You absolutely certain about that reasoning? Because from what I've seen, when automated matchmaking is used, you NEED to play the game like a job just to reach your "correct" ranking and actually enjoy the game. People who don't play it like that are driven away because of it.
If you're curious about the mechanics behind ELO and ELO confidence distributions after X matches, chess ELO is actually a well studied way to learn about the algorithm used by almost all SBMM. After a shockingly small number of matches, your ELO is going to end up being in the right neighborhood for you have +/- 50% WR.
I play games that are so niche that the 'matchmaking' consists of pinging people on Discord. Because we don't have proper matchmaking, we struggle to retain new players because they come in, get pulverized into the dust, and give up.
The point of matchmaking is that even a more casual beginner can find opponents at their level, without having to grind a ton to catch up with those of us who have been playing for years.
Titanfall 2 come to mind here. I bought it well after launch and really enjoyed the campaign. When I went to hop into multi-player, I was often killed as I spawned or within 10s of spawning. I literally was not playing the game at that, just spawning and dying. I never came back, lol.
Yes, I am.
This is just one study I could find quickly but the results are consistent.
Because from what I've seen, when automated matchmaking is used, you NEED to play the game like a job just to reach your "correct" ranking and actually enjoy the game.
This is not accurate. Most people's ELOs don't shift much after settling into your "natural" rank, which should happen after about 50 matches or so. Probably what you're referring to is the publicly available "rank" which is per "season", wherein every few months your rank gets reset. This is FAR less opaque than SBMM but results in lower playtime and lower retention for casual players who don't want to be grinding the 50 matches to settle at their ELO every 3 months.
Actual opaque SBMM (the algorithm you mentioned originally) that never resets creates, on average, much more fun MP experiences for most people.
Most people’s ELOs don’t shift much after settling into your “natural” rank, which should happen after about 50 matches or so.
Ehm, 50 matches seems like a lot to me. Especially if they aren't enjoyable (yet) because of flawed matchmaking.
I pulled that number out of my bootyhole because I knew it was a safe bet for a stable ELO.
US Chess Federation uses 25 games as your provisional ELO stage, many video games will use 10 matches. Assuming a large enough variety of ELO in the player base, you can be confident your ELO is mostly accurate after a shockingly small number of matches.
Would be interesting to see but I would assume most people won't even make it to 10 matches in a game they don't enjoy. The people who spend thousands of hours on a single game are a tiny minority of the tiny minority of people who have the free time to play dozens of a hours a week.
If you can't make it 10 matches in a new game, I don't think SBMM is your problem with the game.
10 matches should be like, between 3-10 hours. Assuming an hour a night, you'll be approximately ranked for SBMM within a week.
Do you understand why people play games though?
Warcraft 3 multiplayer was peak "matchmaking" in my opinion, where people created lobbies with certain rule sets and anyone who was interested in that type of game could just join directly. It was a blast, playing lots of different game modes all the time and meeting a wide range of player types, instead of having to invest an insane amount of time (3-10 hours, vs less than a minute to find a game in WC3) into one single game mode even before you can actually start playing.
What you have described is exactly what I was talking about when I called it "playing the game like a job," where you have to invest plenty of time before you can even hope to enjoy it.
Warcraft 3's custom games were a mess, people left all the time which made team games irritating as hell, and the skill level varied widely from one game to the next so half the games ended up with feeders and a stomping one way or the other.
Do you understand why people play games though?
I understand why I do. I can't speak to your motivations, I'm not you. I can, however, point to studies that discuss groups of people's preferences in aggregate, as I have done. You're an outlier, and that's ok! Play what you want how you want!
SBMM is, unfortunately for you, the current utilitarian optimal for multiplayer PvP gaming. It maximizes both adoption and retention metrics, as well as self-reported enjoyment scores (Likert scale) for the highest number of people. Bummer that it doesn't optimize for you, but the other good part is that there are plenty of games that still support custom lobbies. Find one you like and have fun!
You are never going to answer that question with math and statistics, and attempts to do so are exactly why the industry keeps tanking studio after studio.
It should take about 20 matches or less to give you a decent rating, what games have you played that took longer?
You absolutely certain about that reasoning? Because from what I've seen, when automated matchmaking is used, you NEED to play the game like a job just to reach your "correct" ranking and actually enjoy the game. People who don't play it like that are driven away because of it.
If you're curious about the mechanics behind ELO and ELO confidence distributions after X matches, chess ELO is actually a well studied way to learn about the algorithm used by almost all SBMM. After a shockingly small number of matches, your ELO is going to end up being in the right neighborhood for you have +/- 50% WR.
I play games that are so niche that the 'matchmaking' consists of pinging people on Discord. Because we don't have proper matchmaking, we struggle to retain new players because they come in, get pulverized into the dust, and give up.
The point of matchmaking is that even a more casual beginner can find opponents at their level, without having to grind a ton to catch up with those of us who have been playing for years.
Titanfall 2 come to mind here. I bought it well after launch and really enjoyed the campaign. When I went to hop into multi-player, I was often killed as I spawned or within 10s of spawning. I literally was not playing the game at that, just spawning and dying. I never came back, lol.
Yes, I am.
This is just one study I could find quickly but the results are consistent.
https://www.pcgamer.com/games/activision-secretly-experimented-on-50-of-call-of-duty-players-by-decreasing-skill-based-matchmaking-and-determined-players-like-sbmm-even-if-they-don-t-know-it/
This is not accurate. Most people's ELOs don't shift much after settling into your "natural" rank, which should happen after about 50 matches or so. Probably what you're referring to is the publicly available "rank" which is per "season", wherein every few months your rank gets reset. This is FAR less opaque than SBMM but results in lower playtime and lower retention for casual players who don't want to be grinding the 50 matches to settle at their ELO every 3 months.
Actual opaque SBMM (the algorithm you mentioned originally) that never resets creates, on average, much more fun MP experiences for most people.
Ehm, 50 matches seems like a lot to me. Especially if they aren't enjoyable (yet) because of flawed matchmaking.
I pulled that number out of my bootyhole because I knew it was a safe bet for a stable ELO.
US Chess Federation uses 25 games as your provisional ELO stage, many video games will use 10 matches. Assuming a large enough variety of ELO in the player base, you can be confident your ELO is mostly accurate after a shockingly small number of matches.
Would be interesting to see but I would assume most people won't even make it to 10 matches in a game they don't enjoy. The people who spend thousands of hours on a single game are a tiny minority of the tiny minority of people who have the free time to play dozens of a hours a week.
If you can't make it 10 matches in a new game, I don't think SBMM is your problem with the game.
10 matches should be like, between 3-10 hours. Assuming an hour a night, you'll be approximately ranked for SBMM within a week.
Do you understand why people play games though?
Warcraft 3 multiplayer was peak "matchmaking" in my opinion, where people created lobbies with certain rule sets and anyone who was interested in that type of game could just join directly. It was a blast, playing lots of different game modes all the time and meeting a wide range of player types, instead of having to invest an insane amount of time (3-10 hours, vs less than a minute to find a game in WC3) into one single game mode even before you can actually start playing.
What you have described is exactly what I was talking about when I called it "playing the game like a job," where you have to invest plenty of time before you can even hope to enjoy it.
Warcraft 3's custom games were a mess, people left all the time which made team games irritating as hell, and the skill level varied widely from one game to the next so half the games ended up with feeders and a stomping one way or the other.
I understand why I do. I can't speak to your motivations, I'm not you. I can, however, point to studies that discuss groups of people's preferences in aggregate, as I have done. You're an outlier, and that's ok! Play what you want how you want!
SBMM is, unfortunately for you, the current utilitarian optimal for multiplayer PvP gaming. It maximizes both adoption and retention metrics, as well as self-reported enjoyment scores (Likert scale) for the highest number of people. Bummer that it doesn't optimize for you, but the other good part is that there are plenty of games that still support custom lobbies. Find one you like and have fun!
You are never going to answer that question with math and statistics, and attempts to do so are exactly why the industry keeps tanking studio after studio.
It should take about 20 matches or less to give you a decent rating, what games have you played that took longer?