37 replies to this topic

[Deamonition]

First I'm not going to pretend that I know exactly how MM works in MWO. But, lately, or since a couple of months, I have been teamed with players that were clearly not at the same level as I am. Don't get me wrong, I'm not a pro, probably decent, that's it. I'm mostly referring to the fact that I was teamed with brand new players.

Most of them were really having a hard time against experienced players and were getting stomped.

I also have been teamed with players that were not new, but not so good. Actually, they were very bad. It's quite easy to notice when someone is still having a hard time with this game. Having a look at what they do when you are dead is one way.

Is this because of the size of the players base?

Anybody has some input on how MM treats new players and also how does it splits good or bad players?

[SpuriousLogic]

New players come in to ELO at the middle of the curve - which is most likely way above their ability to start. So if you are a middle of the road player (most of us are), you will tend to get a bunch of new players.

Devs had previously said that they are thinking of lowering the ELO of new players. Think that may have been in the latest Ask the Devs.

[Deamonition]

Well if the game considers me as average, there is no way a new guy is also average. Some are better than others of course, but I do better than most of them.

Moreover, how does the game knows you are average? Is it based on K/D ratio? For example, mine is a bit below 2, so it's average?

[DemonRaziel]

Deamonition, on 12 April 2013 - 10:25 AM, said:

Moreover, how does the game knows you are average? Is it based on K/D ratio? For example, mine is a bit below 2, so it's average?

There is a separate ELO for every weight category, as far as I know. So it's not based on your overall KDR. And personally I hope it's not based solely on the KDR for a given weight category, either.

[Deamonition]

DemonRaziel, on 12 April 2013 - 10:35 AM, said:

There is a separate ELO for every weight category, as far as I know. So it's not based on your overall KDR. And personally I hope it's not based solely on the KDR for a given weight category, either.

Right, you don't kill a lot with a Jenner.

[ohtochooseaname]

Win/loss ratio, since they started the new tracking system, is probably a more accurate indicator of skill. Basically, because the system can't always find good matches, if you are higher up on the curve, you'll tend to be thrown into groups of people with less actual skill than you, making your skill into an increased chance of a win. This will give you a higher win/loss ratio, in general. K/D ratio is not part of the matchmaking system, whereas wins and losses are. Obviously, the game mechanics of having specific brackets,, etc. changes things a bit, but I think this is generally true.

And yes, if you are middle of the road, you will see some fresh noobs until they get dropped to a lower bracket, but they should be relatively few. You will also see some very unskilled players just because the matchmaker is set up to prefer matches to start in under 2 minutes from when it started trying to populate it.

Edited by ohtochooseaname, 12 April 2013 - 10:40 AM.

[Deamonition]

And I don't want to sound cocky. I'm sure I'm not alone in that situation.

I, myself, was not so good when I started playing.

[Deamonition]

ohtochooseaname, on 12 April 2013 - 10:37 AM, said:

Win/loss ratio, since they started the new tracking system, is probably a more accurate indicator of skill. Basically, because the system can't always find good matches, if you are higher up on the curve, you'll tend to be thrown into groups of people with less actual skill than you, making your skill into an increased chance of a win. This will give you a higher win/loss ratio, in general. K/D ratio is not part of the matchmaking system, whereas wins and losses are. Obviously, the game mechanics of having specific brackets,, etc. changes things a bit, but I think this is generally true.

And yes, if you are middle of the road, you will see some fresh noobs until they get dropped to a lower bracket, but they should be relatively few. You will also see some very unskilled players just because the matchmaker is set up to prefer matches to start in under 2 minutes from when it started trying to populate it.

Well while it makes some sense that it works like that, I don't fully agree.

You can't win it by yourself... I had a decent win-loss ratio, but lately it dropped drastically. I just can't carry teams by myself..

It's 8 vs 8,not 3v3... So it's difficult to carry a full team

Edited by Deamonition, 12 April 2013 - 11:08 AM.

[Jess Hazen]

MM in MWO is completely rudimentary it has no brackets from which it can draw like skilled players instead it claims to even the odds by evening the teams the best an elementary system can do by averaging ratings across the teams and it does this from the entire player population...

This current MM system is not by any means fair to anyone, skilled and not yet skilled players alike are all the victims of this system. The current MM installation should be scrapped or patched to a real system asap.

[Deamonition]

Jess Hazen, on 12 April 2013 - 11:04 AM, said:

MM in MWO is completely rudimentary it has no brackets from which it can draw like skilled players instead it claims to even the odds by evening the teams the best an elementary system can do by averaging ratings across the teams and it does this from the entire player population...

This current MM system is not by any means fair to anyone, skilled and not yet skilled players alike are all the victims of this system. The current MM installation should be scrapped or patched to a real system asap.

Sadly, I must agree

[Deamonition]

Moreover, I see a lot of complaining about matchmaking. Anybody knows if PGI said they would look into it?

[DemonRaziel]

They just released a Command Chair post regarding some changes in MM. Apparently there was an error in the ELO calculation that caused many ELOs to be higher than they should have been.

[Deamonition]

DemonRaziel, on 12 April 2013 - 11:55 AM, said:

They just released a Command Chair post regarding some changes in MM. Apparently there was an error in the ELO calculation that caused many ELOs to be higher than they should have been.

Let's hope that the problem doesn't spread to everyone, cause then everybody's elo will shift down and nothing will have changed!

[ohtochooseaname]

Deamonition, on 12 April 2013 - 10:47 AM, said:

Well while it makes some sense that it works like that, I don't fully agree.

You can't win it by yourself... I had a decent win-loss ratio, but lately it dropped drastically. I just can't carry teams by myself..

It's 8 vs 8,not 3v3... So it's difficult to carry a full team

Depends on what you're in. In my experience, the heavier the mech, the easier it is to carry a team (unless it's alpine or Tourmaline). Also, ECM plays a big factor.

I solo pug, and I'm 17 and 4 in my JM6-A.

Regardless, actually, my point was more that your recent win/loss (say, for the past week/month) rate will be statistically impacted by your skill. If you are on average, better than the people you play with, you will, on average, have more wins (you are effectively the only factor changing a 50-50 win rate up or down, if you pug). So, a win/loss ratio greater than 1 indicates you are likely being paired up with worse players (including noobs). A statistical analysis can tell you whether or not your win rate is actually significant. For example, a 3 games to 2 games win/loss ratio is not statistically significant, whereas 3000 games to 2950 games is, even though the ratio is much smaller.

I am guessing most people have positive win rates simply because new players come in at the same spot as average players (now that there is a base population for their analysis, they can actually throw the new players in a lower bracket. Otherwise, whatever start point there is becomes the new average player value). Since the new player statistically lowers their team's chance of winning, they'll tend to lose. Since player retention isn't 100% and it's likely much worse for new players, they'll tend to leave the game more often with net losing records. Therefore, people who play for a while will tend to have positive win rates. This, however, becomes much less significant in higher brackets.

Deamonition, on 12 April 2013 - 12:03 PM, said:

Let's hope that the problem doesn't spread to everyone, cause then everybody's elo will shift down and nothing will have changed!

They're basically recalculating ELO, not reducing or increasing it. They're basically taking all the games you've played, with your starting ELO score, and recalculating the results for every single game to get people where they should be.

[Deamonition]

ohtochooseaname, on 12 April 2013 - 02:07 PM, said:

Depends on what you're in. In my experience, the heavier the mech, the easier it is to carry a team (unless it's alpine or Tourmaline). Also, ECM plays a big factor.

I solo pug, and I'm 17 and 4 in my JM6-A.

Regardless, actually, my point was more that your recent win/loss (say, for the past week/month) rate will be statistically impacted by your skill. If you are on average, better than the people you play with, you will, on average, have more wins (you are effectively the only factor changing a 50-50 win rate up or down, if you pug). So, a win/loss ratio greater than 1 indicates you are likely being paired up with worse players (including noobs). A statistical analysis can tell you whether or not your win rate is actually significant. For example, a 3 games to 2 games win/loss ratio is not statistically significant, whereas 3000 games to 2950 games is, even though the ratio is much smaller.

I am guessing most people have positive win rates simply because new players come in at the same spot as average players (now that there is a base population for their analysis, they can actually throw the new players in a lower bracket. Otherwise, whatever start point there is becomes the new average player value). Since the new player statistically lowers their team's chance of winning, they'll tend to lose. Since player retention isn't 100% and it's likely much worse for new players, they'll tend to leave the game more often with net losing records. Therefore, people who play for a while will tend to have positive win rates. This, however, becomes much less significant in higher brackets.



They're basically recalculating ELO, not reducing or increasing it. They're basically taking all the games you've played, with your starting ELO score, and recalculating the results for every single game to get people where they should be.

First, I would not say that it is easier to carry with a bigger mech. From my experience, when taking my Atlas, everybody tends to focus me. You know... I'm big, so I'm a nice target. Having the entire enemy team focusing you makes it difficult to carry.

Also, while it would seem logical enough that being better than who you are teamed with you should win more. But that isn't true. The reason being exactly what you said. New people decreases the chances of their team to win. So it's impossible that when you look at it from the better player point of view, he's got better chances of winning, but when looking at it from the worst players point of view, in the same team, they would have lower chances of winning.

The team chances of winning depends on the sum of the chances of each players that composes it (it's not really a sum per say, but whatever)

Where I am going with that is: a good player teamed with bad players should not be considered as having higher chances of winning. If he's teamed with 7 complete noobs and going against a better team, well those noobs are just bringing him down.

That problem comes from the fact that people tend to see ELO in such a game as behaving like it would in a 1v1 game (chess for example).

Because it's an 8v8 game, it's way more complicated then just who is good and who is bad. It's pretty much how many good and bad players per team and how "much" good or bad they are. That's how MM should work.

Edited by Deamonition, 12 April 2013 - 02:28 PM.

[ohtochooseaname]

spuriouslogic, on 12 April 2013 - 10:13 AM, said:

New players come in to ELO at the middle of the curve - which is most likely way above their ability to start. So if you are a middle of the road player (most of us are), you will tend to get a bunch of new players.

Devs had previously said that they are thinking of lowering the ELO of new players. Think that may have been in the latest Ask the Devs.

This may not actually be possible, and will likely be quite difficult. ELO is a zero sum system. For someone to go up, someone else must go down. Due to this fact, when you start an ELO system, whatever the starting score for new players, that is the average player's score. So, naturally, new players come in at that level.

To offset this, however, new players are much more likely to lose a bunch of games, and then quit, whereas retention of average and better skilled players is much higher. This causes a net ELO inflation, which allows the developers an opportunity: They can now start new players lower down on the totem pole by the amount of the inflation...which, incidentally, would be starting them at 1300 (since everyone else is boosted up by the inflation). In other words: this problem may work itself out with time as the ELO gets more inflated until the point in which new players start out in a bracket, which is just as likely to stay as leave.


This, however, can lead to abuse: if someone with high skill and a high rating creates a new account, they can rapidly win games up to nearly their actual skill level in relatively little time (less than 50 games, especially if a group of them do this), then start over. This would rapidly bring everyone down. Actually, come to think of it, the opportunity for this kind of abuse currently exists in the game such that a determined team could effectively make new players come in at a higher bracket. it's a crap-ton of work, though.

[Deamonition]

All I mean is, you put 8 super bad players against 8 super bad players and they will have fun, it's gonna be a close game. Put 8 super good against 8 super good, same thing.

2 super bad, 4 average and 2 slightly good against the exact same team... Same conclusion.

It's just that right now it looks pretty random. 1 good with 7 noobs going against 4 average and 4 noobs. That is not a fair game

[Syllogy]

ERMAGERD! Fix coming Tuesday

[Deamonition]

ohtochooseaname, on 12 April 2013 - 02:29 PM, said:

This may not actually be possible, and will likely be quite difficult. ELO is a zero sum system. For someone to go up, someone else must go down. Due to this fact, when you start an ELO system, whatever the starting score for new players, that is the average player's score. So, naturally, new players come in at that level.

To offset this, however, new players are much more likely to lose a bunch of games, and then quit, whereas retention of average and better skilled players is much higher. This causes a net ELO inflation, which allows the developers an opportunity: They can now start new players lower down on the totem pole by the amount of the inflation...which, incidentally, would be starting them at 1300 (since everyone else is boosted up by the inflation). In other words: this problem may work itself out with time as the ELO gets more inflated until the point in which new players start out in a bracket, which is just as likely to stay as leave.


This, however, can lead to abuse: if someone with high skill and a high rating creates a new account, they can rapidly win games up to nearly their actual skill level in relatively little time (less than 50 games, especially if a group of them do this), then start over. This would rapidly bring everyone down. Actually, come to think of it, the opportunity for this kind of abuse currently exists in the game such that a determined team could effectively make new players come in at a higher bracket. it's a crap-ton of work, though.

Well you ask me it shouldn't be an ELO system at all. Correct me if I'm wrong, but ELO was first introduced with Chess and this is NOT a teamplay game.

Somebody somewhere thought it would be a great thing to use in teamplay video games, but it sucks. I've always found it a pain in other games and it will always be.

It should really work with ranks or something. You start at 0 and you go up. There would be some mathematics behind it, but people at 0 would be either bad or new. If the new guys are any good, then they will quickly go up in ranks and end up in a "rank bracket" that fits their skill/quality. At that point, they would start playing for .500 and therefore should not go higher in ranks.

[ohtochooseaname]

Deamonition, on 12 April 2013 - 02:26 PM, said:

Also, while it would seem logical enough that being better than who you are teamed with you should win more. But that isn't true. The reason being exactly what you said. New people decreases the chances of their team to win. So it's impossible that when you look at it from the better player point of view, he's got better chances of winning, but when looking at it from the worst players point of view, in the same team, they would have lower chances of winning.

The team chances of winning depends on the sum of the chances of each players that composes it (it's not really a sum per say, but whatever)

Where I am going with that is: a good player teamed with bad players should not be considered as having higher chances of winning. If he's teamed with 7 complete noobs and going against a better team, well those noobs are just bringing him down.

It's not being better than who you are teamed with, it's being better than the bracket you were dropped into: that, according to what they've said in their posts, is how it works. As far as I can tell, the game creates a game of a bracket, say 1500, then starts pulling in players. The more time it takes to pull in players (based on rating and weight), the looser it is in pulling in player. Therefore, if you are above 1300, you'll tend to get pulled into lower rating games (there are more games below you to be pulled into) than your rating and win more (because your opponents and team mates will still on average, average 1300), whereas, if you are below 1300, you'll tend to get pulled into higher rating games than your rating and lose more. I don't believe it re-balances after it makes the two teams, or my W/L ratio would be much smaller. If it did re-balance then it could possibly set you with a bunch of noobs to offset your score, in which case, you'd tend to lose of that offset were very high because it would assume that you could actually carry a team. I don't believe it does this.

The fact that you are on your team reduces the likelihood that a new player will be on your team, thus, your team should win more frequently. You don't have to carry your team or anything like that: a player on the enemy team is on average as skilled as a player on your team (neglecting people, who just started), so your impact doesn't have to be much to have an effect on your win ratio over a large number of games.