but bowser drops like a 4:58

Has anyone ever seen mario HOVER? Skip to the end of the video and watch what anti-gravity mario does above the axe. He spends several frames suspended nonmoving taking his sweet time to grab that axe... meanwhile the game (other than the axe) finished on normal schedule, because the floor and chain are removed at a FASTER rate.

It's frame 17723 on spreadsheet, with -3.62 --- according to the spreadsheet it's 4:58.991 with 7 lost frames.

We're on the exact same framerule on 8-3 when we're off-sync... Watch the way the Bros act, they're both identical