Players Blame Skill-Based Matchmaking for Losing in Call of Duty

It's also worth noting that this is only half of the complaints that you see on message boards about this game. The other half, or maybe the majority, is people saying, "I just want to be able to get high and use off-meta weapons and not lose to sweats (i.e. people who are trying) all night." This type of complaint is fundamentally a request to be matched against lower skilled players, and has nothing to do with the swinginess of matchmaking.

With Dota2 (a while ago might have been fixed) my complain was more than SBMM wasn't tight enough and unfortunately due to game dynamics a single sub-par player in the team could give a big advantage to the adversaries (and I totally found myself in this quite unpleasant role).

So maybe i'd say elo hell exist in some way as when the matchmaking algorithm doesn't have enough data to match you adequately it must train itself by matching you using a pretty loose standard deviation. I guess the success of "good matchmaking" actually depend on how this training sequence is more or less gracefully handled from the player perspective.

What are the chances this is an 'engagement' hack?

I wonder if players are more likely to continue playing if they have ups and downs.

In a perfect matchmaking system, would players get bored?

I have to wonder if this is intentional, or just the side effect of either perception or an imperfect system.

Also, just to be clear, no implementation is this simplistic. Most models are Bayesian and so incorporate uncertainty around the estimate of skill. And most also include your individual stats (this is game-dependent but I believe Overwatch has a class/map-based ELO that you are rated against...as an example).

I agree with your general point though. Most SBMM systems do not feel close games every time. Usually, someone is smurfing, someone is throwing, someone won't come on comms, etc. (yes, I play Overwatch and I am still angry about it).

1. There are a lot of different game modes for CoD -- Cold War has eleven game modes. This means that the pool of players for each mode will be smaller, making fair matchmaking more difficult.

2. Some of those game modes have unusually large teams -- two of the signature game modes in Cold War have ten or more players to a team. This makes it more difficult for a matchmaking algorithm to fairly assess any individual player's contribution to their team's win/loss. It also makes it more likely that one or more teams will contain a player whose skill has been poorly assessed, meaning that their team has an unfair advantage or disadvantage.

There are servers tiers, but they can't always fill a server with same rank players. How many players are online on the same region that want to join a server at a given time...

So there will be always some randomness to it. As long as you sometimes get to play with good teams and win, it will keep you in the game. You will probably even play longer that way.

Of course same with the notion of "elo hell". I would be deeply ashamed of myself if I would defy all logic and sense to find excuses for my lack of skill.

I have played games without SBMM (Destiny 2 recently switched away from it in certain game modes) in periods when the sweatiness is particularly high and people join as premade teams to stomp some noobs. Let me tell you that as a casual PvP player, it absolutely ain't fun.

I think the current state is a decent compromise. it's totally viable to make a new account whenever you have a new friend that starts playing. as long as you only play with them on the smurf account, it will stay in a close MMR range. your friend has to go through the grind anyway, so you may as well do it with them. the grinding does make it pretty annoying to make new accounts over and over again just to tear through the ranks.

You have huge advantage just playing with your friends in a party because most of the other players don't even use microphone to communicate with their squad, and those that do are just random players matched together.

Also popular Battle Royale games are cross platform which means you play against a lot of console players. A lot the console gamers play on high latency TV screens with WIFI connection. Not to mention that an average console player can't aim as good as an average PC player.

Sometimes you want to have a good game and have that excitement of completely crushing it.

Also, why would a "pro" want to sandbag like this? I've been an amateur athlete most of my life and my goal was never to compete against known weaker athletes - there's no fun in that.

Or, do these games lack the ability to spin up "practice" sessions fully under the control of the player? That seems like the real answer, and would mirror what athletes in meat-space do - train for weeks/months, then compete at well-timed intervals against the best (or, at least relatively equal, if you're doing age group amateur competitions).

The most hilarious thing is when people ask that there be a ranked playlist with sbmm and an unranked playlist without. The one without, they assume, would be more favorable to them, since they would experience a random distribution of players (and they assume they are above average, naturally). But for players in the bottom decile of skill, unranked would have a dramatically worse experience than ranked! So they would leave for the ranked playlist. Now the new bottom decile of skill would be having a terrible experience in unranked, so they would leave, etc. By induction, unranked would quickly become depopulated.

It is really tough to be a below-average player in a game without any matchmaking. You lose all the time and almost never have an impact. You can kiiinda get away with this as a developer in games that have very large teams. For example, World of Warships has no skill based matchmaking (although tier level is somewhat a proxy for skill and performs a bit of a weed-out of the lowest skill players in the higher tiers). There are people who play that game who have only a 30% winrate over a large number of games. This is in a 12v12 game where presumably things tend to average out in the mix of skills on each team, so having that negative of an impact on team performance is in some sense pretty impressive.

The only thing difficult for me to understand is the resilience of the mentality involved in continuing to play that game if you are one of the way below average players. I would not have fun if I never had a positive impact on my team's performance.

It's also worth noting that if WoWS had the playerbase to support it, they'd probably build SBMM too. But it's a bit of a chicken and egg problem. The issue is that their monetization system requires what is effectively forty or fifty different queues (5 tier-based matching bands * 4 ship classes * 2-2.2 game modes, depending on whether ranked is active). Queue times can already be quite extended in some tiers and for some ship ship classes. Developing a system that continues with this alignment of tiers and classes and also substantially reduces skill gaps is probably unsolvable without a much large player base or a more complicated role selection system.

Was generally quite nice, except when you really wanted to play with your mate and the tool kept swapping you over because it didn't find any other way to balance the teams.

Rarely did we ever win though, usually we'd end up getting admin kicked for 'cheating'.

This need not be a hard requirement, so:

- Playing in the middle of the night, no previous teammates around? You get a random crew

- Playing at the same time every day? The system slowly builds a team of likeminded people around you.

STFU!!!! You get paid to play a friggin' video game. Sit down. Shut Up. Get better.

For me, Super Smash Brothers Melee is a great example. Let's say I'm rank 1,000 in my city, and I play someone rank 250. I expect to loose 99% of those games.

I can't say the same thing is true for Call of Duty. It may be perception bias, but I think there's a LOT more variance between skill and victory. What's the breakdown of expected to actual performance? Would it be a 90% win rate, etc?

I have seen some of their recordings, the difference in performance between them and their teammate is much lower than they think.

Most people want to have fun when they play. Players consistently getting curbstomped and having no reliable avenue for improvement is a good way to create a game with terrible player retention. People don't stay around to suffer through game after game just because they really like ninja.

"Ranked" implies a competitive, high-stakes mode, which drives people who want a relaxed game (i.e. one where they _won't_ repeatedly get curbstomped) to naturally gravitate away from a game described as such. You may think that's stupid, but you can't really fight basic human psychology.

Don't forget, "n00bs" are people too.

Asking for the converse is fundamentally a request to coddle experienced players instead.

While this could've held true a few years ago for CoD franchise, I know for a fact that neither last year's CoD installment nor the most recent one discussed in this article (Black Ops: Cold War) has any sort of lootbox mechanics.