As long as there have been fandoms, those fans have posed the eternal (and eternally cliched) question: Could Character X beat Character Y? The oldest archives of online discussion already feature groaning and eye-rolling at Superman vs. Goku or the Enterprise vs. a Star Destroyer. When a child asks whether Wolverine could beat Santa Claus, it's cute. When a 20-something launches into a 15 minute YouTube video explaining exactly WHY Wolverine would beat Santa Claus...it's a lot less cute.
But the problem isn't really the extreme jerks who argue obsessively about the question. It's that the question itself is a bad one to ask, as are variants like "Could Character X beat Character Y if List of Conditions is met?" or "Who are the three strongest characters in Particular Fictional Universe?" They're attempts to put a hard numerical ranking on something that cannot be ranked like that with any validity. Trying to turn everything into a game of High Card Wins, where a ten always beats a deuce.
There's a lot of reasons this doesn't work, but in this essay I'll focus on what I consider the two biggies:
On the simplest level, that means that when you set up a Who'd Win, you need to be more specific because the characters tend to evolve with the stories. Which Superman? Goku at what age? And when you cross properties, you need to figure out how they could interact. How do a Star Destroyer's Deflectors actually work, and would they stop a phaser? Would Goku be able to tap into the Chi of Superman's version of Earth? These tend to be dealt with by specifying which version of each, and postulating that everything works like the user expects it to. Goku can make a Spirit Bomb, phasers and blasters are both charged plasma so defenses work the same, everyone's on a hypothetical Neutral Ground, etc.
But that's not enough. These are not just from fiction, they are of fiction. A Superman plopped in an arena and told to fight isn't just going to work like a playable video game character, he needs a story that brings him to the fight and motivates him. The writer ideally needs a better plot than "it'd be cool to see these two fight." There needs to be a character development (or illusion thereof) arising from the outcome. What a character would plausibly do in a fight depends on the story, which means that as long as the writer can figure out how to make it happen, the winner is whichever one makes the story work. This doesn't always mean the winner is whoever's book the fight happens in, the writer might think their protagonist needs to be taken down a few pegs.
An important thing to consider in light of this is that a fight might not even be the main point of a conflict. Sure, after a while every arc in Dragonball (Z, GT, etc) became some sort of tournament so that the fight was the point. But in most stories, a fight is a means to an end, and a lot of fights end inconclusively because the goal has been attained or thwarted before the fight finishes. This is especially true of heroes fighting each other, as they tend to stop when either they realize they're on the same side, or when they figure out how to foil the outside force that is making them fight. Most people need a reason to fight, and will stop fighting once there's no longer a reason. If you want your fictional characters to ring true, they can't all be Fight Man ("I'll fight until the cows come up, then I'll fight the $%*# cows!"). The reasons the author gives for the fight will have more influence on the winner than anything intrinsic to the characters. A deuce can beat a ten if the ten is throwing the fight to avoid the death of a captive loved one, or the deuce is better prepared and knows a way to cheat.
Let's look at an example of a Who'd Win other than the cliche Superman vs. Goku. Deathstroke the Terminator versus Batman, any versions where neither has a super god-mode plot device (i.e. Deathstroke does not have Wally West's speed, Batman does not have Metron's chair, etc.). Just some version of their Usual Stuff.
No writer who respects the characters is going to just dump them in a fighting pit and have them duke it out until one of them yields or is lying in a pool of blood. If they were to wake up in an arena with a voice from above commanding them to fight or he'll flood the place with nerve gas, they'd work out a plan in five seconds without exchanging so much as a few glances and team up to find a way out (probably involving one of them throwing the fight in order to be taken somewhere hopefully less secure). If Batman is in the way of Deathstroke's objective, the last thing Slade wants to do is let Batman know this...he'll set up some fake side plot, concealing a second fake plot, concealing a third fake plot, so that by the time Batman unravels it all, Deathstroke has finished his business in Gotham (which is what happened in #4-5 of the Rebirth-launched Deathstroke book). And even then, Batman probably figured it out two steps earlier and has found a way to ruin or at least dampen Slade's real plan. And that's a much more interesting story for the reader than 20 pages of fight scene.
The fights may happen. They may even have a clear victor. But they're usually a means to an end in the storytelling, and the next fight is always another story, with whatever outcome the story needs. Goku wins because Superman was about to annoy Lord Beerus and there was no time to talk. Superman wins because Goku was going to try to fight Parasite for fun and that could have been Very Bad. Superman and Goku trade blows for a bit, Goku smiles and compliments Superman, then teams up with a slightly confused Superman to deal with Friezaniac's cybernetic henchmen. The winner, if any, is determined by the story.
"They're fictional" is such a common reflexive response that a reflexive counter-response has developed: "Yeah, but what if they were real?" Okay, what if they were real? Football teams are real, and in a movie about a fictional football team, Al Pacino's character says, "On any given Sunday you're gonna win or you're gonna lose." He was going in for a point about what loss reveals about someone's character, but I want to focus on the fact that neither winning nor losing is guaranteed...at least when the conflict is interesting.
Sure, if you pit any lineup of the New York Yankees against a 5th grade gym class team, the Yankees will win every time. But that's not really a question anyone's likely to ask. I mean, you're not going to ask "Who'd Win: Goku or Harvey Bullock?" I suppose Goku could take pity on the guy, but that would just mean stranding him on top of a flagpole or something. Looking at baseball, the best record in the MLB for 2017's regular season belonged to the L.A. Dodgers, with 104 wins and 58 losses. The worst in their division was the San Francisco Giants, at 64 wins and 98 losses. That's a pretty big spread. So, if the Dodgers played the Giants, who'd win? Let's check the actual record. From May through September of 2017, the Giants and Dodgers played 12 games. The Dodgers did not win all 12, they only won 8. That's nearly identical to their overall record (67% wins, versus overall 64%), and if even one of the close games had gone to the Giants, the best team would have done worse against the worst team than against all teams in general.
Interesting conflicts, even real ones, don't work like Stratego with obvious win/lose/draw conditions. Even if you look at individuals rather than teams, you can point to thinks like the two title fights between Muhammad Ali and Joe Frazier...each man won one of the two bouts, and Ali won a non-title bout in between them. Same two guys, short span of time, both pretty much in their primes, two different results.
Interesting conflicts hinge on small details, and resist reductionist "Who'd Win?" analyses. If it were possible to get a definitive answer, bookmakers would be put out of business. Even if there isn't a story coming from an author, there's always a story that explains why this time the underdog won. Simulations and games that try to treat fictional conflicts as real generally use some sort of randomizer to account for these hidden stories, and the ones that don't tend to do a bad job of prediction when they are applied to real life situations (i.e. wargaming resolution systems where you can only get a clear victory when the margin is 2-1 or better, ignoring plenty of fights between more or less equal forces that did have clear victors). Interesting conflicts are never as simple as high card wins.
To sum up, the question of who'd win is less important than the story being told, and if you're going to divorce it from story the answer usually ends up with a big "probably, maybe?" tacked onto the answer, because there's still a story even when you don't want one.
Don't ask who'd win. Ask how it could make an interesting story.