This guy claimed to have done runs for 6 days prior to the run, presumably practicing for roughly a month considering the time differences between his supposed WR, first uploaded IL and Naegleria's Sonic 1 tutorial. Whilst doing so he claimed to have beaten the WR by almost a full minute from a guy with hundreds of IL-videos and the like not to mention years of experience. With strong claims like that there obviously comes a big demand for proof.

The video in question is an upscaled 320*224 video and he claims that the source recording was split into parts. This completely excludes Fusion as his choice in emulator unless he tampered with the aspect ratio whilst recording in some weird 3rd party program in order to look more suspicious which of course would be ludicrous to assume. Considering that Gens splits recordings into 2 gigabyte files, the audio emulation matches very good with newer iterations of Gens (albeit with a low bitrate which affects sound effects such as this: https://www.youtube.com/watch?v=0qFM6H7Jn1g#t=2m) and that it is the second most popular emulator to use it is fair to assume that he did in fact use Gens as his emulator of choice.

This is important to note because that also means that the difference between legit runs and cheated runs are completely indistinguishable with 1 single built in command. What I mean by that is not that you can't tell when a run is obviously TASed but rather that people passing off runs with savestates as single-segment are literally impossible to tell apart from legit runs. You cannot prove beyond a doubt that a suspected run done on Gens in fact was cheated without a confession. However I believe the burden of proof should not lie on the person analysing a run, pouring hour after hour into researching tiny details in common emulators and their derivatives about sound quality differences, aspect ratios quirks, internal recording processes and their derivatives in order to try to determine if a guy is using an emulator which should be banned anyway because apparently "there's still doubt because you can't prove that he didn't cheat!". The burden of proof should be on the runner, however a video made in Gens is not proof in my book. He could have skipped recording it and just tweeted that he got these times and that would be equally valid proof.

I choose to disbelieve ridiculous claims like these when cheating is so easy (and prevalent) and the runner makes no effort to try to prove himself and in fact actively hides it or avoids the topic after getting called out and questioned.

Now I'll try to detail a few of the reasons why this is a ludicrous video:
GH2: Probably the 2nd easiest zip in the game to do. Says he can't do it yet gets ridiculous zips later in the game. Surviving off camera is hard, zipping/wrapping is not.
GH3: Comment "100%". Sure buddy you continue to think that. Benefit of the doubt and let's say he meant 32 when getting the bounces, quite unlikely still.
M1: Video & audio skips at 2:37. Very large difference in apparent skill compared to GH yet gets one of the sickest most precise jumps in the entire game at 3:16
M2: Stands literally dead center on the sign as it loads. Maybe he found a method to do that on his own but my personal guess is that he savestated and just looked where he ended up to show off since all his sources mention another method with 100% consistency (running when a certain wall passes the left boundary)
M3: Didn't align with the right platform at the start. Not a necessity but makes it more consistent because the timing of the jump can vary by idk 2-3 frames otherwise. Also sick IL boss strats bro 10/10
SY1: Black flash at 5:14. Also sick frame perfect jump at 5:28 it's almost like you can't miss them for some reason. Being able to hold right at 5:35 like that is subpixel dependent not sure of the leniency but it's not great iirc.
SY2: Plays like you'd expect someone who ran the game for a month would play. Didn't jump towards the wall to manipulate his subpixels although he should've known about it if he really did study TeeNTee and Naegleria close enough to beat the WR by 57 seconds. Again nice frame perfect jumps at 6:28 6:32 lol (latter could possibly be 2 but not more I'm too tired to look it up)...
SY3: RISKS LITERALLY HIS OWN LIFE, HIS FIRSTBORN CHILD AND HIS PET GOLDFISH FOR NOT A SINGLE REASON AT ALL. Also he wastes like 5s+ just getting in position in the wall but 1 frame jumps are just too easy for this fella no sweat. 7:53 black flash.
L1: 8:47 YOLOSWAG GOTTA GET THAT SPEED! Seriously that platform is like half the size it looks to be and he ends up going too fast and getting hit anyway like please. Incredibly risky jump for no gain.
L2: Reasonable. Oh and it's the best he has done in a SS run ever together with 3 other levels apparently. No SL3 is not one of them I guess.
L3: Rings are for scrubs anyway who'd even consider wasting half a second grabbing them when you're well on WR pace especially when you're just rushing through the boss not watching the fireballs? 10:31 Audio skips for some reason.
SL1: Subpixel dependent I think like 1/3? Add to that the fact that he zips in between the bomb fallout. That's a 3 frame window +/-1 so again big risk for very little gain. Again with the standing still at the signpost but at least this wasn't perfectly spaced.
SL2: You can manipulate your subpixels here which he didn't which means the trick is essentially random. Pretty precise I think around 1/4-1/5 or slightly less although you can get multiple attempts but it's no problem to Gvozdevik the Sonic 1 master who of course gets it first try. Black flash at 13:10
SL3: Now this is the real kicker in the run. Idk the exact chance of getting the right subpixels for this and I don't feel like doing it tonight but it's probably 1/30 or less. No one who knows how the trick works would go for it in a SS run and I don't believe anyone uploading a YT video from Gens without knowing how the skip works to actually get it in a run. A run this good to boot. It doesn't happen. How much time does he save compared to a good run of the level normally? Around 10 seconds. Definitely worth it. Doesn't complain about a horrible time for doing the wrap or get amazed in his comments for for getting it at all either it's just "like SL2 but harder". A few days later (http://www.soniccenter.org/rankings/sonic_1/times/star_light_3/sonic) he learns an actual boss strat which is insane since no one in their right mind would learn the wrap before learning the boss.
SB1: Going for the zip after everything that has happened is signs of either a mad man or a cheater since by all normal standards less than 1 in 10'000 runs would get even close to this point with these times. But gotta go fast I guess... 14:48 the jump there is not too bad like 2-3 frames I think but getting past the wall he does is pure luck and very rare. I think it even has a chance of killing you iirc. So obviously our man, our hero Gvozdevik has no problem executing it even though his comments make zero sense ("go right-jump"??). 15:03 WHO HAS TIME TO SLOW DOWN TO BE SAFE WHEN YOU CAN GO SLOWER BY JUMPING INTO THE FIRE AND RISK DYING?
SB2: Surprised he didn't go for the wrap here ^_-... Black flash at 15:47.
SB3: Again subpixels. Getting it first try is lucky but not too hard. I hope he bought a lottery ticket after this run though.
F: Fails the pausing strat which is the easiest in the whole run. You literally can't fail it if you just spam pause until you reach the room. Proceeds to do some risky jumps on the boss I'd almost be nervous about doing in IL's. This man has no fear, I wonder why.


I could continue this but I'm too tired you can ask me if you want me to elaborate on something. Sorry if I got a bit silly towards the end there.