Jump to content
Sign in to follow this  
Qbit

Ranking System Discussion

Recommended Posts

Feel free to discuss all things related to ranking systems in this thread. Do you like Halo 5's ranking system? What are the good and bad things about it? How could it be improved? Would you replace it with a new system entirely? What would that system entail? Let's hear what you have to say.

 

It's widely agreed that only having a single matchmaking mode which pairs up both random players and full organized parties alike is both unfair and undesirable. Thus, a simple solution is to provide an option to explicitly search for games that are randoms vs randoms (RvR) or parties vs parties (PvP).

 

For the PvP option, the ideal implementation is to use a zero-sum W/L system which awards 1 point to each player on a team who wins a match and deducts 1 point from each player on the team that loses the match.

 

To fairly match one team against another, a number of options are available. The first is that the number of points that each player on a team has are added together and divided by the number of players to get an average. Teams are then matched based on their player point average and will be matched against the closest matching currently available team.

 

Another option is to match teams based on the player who has the highest number of points from each team. If a team has a player who has 40 points, they'll be matched up with an available team who's highest rated player's point total is closest to that amount.

 

That was simple and straightforward, but things get trickier when we talk about RvR. The problem with W/L in RvR is that lesser skilled players get carried by more skilled ones and higher skilled players get dragged down by lesser skilled ones. This undermines the system. Lets imagine in their first game of Team Slayer a poor player gets matched up with better players. The poor player goes -8, but the better players manage to carry the team to a win. In the eyes of W/L, both the poor player and better players get a point, achieved the same result and are equally skilled. The poor player therefore moves up the ranks and the cycle continues. Was that player matched up again with better players? If so, they stand a good chance of ranking up again, even though they haven't worked for it.

 

The problem is two-fold: how many lesser skilled players did a good player get matched up with? That plays a central role in whether the good player will rank up or down from that game. You can see here that the expected outcome of a game is based on the true skill average of the players involved. If a good player has been matched up with players who's average true skill is less than that of the opposing team, the good player is penalized and will rank down with those lesser players. That's where the problem lies: the true skill of a player and the skill of the player as determined by the W/L ranking system are not one and the same as they ideally should be and could be very different. This ends up penalizing players based on factors they can't control. It's also bad for everyone involved: lesser skilled players continually get creamed and more skilled ones continually get dragged.

 

An individual performance-based system is needed to address these problems. So what's the ideal way to implement it? A good way will address the old problems while not introducing any new problems which are worse than the old ones. A good K/D system is not black and white where it's literally based purely on a binary K/D, important things like assists should count towards a players measured contribution in a match.

 

In contrast to W/L where there is 1 point at stake for each player based on a match win or loss, in an ideal K/D system there are 50 zero-sum points (1 for each kill) up for grabs for the players on each team. When a player is killed, they lose 1 point and that point is divided up between the players who contributed to that kill based on how much each player contributed. For example, if you inflicted 80% of the damage to an opponent and then a team mate finishes off the kill, you get +0.80, your team mates gets +0.20 and the target goes -1.00. Therefore players are always rewarded for contributing to kills and discouraged from handing over deaths to the opponent. If you got 3 assists that totaled to +2.10 without getting a kill and then got killed, you would still be +1.10.

 

At the end of the game once a team has registered 50 kills, each player will gain or lose rank equal to their final point total. If you went +2.30, that's what is added to your rank (a running total). If you went -1.50, you lose that amount. This avoids the problems that a fixed reward system based on whether you go positive or negative would have, such as players trying to get a single kill and then hiding until the game ends. Doing that is futile and not rewarded under this system because the more assists and kills you get, the faster you will rank up. Hiding also forfeits map control so you risk opponents getting power weapons and killing you. It's a much better and more profitable idea to win power weapons and map control to dominate. Additionally, you could also implement a +1 to a non-decrementing counter of won matches in this mode which contributes to cosmetic accolades, much like Halo 3's EXP system.

 

The advantages of this system are easily apparent. The good players on each team will rank up for a good performance and the players who were a liability for their team will rank down. Players in this mode are matched up with others based on their running totals, leading to a much more level playing field for everyone. That's pretty much as fair as you can get and solves a ton of problems inherent in both the W/L system and the black-or-white K/D system.

  • Upvote (+1) 2
  • Downvote (-1) 2

Share this post


Link to post

People on their first placement matches should never match diamond+ players. Seems like nobody agrees with me, but I will never be convinced otherwise. The diamond player is put at a huge disadvantage having people that don't even know weapon spawns let alone power ups. Negates the point of having the ranking if you are playing with or against someone that the game has zero information about. 

  • Upvote (+1) 4

Share this post


Link to post

People should only be place in Bronze, Silver or Gold, and have to rank up from there. I see far too many 1500 Onyx players that go neg every game. People should not be placed in the second highest rank and not allowed to derank from there. 

 

I understand and support not deranking into the previous division, but this when combined with the placement system results in people getting into ranks they should not have been. 

This sums up the main complaint I have about the current ranking system.

 

That and the fact that matchmaking at Champion level are far too strict... it should match you with almost any Onyx if you can't find people right off the bat. Instead it kicks you back to the menu 5+ times until you give up on searching.

  • Upvote (+1) 2
  • Downvote (-1) 1

Share this post


Link to post

I don't agree with the ranking system at all. Even when I lose games the other team is getting absolutely reamed by me even in diamond division.  I understand that win/loss means everything, but I think K/D should be weighted more. If you're losing games and you're still going positive 10-15 heck even 20, but if you go -20 in a game and win you go up? The opposition is clearly not on your skill level and you have bad teammates. So what does the ranking system do? It ranks you down so you still play the clueless and even worse people. What an ingenious idea 343i.  Now if I'm playing with a to4 with onyx players, the games are more competitive and balanced. 

  • Upvote (+1) 1
  • Downvote (-1) 4

Share this post


Link to post

Dude, you dont get enough negs talking about your dumbass ranking system ideas in the h5 thread? You have to make an entire new thread to amass your negs in?

  • Downvote (-1) 8

Share this post


Link to post

So less than 100 players have placed into bronze since Halo 5 launched. And overall less than 32,000 players have been ranked in the Team Arena playlist. This game is more doomed than I thought competitively.

Halotracker only tracks those who have signed up on the website.

Share this post


Link to post

Halotracker only tracks those who have signed up on the website.

 

That's straight from the leaderboards.  I never signed up for that site and my name is still on all the leaderboards.

  • Upvote (+1) 1

Share this post


Link to post

That's straight from the leaderboards.  I never signed up for that site and my name is still on all the leaderboards.

 

People can be added by others. 

Share this post


Link to post

People can be added by others. 

 

So you're saying if I go create a brand new random gamertag and play 1 arena match I won't show up on that list?  That list looks damn complete to me but we'll find out by the end of the day.

  • Upvote (+1) 1

Share this post


Link to post

So you're saying if I go create a brand new random gamertag and play 1 arena match I won't show up on that list?  That list looks damn complete to me but we'll find out by the end of the day.

 

Pretty certain. I just searched a few gamertags of friends who have been playing with me since day 1, but are unlikely to have been on the site. 

Once I searched for them, the rolling KD graphs showed that today was their first day being tracked by the site. 

  • Upvote (+1) 1

Share this post


Link to post

I personally find the OP's ideas interesting and plausible.

 

However, if 343 would just allow people to fall out of a division or start everyone at silver or gold or both I think the problem is solved.  I personally would rather face a deranker than have unbalanced games.  There has got to be a way to track a deranker from a statistical standpoint and ban them.

  • Upvote (+1) 1

Share this post


Link to post

I dislike the fact that I will never make it to champion even if I was good enough, simply because i don't put enough games in daily.

  • Upvote (+1) 1
  • Downvote (-1) 1

Share this post


Link to post

So you're saying if I go create a brand new random gamertag and play 1 arena match I won't show up on that list? That list looks damn complete to me but we'll find out by the end of the day.

No. You wont. Ive had alt accounts that I had to plug in on my own.

Share this post


Link to post

I dislike the fact that I will never make it to champion even if I was good enough, simply because i don't put enough games in daily.

That's not true. I haven't played on my main in 3-4 days, and I think I have only played 130 games on it since H5 launched. If you run with a decent team and win most games, champion is very easy to get. because unfortunately, your full team will match randoms a lot.

Share this post


Link to post

That's not true. I haven't played on my main in 3-4 days, and I think I have only played 130 games on it since H5 launched. If you run with a decent team and win most games, champion is very easy to get. because unfortunately, your full team will match randoms a lot.

So you're promoting the cheap way of matching people and getting free wins? Just run in with a team and match randoms all day. Sorry but I have a full time job, married with a kid. I don't have the time or resources to play with find a to4 that gets on the same time that I get on for the limited time I have.  I don't understand why 343i decided to go down this route?  I'm not bashing you or anything, because I would play with a to4 all the time too.  The jist of my statement is: you shouldn't have to go in with a party all the time to get competitive games, that's the matchmaking system that fails when games are lopsided and individuals are straight up pooping on the other team.

 

If I go 30-1 in a  game and lose an objective game, why am I punished?  Your opposition is clearly not good enough, but lets rank you down and play the same people fair enough.  I understand there would be a gargantuan formula for raising someone's level despite losing in a objective gametype just going by stats.  The game just isn't fun dominating people or playing a to4 going in by yourself or with a buddy.  There's way too many games when I go off and straight up dominate people. It feels like Halo 4 again with no ranks.

  • Upvote (+1) 1
  • Downvote (-1) 2

Share this post


Link to post

So you're saying if I go create a brand new random gamertag and play 1 arena match I won't show up on that list?  That list looks damn complete to me but we'll find out by the end of the day.

In order for a gamertag to appear on the leaderboards, it has to have been searched. Their stats would have to have been viewed, in order to be put into their Database. Leaderboards are based off of data from their Database, not 343's.

 

With that said, anyone can search for anyone, which is probably how most of the site was populated. Lots of people don't care too much about looking at their stats and because of that, we will never really see the full list.

  • Upvote (+1) 1

Share this post


Link to post

In order for a gamertag to appear on the leaderboards, it has to have been searched. Their stats would have to have been viewed, in order to be put into their Database. Leaderboards are based off of data from their Database, not 343's.

 

With that said, anyone can search for anyone, which is probably how most of the site was populated. Lots of people don't care too much about looking at their stats and because of that, we will never really see the full list.

 

Using the API we can see all of the gamertags that have played in the last 25 matches of a single player.  Using this API call on 1 GT can net a result of over 500 gamertags.  We could just run that on all of the known gamertags on halotracker and then re-input back into the tracker.  We're bound to find them all or close to all of them based on the last 25 matches of all known players in halotracker currently.  I'm sure someone could build an actual leaderboard pretty quickly if we had someway to export and import GT's to halotracker.

  • Upvote (+1) 1

Share this post


Link to post

So less than 100 players have placed into bronze since Halo 5 launched.  And overall less than 32,000 players have been ranked in the Team Arena playlist.  This game is more doomed than I thought competitively.

 

only 32k players have been ranked in arena? That seems really low....

Share this post


Link to post

Using the API we can see all of the gamertags that have played in the last 25 matches of a single player. Using this API call on 1 GT can net a result of over 500 gamertags. We could just run that on all of the known gamertags on halotracker and then re-input back into the tracker. We're bound to find them all or close to all of them based on the last 25 matches of all known players in halotracker currently. I'm sure someone could build an actual leaderboard pretty quickly if we had someway to export and import GT's to halotracker.

This is true. Get to work.

Share this post


Link to post

This ranking system straight up does not work for SWAT. I've had 3 games tonight where I was the only positive player on my team.

 

I went 22 for 12 and we lost 49-50.

 

23 for for 15 and we lost 47-50.

 

22 for 9 and we lost 49-50.

 

Why am I getting de ranked for these games? How does that make any sense? How is it in any way an accurate reflection of my skill level?

 

Win/loss is great for objective game modes, doesn't make any sense in straight up slayer.

  • Upvote (+1) 2
  • Downvote (-1) 1

Share this post


Link to post

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×

Important Information

By using this site, you agree to our Terms of Use & Privacy Policy.