I think ultimately, what’s fair will come down to down to what your definition of fairness is and the intention of the overall tournament. For most people, Bracket B will probably seem more fair. Top players can expect to be more likely win than any player that is worse than them. Furthermore, top players can expect their skill to have even greater influence when it comes to reaching the finals and ultimately earning 2nd place. This is especially important at top levels of professional play, where economic outcomes are directly tied to winning or losing. For professional players, placing lower than a less skilled opponent and ultimately being compensated less would definitely seem unfair. Since most tournaments seek to determine the top player as consistently as possible, it makes sense that Bracket B’s seeding is considered “traditional seeding”.

I do think, however, than an argument can be made for Bracket A to be considered more fair. Under such a format, on average, top players still generally can expect to win more often, but lesser players have better odds. Furthermore, below average players have significantly improved chances of reaching the finals and placing second as compared to Bracket B. For some players, they might have better odds of making it to finals than a better player. Under such a seeding system, skill would be a lesser determinant of average result and the choice to participate at all would carry greater weight. These properties probably make Bracket A pretty undesirable to professionals and spectators compared to Bracket B.

The context where Bracket A’s seeding system would potentially be desirable would be in “arcadian” or amateur tournaments. I actually have personally played in arcadian tournaments before while playing Super Smash Bros. Melee competitively. These tournaments specifically denied entry to top level players both regionally and globally so that average level players would have a greater chance to win. These tournaments were somewhat infrequent, but especially popular with attendances often rivaling those of local major tournaments. From a tournament organizer’s perspective, these tournaments set out to better engage the average player who made up the majority of the tournament attendee population.

Because of the number of players and overall low amount of time and exposure within the scene, most players didn’t actually know how good they were relative to each other. This would make it difficult for them to tell what kind of seeding scheme had been employed. Thus, if one employed Bracket A’s seeding, then there was a high chance that the median player would place second and could potentially even win the tournament. Either outcome would be high visibility and would cause that player to be regarded as having high potential within the area among other players. Players that are actually more skillful than the high-placing median player would then play them either casually or in traditionally seeded tournaments, likely beating them and feeling like they’d improved since the arcadian tournament, giving a sense of progression to above average players that they might not normally get. So if this seeding scheme were employed but not discovered, it would actually serve as a source of motivation for increased participation in subsequent tournaments, regardless of whether they followed an arcadian ruleset or not.

Thus, in similar settings where the aim of the tournament is to drive player engagement over skill-accurate outcomes and where this seeding system would be difficult to detect, then Bracket A’s seeding could actually be more desirable. To this day, I’ve never asked the tournament organizer that ran the Michigan arcadian tournaments, but I wouldn’t be surprised if this was at least part of the conversation when seeding the bracket.

So to conclude, we can see that mathematically, Bracket B is more “fair” in terms of what people generally expect from tournaments, with better players placing higher and winning overall more often. Bracket A is more “fair” in absolute terms, in that the average result between players is more equal on average and is less heavily influenced by skill. Depending on context, either bracket could be desirable, and it comes down to the designer to decide on the desired tournament experience and then select the appropriate tools to achieve that.