Search
  • Elliott 'aizyesque' G

Opinion: The Case for Random Seeding

To paraphrase England footballer-turned-crisp-advocate Gary Linekar: Counter-Strike is a simple game; 10 people run around the server headshotting each other and at the end, Astralis wins. Astralis won the IEM Katowice Major with barely a stiff breeze to stop them, as they had done just half a year prior, and it seems nothing was going to change that.


With very little to actually talk about in the Major, due to the fact that ‘who is going to win?’ was never really a worthwhile question, discussion turned to seeding, and how it apparently made for the best playoffs ever where only the best teams made it. However, FaZe got absolutely obliterated by an off-colour Natus Vincere, Renegades got blown out of the water by MiBR and NiP and Liquid also got 2-0’d in the quarter finals.


Much of the community seemed happy with the seeding, but very few people were willing to look at it from a different perspective; so I’m here to provide the alternative angle.

Randomness is inherently fairer than bias - it’s why scientific studies attempt to remove all bias, human or otherwise, to get fairer results. If everyone starts the tournament with an equal chance to qualify, with the only variables being randomness and their own performance, that is fairer than the variables being their ranking (of which is controlled by forces outside of the tournament being played) and their performance. Outside of the server, in my eyes, there should be no inherent advantage to being the favourite going into the tournament, because each tournament is a new challenge. Consistency is nice, it’s important, but in a one-off tournament, I’m not sure you should be rewarded for performances in other tournaments.


For this tournament, the team I choose for this example are AVANGAR, who showed a surprisingly solid level but were given an extremely small chance of making it through, realistically. AVANGAR smashed the Major Qualifier, beat both FaZe and NRG (who started as high seeds, and are ranked very highly by HLTV) whilst losing to Liquid. Their reward? Best of threes against NaVi, the third ranked team, and ENCE, who went on to make the final. Had this been completely random, this would have been incredibly unlucky, but at least one could blame Lady Luck for her dislike of the CIS side. Instead, the system was rigged to make it harder for an unfashionable team to make it through.


This same seeding system protected a team like FaZe, who had been slumping before, during and after the tournament, and had a new player in the line-up - but their reputation belied this. They were given an obnoxiously high seed by their peers, and despite going 1-2 (losing to AVANGAR, the #15 seed as just covered) they were given the #16 seed, CoL, in a survival game, and then C9, who had done nothing to suggest they were a good team at the tournament. Again, had this been random, a little bit of luck for a tournament is nice! - but instead, FaZe’s interests were protected and they were able to make Legend status.


Seeding intentionally makes it harder for teams with lower reputations who are playing well, and protects teams in slumps. It’s not like previous systems had realistically stopped the best teams from winning the tournament - the favourite almost never fails to get out of groups, and with best of threes implemented in the Swiss System this was even less of an issue as good teams would no longer have to deal with ‘random’ best of ones. Random seeding wouldn’t produce random outcomes - the teams who were playing better would go through, generally, as long as we played BO3’s. I would like random seeding to give everyone the same opportunity (not outcome) at the outset, rather than it being biased towards the ‘better’ teams.


One issue with seeding is the sheer volatility of the rankings in Counter-Strike - whilst Astralis have been the top dog for a while now, every ranking below that changes massively over a few weeks, and sometimes it’s not obvious from a numerical ranking when that happens. For example, at the time of writing, MiBR are the #8 ranked team and would probably get a similar seed. Ask yourself if you think MiBR are the 8th best team in the world right now. As there is no offline league format we don’t have a consistent ranking of how good and reliable these teams are in a LAN situation and we go from tournament to tournament which isn’t a consistent indicator of team level, and MiBR’s recent tournaments have been shambolic at best.


Seeding is particularly slow to react to these volatile changes, and it isn’t figured out until a team underperforms - that team is less likely to be challenged if they come in as a favourite but aren’t playing very well, as they will play ‘weaker’ teams and their ranking won’t shift as quickly. Similarly, if decided by the other teams, those teams are more likely to protect the teams they know and have played so many times against, and often other pros have slightly warped views on who is good. For example, would a player like JW be more likely to say a team with his ex-teammate Olofmeister is better because he recognises how good he was (and is) even if there was a better team he hadn’t seen much? That’s not to say players would deliberately mess them up, but humans are biased where randomness is not and you of course would believe the player you know is better than the one you don’t know as much about.


Sadly this argument always occurs from the position of a big team. When Mouz played FaZe in an elimination game, we heard ‘well why should Mouz have to play FaZe in an 0-2 game?’ (the answer is because they lost two games in a row by the way), but nobody asks the same question of CompLexity, who showed a decent level to muscle past NRG, before being given the same task of beating FaZe to survive. Mousesports, by the way, failed at the EU Minor after going 0-3, so maybe their inconsistency was a bigger issue than the format.

In the end, the discussion boils down to: Would you prefer every team to have the same inherent chance and only be judged on their performance in this tournament, or should we slightly rig the system to try and ensure the best teams make the playoffs? It should be remembered that even with the new seeding system, every quarter final was a 2-0 and some of them were extremely one-sided, so even with the system it didn’t ensure amazing back-and-forth quarter finals.


I don’t think a team should be rewarded during one tournament for their performance in another, though I understand some people disagree, I’d like to state that as it seems many people don’t understand my point of view. I love underdog stories, and seeding attempts to make it harder and harder to upset teams as winning a tough game just rewards you with more and more tough games if you’re given a poor seed to start with. I’d just prefer every team, even the ones you’ve never watched before, to start with the same chance. That’s all.

0 views