Grew up with the All-Stars version, have heard that SMB1 plays MUCH better on NES re speedrunning - is there a similar discrepancy in SMB2 or any other significant differences to be aware of?

Just to mention that in the course of adding new runs from Nicovideo and other sources I am coming across some anomalies so apologies for any agitation of the leaderboard while this is ongoing.

I have now resolved the status of previously posted Tokumi2008 runs. One of them is duplicated on a nicovideo channel which duplicates other runs in other courses. Based on the volume of good runs I have found, and research on a mame forum, I am inclined to believe Tokumi (deceased) is the rightful owner of those two times his youtube claims. I have also found another youtube channel which duplicated existing runs without giving credit - no runs will be submitted from that channel nor from any place which does similarly.

If you have any questions or comments don't hesitate to ask. At the very least this will give new runners far more meat of a ladder to tackle in terms of climbing their way up as well as providing more useful information in terms of how top players tackle different corners.

I am adding runs I find on Youtube which meet the board requirements, both here and in SHO Arcade. In light of the fact that a great majority of these in Genesis especially are very casual, I feel like this is a nice way to showcase skill from all sorts of parts of the gaming community and the extra places on the board will help new runners to measure their progress.

In case anyone reading is able and willing I would very much appreciate any help possible in researching runs from Japanese communities. I have found a few runs but it may take me time to work out what is what regarding settings and validity, but similar to Tetris, SHO has a good community particularly in the MAME version in Japan and it would be awesome to find any hardcore Genesis runs to add to our number. If anyone is skilled or able to understand Japanese better then please feel free to link me to anything you find of note.

This is a post to request outside verification of a history of runs which has been questioned re legitimacy. I would welcome guidance from site staff but it would seem apt to have this done by a moderator of one or several Genesis games who is adept at understanding emulated runs including issues such as lags and is also capable of investigating splices, or TAS use.

I am delighted that we had a few really strong new runners join our ranks so far this year! I would like to grow our little game even more since our output seems to be quite seasonal and comes in spurts. I was wondering if anyone wants to do some races, or if anyone is interested in creating extra categories.

For races I would think the best method would be first to complete a course N times, which makes for far easier streaming side by side and avoids our problem of IGT bugs. We could also then harmonise emulator use and get them broadcast on SRL, and it would also open us to being able to enter marathon events in a consistent way. I would love to get SHO featured in a marathon sometime and I feel like a race would be slightly more amenable than a showcase.

Is anyone interested in a marathon category or has any other ideas they'd like? Something like 'N in a row' for each race would be a nice way to measure skill in a more balanced way without traffic/single accidents hampering strong runners' top times.

Timing is defined as starting on the frame you press (1) on the title screen (after you collected CHEA and soft-reset, unless you're some sort of masochist) and ending with the final hit to Maybe.

Take your framecount and use f/60.

When there are frame drops or ambiguities about that in runs, we can reference that point as: -Screen fades from f+3 -Screen goes black at f+11 -Pond appears at f+69

With C.H.E.A.T having become near perfectly optimised, it makes sense that we should adopt an endpoint which can be perfectly defined - it has been difficult in the past to be sure when a final hit was delivered because flashing animations have often stopped and started very quickly, and not all runs showed a frame count.

I would very much appreciate future runs being submitted with 60fps recording, and showing a frame counter. I may be able to interpolate the correct time for a 30fps recording but I cannot guarantee it and I would prefer never to have to deny someone a WR by one absent frame.

The intended 'stop' point of runs already on the leaderboard corresponds to the frame after the 50,000 point bonus is added on. This makes a very easy reference point for future runs and allows absolutely no ambiguity. With that in mind I have retimed the top runs so that they all end at this point. If there are any further points arising from this please do discuss them!

Since the Original Mode WR has gotten so optimised I have added milliseconds on it, in anticipation that future runs will break the record by less than a second barrier. If two runners tie WR to the same second, the runs will both hold first place regardless, since grunt hit luck and speedboat RNG are highly variable.

(Edited) I've started using ms for Arcade Mode runs beneath 10:10.

I am hoping to do a gaming marathon for Macmillan Cancer Support, Monday 17th, 9pm GMT to 3am to try and get UK+US viewers, on Twitch. If successful I would be glad to do more of these with different consoles or themes of stuff to play. Plan is to play various SMS/Genesis games for entertainment and speedruns. I would be super grateful if anyone interested could please spread the word, my Twitch channel is tiny!! 😂 Also feel free to send me challenges/advice/requests! I've never done anything like this before so I'm starting with a goal of £50 donated. Any help you can give to make sure this goes well would be superb. https://gameheroes.macmillan.org.uk/fundraising/petaQ

Hi, posting here in case it was missed in the series forum - thinking of learning EWJ 1 and 2 in future, got good experience in platformers and reasonably technical things, which would you recommend I pick up first?

I've got mostly the hang of everything but is there anything particular we can do in Good Princess Castle with little movements to clip through the floor/ceiling early, or is it down to luck?

Oh and can anyone explain in some better detail the rng manipulation in Lake? I must be losing a few seconds there.

Thinking of picking up the EWJ games on Genesis in future, I have some experience with platformers and pretty precise stuff. Should I look at Jim 1 or 2 first?

I was wondering if anyone else of the top runners was interested in doing a 150% speed speedrun - playing at 90fps. I fiddled around with it in the past and I feel like it's manageable. I put some work into it and churned out a 3'14 in Africa which I was quite pleased with. If anyone wants this as a category even just for giggles then let me know.

I'm having typical Windows user gripes with my machine causing lagging making speedrunning difficult, and I was thinking seriously of investing in physical equipment to do my runs on. Some input from SMS/SMD runners with expertise in the accuracy comparisons of region mods v native consoles would be greatly appreciated. I am told that an NTSC console would be more accurate, but do not know just how much the difference is. The games I run have very small communities so I should prefer to be maximally accurate as far as timings go since a good investment will pay for itself. And if I go ahead with one or other of these things, what else would I need to handle recording and hookup to a modern tv? Thankyou for reading.

CoKami and I had a best of 3 race on Friday, and we plan to do it again so if you're interested give us a follow on Twitch/Twitter to hop in on the fun! I personally would be happy to race against anyone and Zeroth, we are both now sat tantalisingly at 14:10 - wondering if you'd like to race me at some point! If anyone wants to do this just shoot me a message somewhere. Heck even if I have to get buttstomped by btrim or thiaguin I'd value the opportunity!

This is possibly a very stupid question but I'd rather ask and be sure. Based on the timing rules currently shown I appear to have a run without pauses that comes to 9.84, and as an example, if I frame advance 16bitpanda's run it appears to end earlier than his notes state which would corroborate that. Is there some software I can use that will retain 50fps recording as opposed to 30 with vlc? I by no means doubt what has been said about the timings but I can't explain the difference. I can try to replicate it with inputs showing but as I speak any attempt to record that is killing the 50fps on bizhawk. I hope I'm not missing something very obvious.

Frame 770 for the start and frame 1262 for the blue screen. I imagine one of you who mods can sync this up with a 9.84 if need be.

Just wondering if anyone had a similar experience, and can give any advice, not necessarily from speedrunning. A month ago I was running and a tiny break in the side of my thumbnail grew out and turned out to be the beginning of a separation of the nail from the bed, so I had to mostly stop gaming for a while. I was recommended to get a cushioned thumb cover. I can't be certain the problem came from thumb on d-pad, but I have noticed that my thumb stays nail-up when moving left and then rolls over when I have to switch to right, which I doubt would be fixable really. Has anyone got any experience of buying such covers, can they recommend a particular product?

They both seem entirely legit. I attempted to contact someone who knows the player to get permission to add the runs to the site, and to find out whether they were done on an actual bike, or via MAME. Site rules say that a top time can generally be listed without permission. Does anyone have any objection or point of order to raise about the validity of these runs besides the method?

Did anyone who has a time on the leaderboard do it with an actual bike? I would greatly value any feedback on whether that makes it a bit easier to react on such a fine level as you see above without shintysix squillion resets.

To streamline the board and avoid confusion I have created a Variable to show the romset/track layout that runs were done on. Sitdown/Upright is the default, to view Limited Edition/Mini Ride-On you can use the Filter. By doing this we can group all runs across arcade and modern console conversions in the same place without overcomplication. Please do not worry about selecting the right option in your submissions, I will always double check everything.

Limited Edition/Mini Ride-On is the same track layouts that are present on the Genesis version. LE is generally considered the easier option as it has a longer time limit and tends to have less traffic of AI riders. Sitdown/upright has a more challenging beginner's course and tends to be more ubiquitous and popular. The console ports for PS, Xbox, and Wii all have the option to use either set of track layouts, and you are not restricted in what you choose! The Asia course is the same in all layouts, personally I think it's the best one as it's short enough that you can push hard without having to play a long tactical game or avoid risk. Or am I just saying that cos I stink at weaving in traffic 😂

On timing: Super Hang-On has a bug in the In-Game timer, present in: -Genesis (Rev 00 and 01) -Arcade (all versions) -PS3 -Xbox -WiiVC

across both real and emulated runs.

The second part shows a comparison on the Arcade version with a slower sector 1 against a faster. Both runs follow precisely the same path with speed locked, yet end up with different sector 2 times. I have also verified this bug by syncing playback of different runs' sectors where IGT says run A will reach checkpoint first, but the real time and visual evidence show run B reached it first.

Here are a couple of quick clips of the PS3 and Xbox versions. The timer goes from 31.9 to 32.88 in one frame.

The timer goes from 1:26.29 to 1:26.55 in one frame.

When a checkpoint is passed, the clock is zeroed and that time is written to memory. But then the timer starts counting from whatever fraction of a second was left over.

Due to this, all runs' times have been corrected. It's very simple to work that out: just remove the fractions of a second from each sector time that shows up, except the last sector of a race which doesn't show anyways. Add these up, and subtract them from the IGT. Then you have your True Time. I will always double check these. You may in future wish to complete any run which is close to your PB or a WR: typically an IGT is off by 2-3 seconds in Africa, increasing to 7-8 seconds in Europe.

Here's an example of how to do this from a race in Africa:

S1 35.93 S2 33.1 S3 34.8 S4 31.93 S5 35.8 (S6 time not shown on screen)

IGT showed 3'29"59. 0.93 + 0.1 + 0.8 +0.93 +0.8 = 3.56 3:29.59 - 3.56 = 3:26.03

Thanks very much for reading!

Super Hang-On has a bug in the In-Game timer, present in both Rev 00 and Rev 01, in both console and emulator.

The first part of the video shows a playthrough of identical runs at 265 km/h (Left) and 263 km/h (Right). Notice the sector times look alright at first, but then start to make no sense, as the slower run gets faster splits and a faster IGT overall.

When a checkpoint is passed, the clock is zeroed and that time is written to memory. But then the timer starts counting from whatever fraction of a second was left over. For example, sectors in the first 265km/h run start from 00.80 each time, accounting for the mess.

I have now edited all runs current and obsolete on the leaderboard to show the true times. Only a couple of minor places have changed hands and all times are now totally accurate. One item of note is that Mc166's Expert run was never actually the WR, but since it spurred Dev to pull out a bonkers fast time, I don't think he will mind ;) Also let me just congratulate Seymour whose Beginner run was 3.5 seconds faster than it appeared, giving him a corrected and well-deserved 8th place!

So, how do you work out your actual time from a race? It's not so hard.

Write down the fractions of a second that show when you passed each sector, except the last sector of a race which doesn't show anyways. Add these up, and subtract them from the IGT. Then you have your True Time. I will always double check these. You may in future wish to complete any run which is close to your PB or a WR: typically an IGT is off by 2-3 seconds in Africa, increasing to 7-8 seconds in Europe.

Here's an example of how to do this from a race in Africa:

S1 35.93 S2 33.1 S3 34.8 S4 31.93 S5 35.8 (S6 time not shown on screen)

IGT showed 3'29"59. 0.93 + 0.1 + 0.8 +0.93 +0.8 = 3.56 3:29.59 - 3.56 = 3:26.03

Thanks for reading!

I did my last any% run with a mega drive controller, wanted to check whether runs need to be done like this as opposed to mapping controls to a keyboard. Don't know whether all leaderboard runs were controller and I'm not pushing for keyboard to be allowed - I feel like it'd be an unfair advantage actually since I would be way faster on it. Thanks