As many of you have probably noticed, there have been some changes to matchmaking in Season 2.
In season 1, matchmaking was done in order of highest-rated in the bot to lowest-rated. Lower rated and unrated players were consistently kept from playing.
In a league that claims to pay more than just lip service to inclusiveness, this is surprising. But in a league where the number of games played has some positive correlation with your final score, it's unfair. The more games you play, the better chance you have of a strong performance, an advantageous team stack, a lucky streak, etc. Everyone knows this. No-one contests it. So by preferencing those who are rated highly in matchmaking, absolutely guaranteeing that the highest rated always play if they show, we further stack the odds against lower and mid-range players.
Remember, this isn't a group-oriented league, where you'll always want to play your strongest players. It's based on individual performance.
I saw something that undermined the integrity of the zone, and made a call to step in and change it. This is a sysop's prerogative. The first attempt wasn't a great one, relying on games played, preferencing those who haven't played as many games in the matchmaking. But this resulted in the same problem as before, with the same people being consistently benched. It also de-incentivizes showing in general. (Why show if you know you'll be benched?)
In lieu of a better solution, since last Sunday, bench selection has been fully random, using no criteria whatsoever. It's hardly ideal, but at least it doesn't provide a reverse handicap to top-rated players as happened in Season 1.
We can to come up with a set of criteria to determine bench selection that does not preference players unfairly. For example, the number of games you've played in the particular league out of the number needed to qualify could determine 20-50% of the decision, with the rest being random. Or other criteria could be added and used.
We can also look at trying to come up with the most evenly-matched teams, given the players available, benching those not used in matchmaking. This is a bit difficult, but I've been looking at some algorithms.
Or, if none of these options are acceptable, we need to completely rework the league such that the number of games played is not positively correlated with final score.
Or, finally, we can redefine the purpose of the league. We can call it a vet's league like every other one, and explicitly acknowledge that it unfairly preferences the best-rated, is not meant to be inclusive as claimed, and then work on solutions based on that premise.
Thoughts?
In season 1, matchmaking was done in order of highest-rated in the bot to lowest-rated. Lower rated and unrated players were consistently kept from playing.
In a league that claims to pay more than just lip service to inclusiveness, this is surprising. But in a league where the number of games played has some positive correlation with your final score, it's unfair. The more games you play, the better chance you have of a strong performance, an advantageous team stack, a lucky streak, etc. Everyone knows this. No-one contests it. So by preferencing those who are rated highly in matchmaking, absolutely guaranteeing that the highest rated always play if they show, we further stack the odds against lower and mid-range players.
Remember, this isn't a group-oriented league, where you'll always want to play your strongest players. It's based on individual performance.
I saw something that undermined the integrity of the zone, and made a call to step in and change it. This is a sysop's prerogative. The first attempt wasn't a great one, relying on games played, preferencing those who haven't played as many games in the matchmaking. But this resulted in the same problem as before, with the same people being consistently benched. It also de-incentivizes showing in general. (Why show if you know you'll be benched?)
In lieu of a better solution, since last Sunday, bench selection has been fully random, using no criteria whatsoever. It's hardly ideal, but at least it doesn't provide a reverse handicap to top-rated players as happened in Season 1.
We can to come up with a set of criteria to determine bench selection that does not preference players unfairly. For example, the number of games you've played in the particular league out of the number needed to qualify could determine 20-50% of the decision, with the rest being random. Or other criteria could be added and used.
We can also look at trying to come up with the most evenly-matched teams, given the players available, benching those not used in matchmaking. This is a bit difficult, but I've been looking at some algorithms.
Or, if none of these options are acceptable, we need to completely rework the league such that the number of games played is not positively correlated with final score.
Or, finally, we can redefine the purpose of the league. We can call it a vet's league like every other one, and explicitly acknowledge that it unfairly preferences the best-rated, is not meant to be inclusive as claimed, and then work on solutions based on that premise.
Thoughts?
Comment