Right off, Version C seems awfully impractical. I know that best emulates what one would normally consider a tournament system, but until there's a system implemented directly into the game to handle it, I wouldn't bother. Just too clunky.
Version A and Version B seem like essentially the same thing, with t = 3 hours for the former and t = 3 days for the latter. If there are intended to be more explicit differences than this, I'd appreciate some clarification in that regard - that said, it's worth discussing regardless.
The biggest obstacle I see is that any of these systems would have a hard time encouraging players to keep playing once they've hit the prerequisites for the next round - in order for other players to participate they need people to play against, and if I stop playing as soon as I hit my requirements, that potentially ruins their chances. If players were able to carry some advantage over from the first round into the second, that would encourage them to play for 'first place' even before they're actually capable of winning the tournament outright. And since the goal of the whole thing is get people actively playing PvP, keeping them playing seems important.
In general, I'd probably favour larger t values over smaller ones - while the focus that smaller ones provide is nice, the accessibility of the larger ones is probably more relevant in this context. I think the game would need a larger playerbase for the other to work best, so the option that allows more of the available players to participate seems ideal. Something interesting to note about consecutive long tournaments: since a successful player moves on to the second day of one tournament, he wouldn't be able to access the first day of the next tournament; this introduces some natural variance to the tournament winners, since the winner of one tournament naturally can't have participated in either of the following two. This is mostly a good thing, but also requires offering some natural compensation to players who make it through the tournament system; something like a Silver Pack for making it to day 2 and a Gold Pack for making it to day 3, most likely. This would also give a good excuse to send players a reminder for tournaments they've participated in whenever they log in, and would incentivise playing through part of a tournament even if you know you won't be around for the final days.
Another thing I'd like to see is a balance between rewarding lots of play (e.g. going 24-12) and rewarding consistent play (e.g. going 12-0); I'd probably have the first round of the tournament base its cutoff on number-of-wins, while the second round qualifier adds in a consistency requirement (probably 2-1 or better). That way the tournament system doesn't simply favour players who have the time available - or live in the right time zones - such that they can simply play with larger percentages of the roster. Obviously any system will favour them somewhat, but that limitation would limit the effectiveness of simply fighting every player available.
Altogether, I'd probably consider a system somewhat like this:
Quote:
Round 1
Uses the Version B 'special queue' for unique matches. Players get 10 'points' for each win. At the end of the round (hour or day), the top ~50% of players (obviously depending on your intended size and scope) move to Round 2.
Round 2
'Unique match' tracker rests. Players get 25 'points' for each win; points from the first round 'roll over' (to encourage players to keep active during it), but second round wins are still more valuable. At the end of the round, the top ~50% of remaining players with a 2-1 record or better move to Round 3.
Round 3
No longer uses Version B 'special queue' - players may face each other any number of times. Players earn 50 'points' for each win, plus some portion (10%?) of your opponent's 'points' (which they lose). At the end of the round, the player with the most points wins.
The rounds are each designed to fit specific goals: the first round weeds out inactive players and allows especially active ones to gain a head start; the second round weeds out players who got through the first round primarily based on activity; and the third round has the mixed role of encouraging high level play while discouraging things like hasty concessions (the point loss) and 'sitting on' a high score (repeatability and large point rewards) - basically trying to ensuring that the tournament is decided by actual play, rather than gaming the system.
The one concern I have is that because the Round 3 point system is generating points, players could theoretically form coalitions to farm them - if two players agreed, for instance, that each time they were matched up the one of them with more points would forfeit, then they would be able to introduce more points to the system whilst only moving them primarily between each other. The obvious fix is to eliminate that feature (e.g. just having the 50 points you win taken directly from the other player), but not being able to generate points encourages players who are already winning not to participate (lest they risk reintroducing those points to the system). I'm not sure which of the two I'm really more concerned about. It would also be possible to solve this by restricting players to playing each other only once, but that seems like it would result in more 'lame duck' players (can't win, but still playing) and would limit the number of observable high-end PvP matches, which rather defeats the purpose. sweatdrop
I'll spend some more time thinking about it, and see if I can't come up with something better...