In bracket-styled tournaments, why is seeding not randomized?

75 views
0

Namely in fighting games, I’ve noticed that bracket seeding is often determined prior to the tournament through rankings — but, why not let it be randomized? What difference does a seeded and non seeded tournament make?

In: 2

The top-ranked individuals/teams are rewarded for their high ranking by getting an easier path to the championship.

You don’t want the two most likely winners to meet each other in the first round and then have the finale be an anticlimatic average player (who by chance only encountered the bottom half of players) meeting the almost guaranteed victor.

So they deliberately sort the best players into different brackets to have them meet up late in the tournament.

Very often, you also give some prizes to players that are not first place. Maybe it’s a per-win bonus, maybe they get new fans, maybe they just outright get a 2nd place trophy. If best player eliminates second best player in round 1, your 1st place may not be affected, but your 2nd place and further are.

Seeding ensures that everyone ends the tournament where they should if all games were won by the better-ranked side.

Seeded tournaments are set up so that, should all the favorites win, the top four teams will meet in the semifinals.

Of course upsets can happen, but in many competitive events history has shown us that the lowest seed is highly unlikely to upset the top seed.

In the notoriously top-upset-prone NHL, the lowest 8 seed wins around 1/4th of the first round matchups against number 1.

In the notoriously *not* top-upset-prone NCAA tournament the number 1 seed has only lost once in the first round, in 144 attempts.

So you can see that seeding doesn’t mean much in the NHL – the “worst” playoff team isn’t that much worse than the “best” and can routinely win, or at least produce a good showing.

Conversely, seeding the tournament is critical in the NCAA tournament – the best teams trounce the worst more than 99% of the time, and you’d hate to have such a pointless matchup occur several rounds in after much better teams have been eliminated.

So one major argument is, as others have mentioned, about making the tournament more exciting (for the viewers *and* for the competitors, who theoretically face harder and harder opponents in each successive round). Another is that, in a tournament of skill or ability, the “proper” outcome is for the most skilled competitor to get first place, the second most skilled competitor to get second place, and so on, including having the worst skilled competitors score last. It turns out that arranging brackets by seeding* is the best way to make that happen.

Imagine a 64-player fighting game tournament. The “proper” outcome, if we perfectly knew everyone’s skill level, is for the worst 32 players to get eliminated in round 1 (for simplicity, let’s pretend it’s single-elimination for now, although double-elim tournaments have the same issues). The way to make that outcome happen is to match each of the “worst 32” competitors against someone from the upper 32. If, at any point, you match two of the “upper 32” players against each other, then one of them must necessarily be eliminated, and some other match, which ends up being two players from “worst 32”, will qualify one of those “worse” players to round 2.

In the most extreme case, the seed 1 and seed 2 players might play each other in round one, causing the second most skilled player to actually take last place. This is an outcome which is at odds with the purpose of the tournament (that is, for players to place according to their relative skill). Also, depending on the prize scheme (e.g. second place prize), there could be actual financial impact on the players. So, to prevent this, brackets are arranged by seeding (and occasionally a few other minor factors).

* Seeding is supposed to be based on a measurement of skill. In an ideal world, seeding would reflect a competitor’s true skill with complete accuracy. In practice, this is obviously impossible to get perfectly right. Also, for very large tournaments, it’s often too much work to evaluate and seed every competitor. In those cases, a tournament might only seed a certain number of top competitors, and everyone else just ends up in an “unseeded” category.