No More Win Based Skill Ranks

(NOT A RANT) I’m going to be blunt, I really really dislike the way competitive ranks have been implemented, I love the idea of competitive ranks, but the way they were made in halo5 is awful in my opinion. I know the majority of people that play halo probably couldn’t care less about csr and that’s completely fine, it’s not a good game without your casual players.

However I’ve seen it over and over again, really good players stuck in silver and gold tiers, their kd’s are good, they aren’t lacking in skill, but it’s getting paired up with a team who is well…below par. It’s the same reason I go back and forth between onyx and diamond 6, sure sometimes it’s my B and that’s my fault, but lots of times the team lets you down. Stuff happens which is why the way the system is set up is terrible.

Ever since I got my first rank after 10 games I have been dumbfounded as to why an individual’s skill tier was almost completely reliant upon how the rest of his team performs each and every game. It seems obvious to me that an individual’s skill should be based on…well…an individuals skill. This is not a rant I want actual ideas and solutions to this, because I truly don’t want this same system to be implemented in Halo 6 and only the community is going to convince 343 to change it.

I think they should make it have some to do with winning and losing, and some to do with individual performance. So lets say your team loses, but you did really well and you were the only reason you team had a chance, your rank would only go down a little or not at all. Or if you did really but, but your team pulled the W, you would only go up a little. I believe Overwatch’s ranking system does this and I think it works great for the most part.

Also, a little off topic here, but they should also change just how matchmaking is, make it so that teams are paired with teams, and solo players are paired with solo players. Because I’ve been in games where I know my team is all random people and the enemy team is a party of 4 and they destroy us, even though we may have a higher degree of skill, they won because they were able to communicate better then us and they could clean up kills after a fight.

I would agree with this. It should have less to do with the Win and more to do with your personal skill performance in each individual game. I would say that if you lose a match, but you had 30 kills and the other 3 team members never broke double-digits that your rank should go up significantly. But the opposite team that beat you and earned the Win had a teammate that got carried and only dropped 2 kills - that player should lose rank even if they won. Or lets say you earned the Triple Double medal (double-digit kills, assists, and headshot) that your rank gets a boost for earning that as well.

The biggest benefit of this would be no more boosting. Great players can’t boost the lesser skilled if this were the case because the lesser skilled would still do poorly and lose rank.

I think this would be a far more accurate skill rank method to create. I don’t know how difficult this would be from a developers perspective though.

> 2533274905191374;1:
> (NOT A RANT) I’m going to be blunt, I really really dislike the way competitive ranks have been implemented, I love the idea of competitive ranks, but the way they were made in halo5 is awful in my opinion. I know the majority of people that play halo probably couldn’t care less about csr and that’s completely fine, it’s not a good game without your casual players.
>
> However I’ve seen it over and over again, really good players stuck in silver and gold tiers, their kd’s are good, they aren’t lacking in skill, but it’s getting paired up with a team who is well…below par. It’s the same reason I go back and forth between onyx and diamond 6, sure sometimes it’s my B and that’s my fault, but lots of times the team lets you down. Stuff happens which is why the way the system is set up is terrible.
>
> Ever since I got my first rank after 10 games I have been dumbfounded as to why an individual’s skill tier was almost completely reliant upon how the rest of his team performs each and every game. It seems obvious to me that an individual’s skill should be based on…well…an individuals skill. This is not a rant I want actual ideas and solutions to this, because I truly don’t want this same system to be implemented in Halo 6 and only the community is going to convince 343 to change it.

Ya I agree for the longest time I was stuck in Gold/high silver. But I got pretty lucky and found a team and now I play with a pretty good team if 4.

> 2533274925727172;2:
> I think they should make it have some to do with winning and losing, and some to do with individual performance. So lets say your team loses, but you did really well and you were the only reason you team had a chance, your rank would only go down a little or not at all. Or if you did really but, but your team pulled the W, you would only go up a little. I believe Overwatch’s ranking system does this and I think it works great for the most part.
>
> Also, a little off topic here, but they should also change just how matchmaking is, make it so that teams are paired with teams, and solo players are paired with solo players. Because I’ve been in games where I know my team is all random people and the enemy team is a party of 4 and they destroy us, even though we may have a higher degree of skill, they won because they were able to communicate better then us and they could clean up kills after a fight.

Completely agree with your second point, going 4 randoms vs a full squad is brutal.

> 2533274905191374;1:
> However I’ve seen it over and over again, really good players stuck in silver and gold tiers, their kd’s are good, they aren’t lacking in skill, but it’s getting paired up with a team who is well…below par.

And if their teammates truly are consistently worse than them, so are their opponents. And if everyone else in the game is consistently worse than them, then which team they get put into will, on the average, have a higher porbability of winning. And if you win more games than would be expected at that skill level, your rank will go up. The cold hard truth is that wins are all that is needed to determine your skill level in a team-based game. If you can’t advance beyond a certain level, it’s not because you’re getting bad teammates, but because you’re not performing well enough to progress which, considering your complaining about your teammates, probably has to do with lack of teamwork.

What you should also understand is that depending on your level of skill, you may actually be performing better than the players you face without actually being better than the level that has been assigned to you. Strictly speaking, when you get faced against opponents around your actual skill level, you’re only expected to perform as well as the opponents you face if the skill distribution of players is uniform. If the distribution is at all slanted, then there might be less players who are better than you around your level, and more who are worse than you, making it more likely to get matched against players below your skill level than above. This is intuitively clear if you imagine that you’re the best player at the game, but it happens to lesser extent at any level that’s above average.

> 2533274905191374;1:
> Ever since I got my first rank after 10 games I have been dumbfounded as to why an individual’s skill tier was almost completely reliant upon how the rest of his team performs each and every game. It seems obvious to me that an individual’s skill should be based on…well…an individuals skill. This is not a rant I want actual ideas and solutions to this, because I truly don’t want this same system to be implemented in Halo 6 and only the community is going to convince 343 to change it.

“Individual skill” is a meaningless notion in a team game. If you want your individual skill—i.e, your ability to play as a one-man army—to be judged, you should play FFA. In a team game you are expected to work as a team, and if you don’t, you’re not performing as well as you could, and as the skill assigned to you is only as good as your performance, your skill level will be adjusted accordingly. After all, the rating system can’t know that you’re not working as hard as you could. But if you want it to recognize your skill, you need to work harder.

I can’t speak for the Halo 5 system, because like everyone else, I don’t know how it works. However, I can speak for win-based rating systems in general and say that they are completely par for the job. The community should, in fact, have no say in what sort of rating system the game will have because they have neither the knowledge nor understanding of the existing system, let alone how to create a better one. Suggesting that the system should be based on “individual skill” is useless unless you can describe how that “individual skill” should be introduced into the system and explain how that gives, in any meaningful way, a representation of a player’s skill.

I’ve written an explanation of the principles behind skill based rating systems. I serously recommend reading it—at least the first section, and glancing at the later two. I also recommend Jeff Moser’s article which is a great explanation of the problem of quantifying skill. After all, you can’t truly criticize something unless you understand what it is that you’re criticizing.

> 2533274825830455;6:
> > 2533274905191374;1:
> >

Yes, but shouldn’t you be given less of a beating if your team loses but you alone got like half the kills?

Let me pain a picture, you’re put into a game with 4 random people, the enemy team is in a part of 4, but individually, they are not as good as any of you. The team you are facing off against has a great level of communication and can clean you up after a fight if you can’t get away or find another team mate to assist in defending you. The 3 other people you are paired with have a slightly below average skill and your own skill is slightly above them, but not by much.

The enemy team begins to make a push for the sniper rifle on Coliseum, Your team goes for the rocket launcher. You know that if you let the enemy team get the sniper, they can hold down the rocket spawn with the sniper. You try to tell you team to come back and attack the sniper spawn, you’re able to get one person to come help you, but you will not win a 2v3 or 2v4 battle.

The fight for the sniper becomes a 2v3, you have no clue where the 4th guy is, but either way, you and your buddy push for the attack. The 4th guy got behind you from the bridge and was able to flank you, they had no loss. They got the rest of your team going for the rockets. They won the first fight with 4 to 0. But now, for the rest of the game, the enemy team has such great communication that you can’t get a grasp of either the rockets or the sniper, your 3 team mates trickle in to make a push for the sniper, but are defeated very easily. The game is eventually lost 50 to 30. You were the only one in your team to not push for the power weapons and instead you used your pistol and AR and other weapons on the map to get picks. You are the top of the leaderboard with 13 kills and 9 deaths. Your buddy at the beginning also did this, but was new to the idea, he went 11 and 10. He will likely be using this plan in other underdog leveled games.

Even though the enemy teams individual skill was not as good as yours, you lost and went down with your rank. But not any less then your other teammates, you went down just as much as your thickheaded teammates. That I believe is unfair and should be changed. Eventhough you had good communication with your team, they ignored it, eventhough you had a higher degree of skill then the enemy team individual skill, they called you out and cleaned you up. You lost a battle you couldn’t win.

And due to the low player base in the playlist, you get matched up with the same team in the next game. This time, someone on your team leaves. You know how this fight ends, you lose, again.

Now that scenario is possibly based off of an older version of Coliseum, but you get my point.

=============================================================================================

Now what I propose to help this is to do one of 3 things:

  • You implement a matchmaking system that tires to make solo players pair with solo players and teams pair with team, as I stated above. - You implement some sort of underdog system, where if you lose against a team, you don’t go down as much or if you beat a team, you go up more then you would before. - You implement a leaderboard rank into the mix, where you may have lost the game, but because you did the best on the team, you don’t go up very much. If you win but did the worst, you don’t go up very much. This would also have to include what % of kills were with power weapons, how many times did you pick up a power weapon, what was you overall accuracy, how many assists did you get, (if possible) how often did you stick with your team mates or use your mic?Adding at least one of these things would help the ranking system greatly IMO, and would help with these alleged clams of “My team sucked and I went down??? WTF 343” blah blah blah stuff.

And I realize that Halo 5 might be too far in its lifespan to add a new matchmaking system or to revise the ranking algorithm, but please, at least make it in Halo 6.

(And good god, that has to be one of the longest thread comments I’ve ever made in all my time in the Waypoint forums. Hope I didn’t make any silly grammar errors.)
(Also had to delete most of what you said, sorry about that.)

> 2533274925727172;7:
> > 2533274825830455;6:
> > > 2533274905191374;1:
> > >
>
> and would help with these alleged clams of “My team sucked and I went down??? WTF 343” blah blah blah stuff.
>
> > Hope I didn’t make any silly grammar errors…

I found your silly grammar mistake :stuck_out_tongue: “alleged clams”

> 2533274825830455;6:
> > 2533274905191374;1:
> >
>
>
>
> > 2533274905191374;1:
> >
>
> I’ve written an explanation of the principles behind skill based rating systems. I serously recommend reading it—at least the first section, and glancing at the later two. I also recommend Jeff Moser’s article which is a great explanation of the problem of quantifying skill. After all, you can’t truly criticize something unless you understand what it is that you’re criticizing.

Thanks for the links - both articles were extremely useful, although I must say that yours made me cry before I’d got past the second page! I haven’t had to think that hard in years and it most definitely did not agree with me. To all the kids: use your brains now while you still have them.

Thanks for the articles @Tsassi ,however I’m going to have to disagree with you on the point of “it’s not about individuals” sure, in general to win Halo matches teamwork is required more so than say CoD. However while the gameplay isn’t strictly for the individual, the majority of people always look at the individual instead of a whole team, and teammates look at individuals to judge how well they can perform to help the team. Companies (including mine) often recruit with one of the parameters having to be above a certain CSR, and they are looking at the individual when recruiting not a team. I think 343 got it wrong so what sue me, all I’m saying is that in Halo 6 having CSR should be based on the individuals performance which could also then be used to match teams more accurately.

Teams are made up of individuals.

Poor preforming teammates drag you down, despite having a positive K/D in the majority of matches. Don’t know why I should be punished based on a win/lose scenario. Also doesn’t help with the introduction of new players after every holiday, especially as they don’t know what they are doing most of the time, just standing around. This is rage inducing. You can’t correct the mistakes of the team, and lead them to victory, your skill means nothing. Halo 6 multiplayer needs to be based on individual skill, not on win/lose.

> 2533274835379871;8:
> > 2533274925727172;7:
> > > 2533274825830455;6:
> > > > 2533274905191374;1:
> > > >
> >
> > and would help with these alleged clams of “My team sucked and I went down??? WTF 343” blah blah blah stuff.
> >
> > > Hope I didn’t make any silly grammar errors…
>
> I found your silly grammar mistake :stuck_out_tongue: “alleged clams”

haha, theres one lol

> 2533274925727172;7:
> Yes, but shouldn’t you be given less of a beating if your team loses but you alone got like half the kills?
>
> Let me pain a picture, you’re put into a game with 4 random people. . .

The problem is, you’ve already gone astray when you try to use one match as an example of what you think is wrong. It doesn’t matter what happens in one match. What matters is how your team performs on average, and if you truly are better than the players in your skill range, then your team will win more often. There’s no preference to put you in the worse team.

> 2533274925727172;7:
> And due to the low player base in the playlist, you get matched up with the same team in the next game. This time, someone on your team leaves. You know how this fight ends, you lose, again.

When the player base is small, it’s not the system that’s your problem. It doesn’t matter what system you use. If it evaluates your skill relative to other players in a proper manner, and matches you based on that skill, if the player base is small and you’re well above average, the average level of players you match will be lower than your own. Nothing will change that.

> 2533274925727172;7:
> Now what I propose to help this is to do one of 3 things:
> - You implement a matchmaking system that tires to make solo players pair with solo players and teams pair with team, as I stated above. - You implement some sort of underdog system, where if you lose against a team, you don’t go down as much or if you beat a team, you go up more then you would before. - You implement a leaderboard rank into the mix, where you may have lost the game, but because you did the best on the team, you don’t go up very much. If you win but did the worst, you don’t go up very much. This would also have to include what % of kills were with power weapons, how many times did you pick up a power weapon, what was you overall accuracy, how many assists did you get, (if possible) how often did you stick with your team mates or use your mic?Adding at least one of these things would help the ranking system greatly IMO, and would help with these alleged clams of “My team sucked and I went down??? WTF 343” blah blah blah stuff.

The first option is a completely sensible one. The second essentially proposes that the skill of a team is more than the sum of its parts, which in itself is not necessarily untrue, but impossible to implement in practice to any consistent degree. The last option just puts less emphasis on teamwork, for which there’s no rational basis besides making those players who play team gametypes by themselves feel better about themselves.

I’ve been saying for years that there should be heavier restrictions on parties matching parties, because parties generally are better than teams of randoms, so avoiding matches between these two groups would lead to fairer matches, thus supporting the whole goal of matchmaking. However, introducingstricter party restrictions would mean significantly longer search times, which players would complain about. I have no problem with that, but many players might.

> 2533274925727172;7:
> (Also had to delete most of what you said, sorry about that.)

It’s a short post if you don’t need to.

> 2533274873843883;9:
> Thanks for the links - both articles were extremely useful, although I must say that yours made me cry before I’d got past the second page! I haven’t had to think that hard in years and it most definitely did not agree with me. To all the kids: use your brains now while you still have them.

Don’t know if that’s my style, my writing skill, or the topic.

> Thanks for the articles @tsassi ,however I’m going to have to disagree with you on the point of “it’s not about individuals” sure, in general to win Halo matches teamwork is required more so than say CoD. However while the gameplay isn’t strictly for the individual, the majority of people always look at the individual instead of a whole team, and teammates look at individuals to judge how well they can perform to help the team. Companies (including mine) often recruit with one of the parameters having to be above a certain CSR, and they are looking at the individual when recruiting not a team. I think 343 got it wrong so what sue me, all I’m saying is that in Halo 6 having CSR should be based on the individuals performance which could also then be used to match teams more accurately.
>
> Teams are made up of individuals.

Teams are made up of individuals, and an integral part of their skill is how well they can work as a team. Sorry, but no determination of skill is complete if you don’t pay attention to all aspects of a player’s skill. Ultimately, a win-based rating captures all aspects of a player’s skill, not just their ability to slay other players, but also their abilities from working as a team, controlling the map, playing objective, and so on. The win is a combined result of all the skills a team needed to win. Any combination of player’s kills, assists, deaths, or whatever, is just that: a selection of arbitrary statistics that only capture a small part of total skill. Not to mention, when you’re trying to implement these individual statistics, you have to worry about weighting, which is an entirely arbitrary process.

Win-based ratings capture everything about the player’s skill in a way that’s undisputable. There is only one statistic, and therefore no arbitrary weighing factors. Moreover, if you really just want to know how good a player is working all by themselves, ignoring all aspects of teamwork, you can always look at their FFA rank. There your have your individual if that’s what you desperately need. However, in a team game, players should also be judged based on their ability to work as a team.

You believe that rating players based on their “individual skill” would lead to more accurate matchmaking, but why? You have no evidence for this really, you only believe it because you want to.

It’s hard to take your points seriously when you have never even touched the game.^

> 2533274905191374;14:
> It’s hard to take your points seriously when you have never even touched the game.^

His point is still valid. Ranked playlists, beside FFA, are based on Team Skill not Individual Skill. Sure, one player can doom a team in a game, but that doesn’t put them entirely at fault. Instead of the team playing to the strengths of their weakest link, they continue to just blame the state of the match on that one player instead of changing their playstyle to suit them. It’s easy to choose someone at part for your loss, but you lost as a team because you all failed to work cohesively together.

I’ve had my fair share of matches with teammates that tend to do ridiculous things, but if you want any remote chance of winning under those circumstances than you need to do something about it besides playing the blame game.

> 2533274825830455;13:
> > 2533274925727172;7:
> > Yes, but shouldn’t you be given less of a beating if your team loses but you alone got like half the kills?
> >
> > Let me pain a picture, you’re put into a game with 4 random people. . .
>
> The problem is, you’ve already gone astray when you try to use one match as an example of what you think is wrong. It doesn’t matter what happens in one match. What matters is how your team performs on average, and if you truly are better than the players in your skill range, then your team will win more often. There’s no preference to put you in the worse team.
>
>
>
>
> > 2533274925727172;7:
> > And due to the low player base in the playlist, you get matched up with the same team in the next game. This time, someone on your team leaves. You know how this fight ends, you lose, again.
>
> When the player base is small, it’s not the system that’s your problem. It doesn’t matter what system you use. If it evaluates your skill relative to other players in a proper manner, and matches you based on that skill, if the player base is small and you’re well above average, the average level of players you match will be lower than your own. Nothing will change that.

I gave you that scenario because it’s happened to me more then often times. Every other game I play I’m facing off against a team of 4 and I get matched with them again, and again, always loosing until I give up and go play something else.

It may have been an example of one game, but similar things such as this happen very often to me. Even my friends have claimed to be in games like this where the tides have been turned in favor of the team of 4, with no chance of the 4 random players winning.

> 2533274905191374;14:
> It’s hard to take your points seriously when you have never even touched the game.^

I mean, I have, but that’s not really relevant to this discussion. Nothing regarding how skill ratings should or shouldn’t be handled requires information specific to Halo 5. To be frank, the fact that I understand how performance predicting skill ratings work in general probably makes me more prepared for this discussion than any amount of playing Halo 5 could.

I was trying to come up with a way of making the problem of skill ratings more approachable after my last reply. So I give you this: your problem is to match player together in a fair manner. The first question you need to ask is: what does the outcome of a fair match look like? How can you quantify this? (Hint: you want to minimize something.) When you can quantify a fair match, you can quantify performance. Now that you have a performance metric, how can you use it to optimize the outcomes of matches? (Hint: you need to be able to predict outcomes in order to know what’s a good match.) Bonus points if you can do this in the general case where you have an arbitrary game competitive game.

> 2533274925727172;16:
> I gave you that scenario because it’s happened to me more then often times. Every other game I play I’m facing off against a team of 4 and I get matched with them again, and again, always loosing until I give up and go play something else.
>
> It may have been an example of one game, but similar things such as this happen very often to me. Even my friends have claimed to be in games like this where the tides have been turned in favor of the team of 4, with no chance of the 4 random players winning.

But that’s a matchmaking problem, isn’t it? I mean, ideally you’d want to match randoms with randoms and teams with teams regardless of how you rate your players because teams will always be better than randoms. Chances are that if you’ve reached a high enough level, you’re running into the mass of players who actually want to win and choose to use teamwork. Ultimately, you’re not using your full potential as long as you play as a random. It’s really only sensible that you’re not going to rank as high as the players who do use teamwork.

Don’t you think it’s kind of silly to complain about not getting a higher rank because you are running against players who are better prepared than you are? I mean, there’s obviously a reason they are doing better than you, but instead of working on that you want the bar to be set lower.

> 2533274905191374;1:
> (NOT A RANT) I’m going to be blunt, I really really dislike the way competitive ranks have been implemented, I love the idea of competitive ranks, but the way they were made in halo5 is awful in my opinion. I know the majority of people that play halo probably couldn’t care less about csr and that’s completely fine, it’s not a good game without your casual players.
>
> However I’ve seen it over and over again, really good players stuck in silver and gold tiers, their kd’s are good, they aren’t lacking in skill, but it’s getting paired up with a team who is well…below par. It’s the same reason I go back and forth between onyx and diamond 6, sure sometimes it’s my B and that’s my fault, but lots of times the team lets you down. Stuff happens which is why the way the system is set up is terrible.
>
> Ever since I got my first rank after 10 games I have been dumbfounded as to why an individual’s skill tier was almost completely reliant upon how the rest of his team performs each and every game. It seems obvious to me that an individual’s skill should be based on…well…an individuals skill. This is not a rant I want actual ideas and solutions to this, because I truly don’t want this same system to be implemented in Halo 6 and only the community is going to convince 343 to change it.

First, a Mercenary playlist for solo searchers would be ideal.

Second, IMO the win/loss system is perfect, as winning is the only thing that matters. Think chess. If you lose your queen in the 3rd round yet still win the match, does it matter that your opponent’s queen took out half your board after you already lost yours? No, it doesn’t matter at all, because you still won. Everything that happens in between is just noise.

Reach’s rank system was originally based on individual performance. It lead to extremley selfish play and pretty much killed the team aspect of the game, thus things went back to win/loss. If you want to be ranked by individual performance play ffa, a team game should rank teams not individuals. If you can’t work well with a team then you don’t deserve to rank up. If random teamates are the issue, simply party up with good teamates.

I don’t know what a perfect ranking system would look like but I do know that being the only one who went positive on your team but still losing, then going down in rank is not fun.