Ok, so I think most of us who play ranked formats and cannot quite get to Onyx rank can relate to this: solo play is impossible because you have no idea who you’ll be matched with, even with focused on.
There are a few upsides and downsides to each matchmaking preference that I would like to discuss, starting off with Balanced.
First, balanced is not necessarily balanced. You often find that in team games you get aligned with people who are either still qualifying or people who around the same ranking as you. This in and of itself is not a bad system, as it does attempt to place you with people who are around the same skill level as you. However, you often find the problem in those people who are still qualifying to be the problem, as there are people who have played one match of qualifications, on your team, who end up going 2-15 (or something tk that degree). While you cannot necessarily blame the person, it can be frustrating when it becomes a consistent problem, and sometimes you can find out later that someone on your team qualified to be Bronze 5 because they
Now, I would like to clarify that I do not know how people get matched, but however it is made, I wish that it could be adjusted in one manner (which I’ll get to in a bit).
Now, Focused is really the ideal system, but sometimes you find that your hidden rank is much higher than your actual rank because you simply got matched with some players who aren’t necessarily at your skill level. In addition, you find the opposite problem, which can negatively impact the players around you, and you negatively impact players who have an abnormal rank in relation to their hidden skill, but often times this is simply a manner of necessity. Another problem with Focused, though, is that a lot of games come down to luck, and your CSR can sometimes stagnate based on who you get matched with. This isn’t necessarily a problem with CSR or the matchmaking system, but rather a problem of players of roughly the same skill level being matched together.
Now, Expanded is really only ideal if you simply want to play a game and be done with it. I think that the problems of it are self explanatory, but I will at least say this: too random to realistically expect good teammates a good portion of the time, casuing it to come down to who the best player is.
Now, to get onto the main point: how do we adjust CSR so that it can be a more fair system that doesn’t rely on luck for solo queuers? (Note: this does not necessarily apply to high ranks like Onyx+. It is merely discussing how to adjust the CSR system so that it doesn’t punish better players for being matched with worse players).
I propose two changes: one- make qualification matches seperate and follow an ELO-esque system; two- do not reward the worst player on the winning team and punish the best player on the losing team.
For point number one, I imagine something like this: your hidden rank information sets your base ELO, and depending on factors from each game (your Kill spread, objective points, assists, etc.), your “ELO” goes up or down based on how well you did. The reason why these matches ought to be kept seperate in my opinion is because of two reasons: because matchmaking cannot necessarily assert the skill of a player in qualification matches with a reasonable amount of certainty, and because the prior simply can become an unpredictable hindrance upon those who qualified into their rank and are doing everything in their ability to improve it. This is a problem that we cannot always avoid, but by assuring players that a player has the skill to reach a certain rank (rather than luck), they can be more confident in the abilities of others overall.
As for my second point, I believe that punishing a consistently good player in relation to everybody else in the game because it suddenly became a 2-4 because some people went afk for half the match isn’t a healthy thing to do. In addition, if you have a team that wins (in what would otherwise be a blowout manner) a very close game, if there was a player who had some ridiculously horrendous spread, then they should not be benefiting from CSR. (Note: I am unsure if this point applies. If it doesn’t, then ignore it, since I will have retracted it).
I do acknowledge that this system has many problems and would be difficult to actually implement, so I beg of you to insightfully analyze and discuss this topic.
If you have any constructive criticisms, please do discuss them below.