It would be more useful as a theoretical thing if you actually can easily tell match length of a casual play/blowout easily but you really cannot.
Immaterial and Missing Power runs literally over 10 minutes long for full matches on high end play. Low end tends to manage about 4. Marvel vs Capcom tends to cut the opposite way, as hitting successfully once can mean death from full life...if you know the combos.
I don't really object to the concept, but unless it's a game people have enough experience with to say stuff about, it's going to be unfeasible.
For my guesses based on knowing some of the basic mechanics, Tekken and Melty end up about what you'd expect of high end play for time, on the low end. Street Fighter ends up shorter. (The large boosts of damage from combos tend to be balanced out by the fact that low end players won't zone the same way as high end.)
MvsC2 and TvsC end up notably longer, there's no real zoning element so there will simply be much more back and forth in standard play of them due to the lack of killing combo knowledge. In fact, swapping characters allows for fairly huge amounts of healing in MvsC2, but tournament level players often kill before this fully goes off. Casual players do not. I wouldn't be stunned if MvsC2 took twice as long, if not longer, if both sides grasp the basic concept of "Swap at low health". (And...thinking on it this is backed up by what I've seen of low end players screwing around.)
TvsC is less so, two characters instead of three, but we're not talking about fast stuff here either.
2vs2 Brawl...you know I was going to say this ends up longer, but thinking back to the elements that make Brawl long...perhaps not. The slow paced dodging and maneuvering for position flies out the window somewhat when anywhere you dodge could have a smash from a second person intersecting it. Hmmmm. Well, hard to say.
Tetris Attack and SPF2T both will be shorter with good players. Especially Tetris Attack. Unless both sides are good, then Tetris Attack can run into three+ minutes a round, but then again people will love seeing that if it occurs and it'll occur relatively little, so.
Anyways, the point is that these are estimates, though; I most certainly can't give exact numbers because they are going to be hard to give without researching, and to research you'd basically have to find matches of people who are clearly not at all knowledgeable on the game, matches of them vs people who are somewhat knowledgeable, and so forth.
But while it sounds good in theory to math out the time, chances are it's going to end up wrong with large gaps between games if you try to strictly adhere to a plan like that, as near as I can tell.
