Regarding My World Record Time
1 year ago
Massachusetts, USA

OK, so today it was brought to my attention that my old Super Mario Bros speedruns were timed incorrectly.

What does this have to do with Pro Wrestling? Well, the reason SMB1 was timed wrong was because it was done using a VCR. My Pro Wrestling was also done with a VCR.

As I've always suspected and as far as I'm aware VCR recordings add between 0.3-0.4 seconds per 5 minutes of recording. Pro Wrestling is about 10 minutes in real time so my actual time would be 10:02.

Why do I care? Well, Metroidmcfly had previously beaten my run in real time with a 10:02. The run is now removed, but taking into account the VCR issue the runs would be virtually identical in time. It can't be determined. The only reason SMB1 can be determined is because of Bowser's patterns, but with Pro Wrestling there isn't any way to know.

There's also the fact that the standard timing for this game was in-game time initially with Twin Galaxies and then with Speed Demos Archive.

My Pro Wrestling run also happens to be the oldest standing NES world record as far as I'm aware.

So I just have a lot to say about this even though it's a silly speedrun because technically my in-game time was never beaten and it's inconclusive whether it was ever beaten in real time either because of the VCR issue.

Conclusion is this: My time is definitely 10:02 and not 10:03 because of this issue. I feel like in-game time should be the standard until someone beats my time. I just wanted to mention all of this. You can do whatever you wish.

Ultra64 likes this
Massachusetts, USA

Things can remain as is. I do realize in-game time is also listed and that I'm being a little weird.

I just like accuracy. That's why I'm here. The real problem is that my time is likely 10:02 and not 10:03, but it's impossible to prove it.

Feel free to ignore the rest of this. I'm bored, curious, and I like math so I calculated times based on what I understand about VCR lag.

A VCR records and replays at a lower FPS than an NES which is why this discrepancy exists and would exist for any VCR recording video games.

Normally a VCR records at 59.94FPS, but It's my understanding that a VCR can record at a maximum of 60FPS under normal conditions, but can record more if the VCR happens to be able (basically overclocking the VCR). The NES runs at 60.0988139FPS. The video then plays back at the standard 59.94FPS, so the minimum amount of lag when recording a video game with a VCR should be exactly 99.9% (59.94/60) of the actual game time.

So in the case of my Pro Wrestling run the minimum amount of lag would be: Time = 10:03.233 Time = 603.233 seconds 603.233*(59.94/60) = 10:02.629

The maximum amount of lag would result in: Time = 10:03.233 Time = 603.233 seconds 603.233*(59.94/60.0988139) = 10:01.638

Metroidmcfly's (former SRC) time is 10:02.467.

Ultra64 likes this
Massachusetts, USA

What's also funny is that if you apply the same lag issue from my SMB1 run to the Pro Wrestling run, then I think my time would be 10:02.47 if I'm calculating it right. That would mean our runs could be the exact equal length. The only difference being that my run's in-game time is 6:00, where his is 6:01.

Ultra64 likes this

I agree with your maximum amount of lag math if that is actually what is happening. NTSC NES and SNES have the same framerate so this is a wider issue.

How do you know the VHS doesn't just sample the ~60.10 Hz footage at 59.94 Hz or 60.00 Hz for no time difference? Like what capture cards do that record in digital video. VHS was very substandard Composite video. Nicer VHS recorders used time-based correction circuitry.

I think someone should record 10 minutes of VHS footage of the game at the title screen. Any game. Can passively split Composite video + audio with cheap Y cables. Record the other line with capture card or simultaneously record the television with cell phone camera. Close enough for audio. Then compare the video and audio at the end.

By the way, a direct comparison of AndrewG's video and Metroidmcfly's 2018 was linked on Meester_Tweester's Reddit thread and Metroidmcfly's is faster as stated. Runs are removed by runner request or moderator action if the runner is banned or makes the video private.

I'm curious if the mods in hyper-competitive Super Mario Bros undid people's VCR lag penalty. Which I'd like to see proven. I don't think saying a clearly slower run in real time is actually faster is in the spirit of speedrunning but I don't play competitive games. I think it's a conflict of interest that AndrewG is a mod and would have a record of longest unbeaten run if Metroidmcfly's run is mathmagically slower. I'd prefer real time with priority over in-game time, given advancements in recording technology. I don't play this game though.

I found an ancient (2012) SDA thread about obsolete but powerful VirtualDub capture card software having settings to let it record incorrectly at 59.94 Hz for NES and SNES versus dropping frames. Same timing issue presented here. The poster is wrong about Super Game Boy 2 framerate, but 2012, can give a pass. There could be lots of games recorded with added time but most games aren't competitive to the second.

If we want to be exact, or pedantic, there's a mistake here and in the thread:

NES: 21477272.72 / ((2623412-1)*2) ~= 60.09881387708959) ~= 60.09881387708959 SNES: 21477272.72 / (261 * 1364 + (1360 + 1364) / 2) ~= 60.09881387708959

The crystals themselves used to generate the master clock are 21.47727 MHz ±30ppm and can drift 5pmm in the first year. My SD2SNES flash cart displays the master clock frequency and it's slightly different on every bootup. In other words, using 9-18 significant figures is fake accuracy.

21.47727 MHz ±30ppm is 21.47727 * (1 ± 30/1000000) = [21.4779143181, 21.4766256819] MHz. Given the crystal's printed 7 significant figures, we have [21477910, 21476630] Hz with rounding. Using either equivalent formula, NTSC NES and SNES are [60.09702, 60.10060] Hz and not 60.09881387708959 Hz or else with some arbitrarily greater precision. Some people's master clock drift is so great that they lose color in Composite and S-Video. As in, outside that range.

Then max lag of 603.233*(59.94/[60.10060, 60.09702]) gives [601.6210, 601.6569] seconds for 10:01.621 to 10:01.656 possible real NES time for a VCR recording of 10:03.233. Timing to whole frames, 0.621 to 0.656 is a 2 frame spread. 1 frame of uncertainly for every 280 seconds. Close enough to 5 minutes. Assuming VCR lag is proven.

Massachusetts, USA

I'm totally fine with metroidmcfly having the better real time as it's the only thing we can determine for sure, but the VCR thing definitely exists as seen with SMB1 becausse the timing is made by using Bowser's patterns and knowing the exact frame counts that way. i

In-game time was the standard to which I did my run is the more important thing. Until someone beats me to the standard in which I did it I don't think you can technically say it's been beaten.

I am also not a mod and if I were I would possibly change it to IGT for ranking and not vice versa for this reason. Or once it was beaten I would revert it because real-time is the more accurate measure. That's to say, yes it makes more sense, but that's not what I was aiming for.

Another thing to note on that topic is that I am fairly sure I got a number of 6:01s and 6:02s in the process of getting this run. They could have been faster in real time, but that wasn't the goal! I was aiming for 6:00 flat or lower. Otherwise I would have tried to precisely time it this way.

https://speeddemosarchive.com/ProWrestling.html

https://www.twingalaxies.com/game/pro-wrestling/nintendo-entertainment-system

Edited by the author 26 days ago

I'm sorry I thought you were a mod. I don't know why I did. There's no conflict of interest. You have an interest for holding the oldest unbeaten record since you posted on Twitter about it. I think anyone would.

You have the in-game time record. That's fair. Never been tied or beaten. I see Twin Galaxies uses it and Goldeneye and rest of games here moved to RTA. That's also fair that you probably did get a faster RTA run with slower IGT that you immediately taped over.

The counterpoint is Metroidmcfly stopped playing when he beat your RTA time. If the time had been adjusted or IGT had been the standard, he would have tried to beat that instead. He was #1 on this leaderboard at least in 2017, which the run history doesn't show ever existed. I guess because he deleted his account. Pic at the bottom to save a click.

I messaged mods with the comparison video asking to add the run back. Mods can submit runs on behalf of others. Now if Metroidmcfly asked to have runs removed, that's different. The mod then, RantronBomb, hasn't logged in for over 4 years to ask. You're fine with him having the faster RTA anyway.

I think the VCR difference is more interesting:

Adjusting for VCR delay, I can't find that done in a single game. I can believe it was done in SMB1, maybe the most competitive game in existence, which times to the frame with Bowser's pattern as you mentioned. I can barely even find a discussion. This Twin Galaxies thread from 2004 is interesting:

  • Users rdrunner, The_Pro and sdkess NES show a 6 second delay per hour from VCR recording with a total of 7 different VCRs. sdkess found the same thing with SNES. This can be calculated with 60.10/60.00 * 3600. As in, all 7 VCRs recorded at 60 Hz. Recording at 59.94 Hz would be almost a 10 second delay.
  • The_Pro and rdrunner showed PS1 has an advantage recording with VCR. PS1 is 59.94 Hz, which would make a VCR recording faster than real console if the VCR tapes at 60 Hz or faster.
  • LucidFaia stated their Super Mario World time of 11:21 is the same down to a tenth of a second with their VCR. We know NES and SNES run at the same ~60.1 Hz.
  • MKM stated their VCR is dead on accurate with stopwatch timing. Consoles not stated. VCR is a Toshiba VCR from the late 90s.
  • One person found GameCube on VCR to be the same speed as the live play, another slower like NES, another faster. Not sure how to interpret that when it's a 60.0 Hz console.

I'm curious if you had a 59.94 Hz VCR when they all seem to be 60 Hz, or even contained time-based correction circuitry (TBC) that correctly handled NTSC NES/SNES.