Seems that Delta takes around 0.2s longer to run out of time in 1-1 than my AV Famicom:

Going to guess this runs at 60 flat or something comparable? I'm sure someone could perform more exhaustive research on this topic. Can also confirm that opposing D-Pad directions are disallowed, at least with touch controls.

Delta's NES core seems to be powered by Nestopia UE v1.48, for what it's worth. That should meet this game's accuracy threshold, although as Roopert expressed, many other factors would need to be considered.

And did this require a forum post?

Considering all the abandoned projects this community has put out, I wouldn't get your hopes up.

The former, visual artifacts in the UI are merely a means to determine the frame that the game has crashed.

Private submission to a moderator/verifier is fine as long as they can corroborate that the full session was received. Our objective in requiring full session isn't to require the footage be displayed publicly, we understand that for various reasons a user may not want to have that footage viewable by the public eye.

Yes, Pellsson's practice hack uses mapper 1 (MMC1), which is supported by KrzysioCart.

We once had individual levels on the leaderboards, but they were timed using IGT and it just didn't work. I don't think individual worlds would be much better unfortunately, even if more appropriate timing conventions were used. Unless there's substantial interest in such an idea, we have no intention on adding individual worlds to the leaderboards.

Hey everyone, just wanted to let you all know that there is an 11-frame framerule in this game that dictates various long timers. I haven't seen this documented anywhere so I am assuming this wasn't previously known to the speedrunning community. What this means for the speedrun is that the game can only proceed to the next stage in intervals of 11 frames. Therefore, the barrel stage and the elevator stage can be played imperfectly and still take just as long as a perfect stage. Since timing stops when the final rivet has been removed from the rivet stage, every frame matters there for 1 Loop speedruns. However, it's worth noting that the cutscene of Donkey Kong falling after removing all rivets depends on the framerule, so only on the final rivet stage completed does each frame matter.

Zero page address $34 is the 11-frame framerule that controls all long timers. This address is decremented every frame and when it wraps around from $00 to $FF, it is forced back to $0A (10) and all long timers are decremented. The long timer that controls stage completion is zero page address $43. It's set to $8D once you reach Pauline and transitions to the next stage once the timer reaches $7A. With this knowledge, it's recommended to split when the black screen appears after completing a stage rather than the moment Mario reaches Pauline.

To give runners an idea of how quickly a level must be played to match the optimal framerule, I created a quick patch for the Revision 1 ROM that displays the numbers of frames remaining in the framerule when reaching Pauline; you can grab it from the Resources tab. Hope this thread proves useful for speedrunners of this game.

ChromeOS is based on Linux, which is supported by some NES emulators. I no longer possess a Chromebook so I have no means of testing this myself, but if you are able to run Linux applications on your Chromebook, then you should be able to run the Linux builds of NES emulators on your device.

@joeth3br0 Yes, thanks for catching that, late night brain. It's fixed now.

Yeah, that's a bug present in the original NTSC release of the game, not an issue with the practice ROM. This occurs because springs are allowed to load into the sixth enemy object slot (usually reserved for power-ups and the flag atop the flagpole), so the spring in 2-1 will get overwritten by the flagpole's flag if it loads into this slot. It was fixed in the PAL version, SMB2J, and the All-Stars re-release.

@ArcadeZeus1 To give people time to learn about the ban from either the News tab or the Discord server before enforcing it. Banning an emulator is a big deal and we wanted to give a two week grace period to allow any FCEUX runners to get accustomed to an approved emulator.

@KilleDragon just use a different emulator man

I think we should make a category called "stop necroing this thread"

1. You are allowed to continue a run from a game over; any number of game overs are permitted as long as the speedrun is completed.

2. I do not believe an autosplitter has been devised specifically for this game. I do not have knowledge about how the autosplitter works, but this game shares many RAM addresses with SMB1, so there's a possibility the autosplitter will work with SMB2J as well. If not, then you'll have to split manually.

Speed Mario Bros. runs go on the Ultimate NES Remix leaderboard: https://www.speedrun.com/ultimate_nes_remix_3ds/leaderboards?h=Speed_Mario_Bros.&x=wdm489od

It wouldn't make sense to allow a 1.5x speed version of SMB1 here, official or not.

In the Japanese version, there's a glitch that lets you instantly warp to the right side of the flagpole when grabbing the flagpole in certain stages, which saves time due to Mario not needing to turn around when getting off the flagpole. It's generally called backwards flagpole because Mario gets on the other side of the flagpole early. The reason this happens is a bit technical (a decent explanation can be found here: https://www.speedrun.com/smb1/thread/xsghm ) but hitting a block containing a power-up can enable this glitch in certain stages, while in others you will automatically be set up for the glitch.

Also, if there was any confusion, backwards flagpole in SMB1 is exclusive to the Japanese version of All-Stars, but there's an unrelated backwards flagpole trick possible in The Lost Levels which can be done on any version of All-Stars.

A year late, but we've added Both Endings and All Stages as categories to maintain parity with SMB2J's leaderboard. We will add your run to the All Stages leaderboard.