Haven't watched the video yet, but that's really not the main goal of comeback mechanics, at least not anymore. A large part of it is to add excitement to the viewership and hype for the players themselves. I think many mechanics in these games exist for that purpose. X-Factor, V-Trigger, Fatal Blow, Sparking Blast, etc. make the viewers excited and bring swings to a match that make them more interested in watching.
It's the same reason certain character archetypes and playstyles have been deemphasized in modern fighting games, building games around rushdown and getting in on the opponent, because that's what viewers like to watch the most.
FighterZ is pretty much the epitome of all of these issues put together, but people like to bag on SFV for it.
It's a bit more complicated than it just being a viewership thing. Part of it is also fundamental game design, you're trying to avoid
-Lame duck scenarios where the game is basically over, but the players have to play it out and prelong the inevitable
-Slipper slope scenario where the winning player can easily boat race someone to the finish because they gain too much of an edge (character advantage in tag games, meter advantages with a life lead in a game with more defensive neutrals, etc).
The problem then becomes balance. I agree with Daigo in that they aren't needed, but I also don't think for instance Sparking is necessarily that bad of a set up. Something like Danger time in guilty gear is more stupid because of its randomness and the players inability to plan around it. Goes without saying X Factor was 10 times more bullshit. Especially with our lord n savior, and totally fairest character in the world Vergil.
The reason SF gets shit is because SF's neutral has never really been some air dasher 50/50 fest, tag games on the flip side are totally that way. MvC3 is busted as all hell, and Skullgirls rewards setting up resets to steam roll goons. MvC2 is a prolonged scramble.
SFV in its effort to correct the overly defensive nature of 4 and it's absurd OS's over swung the game and gutted shit, it was laggier, they made DP's n reversals not invincible, crouching jabs anti-airing stuff, crush counter+vtrigger (would argue crush counter is the more offensive system out of the two) compromising the thing that makes Street Fighter so ace, which is it's neutral n footsies game.
Ultimately I'd say both are good games, but wildly outclassed by deeper fighting games, but there is a reason SF gets more shit. You don't get to be the face of the genre and make those mistakes. DBFZ benefits from decades now of not getting a good fighting game in that series. But I do think that game warrants a year later discussion on its many short comings like super dash, the lack of air options against super dash, the homogenized nature of characters, the issue that has come from everyone only having one assist, how all the characters being so samey makes balancing the game a bit of a nightmare without basically making the game more dry.
I include all fighting games in this. It doesn't matter that MK11 is less rushdown heavy than MKX was: the game still incentivizes that playstyle the most. The game has maybe 2 legit zoners out of its entire cast right now. The game is still all about the 50/50s up close. Look at all of the top tiers (Sonya, Geras, Erron, and Scorpion) and you'll see exactly what this game values and where the meta revolves around.
To be fair, I think MK's zoning issue is an entirely different thing. It is pretty amazing how badly NRS misunderstands the fireball game in every other 2d fighter.