There have been a number of questions about seeding for the weekly playoffs that are worth addressing in long-form rather than ye olde 140 character tweetage.
Good Seeding is the Foundation of Good Competition
First, at Gamestah we understand that good seeding is the foundation of good competitions. We know that the purpose of seeding is to provide the best possible chance for the two top teams in a tournament to meet in the Grand Final, and not before. This means that seeding is especially important in single elimination brackets. With Double Elimination, you can still have the one and two seeds meet before the Grand Final, and have the loser work their way back to a Grand Final berth through the lower brackets. That is simply not possible with single elimination.
This is why, before JAM-STAH started back in May, Gamestah convened the Overwatch Oceanic Power Rankings Panel. We needed some basis for seeding the JAM-STAH tournaments, especially considering they were all single elimination. The Power Rankings has been tremendously helpful in this regard, not just for our own competitions, but also for others in the region. We also used the Power Rankings to determine our first four auto-qualified teams in Week 1 of the ZOWIE Gamestah Overwatch League.
ZOWIE has a unique tournament structure that requires different approaches to seeding for each bracket on a weekly basis.
The Weekly Qualifiers use a six-round Swiss format bracket. The Qualifiers are open, so each week there is a mix of teams that are new to the competition, and teams that have played before. The latter will have established League Points and a position on the ZOWIE Leaderboard. The Leaderboard is our primary resource for seeding the open Qualifiers each week. We supplement this with reference to the Power Rankings. For example, if a team enters the Qualifiers in Week 5 and doesn’t have any League Points, but has been ranked by the Power Ranking panel at no. 2 (for argument’s sake), it would be foolish for us to give the team a bottom seed and have them face the top seed in Round 1.
The Swiss bracket is played out and at the end of the sixth round, battlefy generates standings based on the results of all six matches each team plays. From these standings, the top four teams qualify for the Playoffs.
It should be noted that Swiss format brackets require tie-breakers to determine final standings. While there are multiple methods of breaking ties, battlefy’s bracket software uses “the average of the percentage of 1. Rounds your opponents won 2. Individual games you won.” We accept this method without question as it takes into account battlefy’s round-generation algorithm.
At the end of each Weekly Playoff, battlefy generates a standings table based on results. First place goes to the Grand Final winner, second place to the Grand Final loser, third place to the Semi-Finalist that lost to the Grand Final Winner and fourth place to the Semi-Finalist that lost to the Grand Final loser. These four teams automatically qualify for the next week’s playoffs, and we seed them based on their standings as generated by battlefy.
The four teams that come through the qualifying Swiss bracket are seeded according to the standings table generated by battlefy. This means that the first-placed qualifier is seeded fifth, second place is sixth, third place is seventh and fourth place is seeded eighth in the Playoffs.
Your Tie-breakers Suck and You Should Use Mine Instead
A number of players have approached us with this notion. Many methods suggested are quite valid. They just aren’t used by battlefy’s system. To be clear, Gamestah did not sit down and discuss whether or not we would use head-to-head, or opponents-win-percentage, or least-opponent’s-score-overall, or Bucholz or any other method to break ties. We simply noted that battlefy implemented a tiebreaker system, that the system used two accepted Swiss-format tiebreaker algorithms, and we decided to use that system.
The battlefy system has been used to determine rankings for every Qualifying stage from the start of the League. Those who have recently qualified for a playoff round, and who would have preferred a different method of tiebreaker because it would have resulted in a higher seeding for their team (excuse me, would have been “more fair), should be aware that applying a unique tiebreaker to this week’s Qualifier standings would unfairly disadvantage all of the teams from the previous rounds.
In short: we have a valid system for breaking ties, it has been equally and thus fairly applied to all teams throughout the League, and we won’t be using a different method to suit special snowflakes.