Jump to content

- - - - -

Mwo Ranking System


24 replies to this topic

#1 Crankshaft87

    Rookie

  • Knight Errant
  • 3 posts
  • LocationGermany, Bavaria, Ingolstadt

Posted 09 June 2014 - 02:37 AM

Hi Pilots :)

Im not sure, but i can't find a reliable information about a ranking-system in MWO. Does it have one, or does it not have one? Is it just randoms playing against randoms, now matter if they play one day or one year? It would be nice, if anyone could teach me that ;)

Greetz

#2 Appogee

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 10,966 posts
  • LocationOn planet Tukayyid, celebrating victory

Posted 09 June 2014 - 02:41 AM

I would be happy to welcome and help a new MechWarrior.

There is a hidden ranking system in MWO, which is based on the Elo system. Each player has 4 personal Elo scores - one in each of the 4 weight classes of Mechs - which range from 0 (lowest) to 2500 (highest). A player's personal Elo in a weight class is ratcheted up or down based on his/her team's wins and losses.

To try to create a relatively even match, MWO's matchmaker adds up the 12 personal Elos of the individuals on your team, to come up with a team average Elo. It then tries to match your team against another team which has a similar 12-person average Elo.

Important note: this is NOT the same as matching 12 people who have the same average Elo. A team often comprises people with quite different individual Elos.
  • When you first start playing, your personal Elo for each weight class starts out in the middle of the range: 1500.
  • Your personal Elo for a particular weight class is increased a little whenever your team beats another team which had a higher average Elo.
  • Your personal Elo for a particular weight class is decreased a little whenever your team is beaten by another team which had a lower average Elo.
Your personal Elo is not affected at all by your individual performance. Ie. It doesn't matter how many Mechs you personally kill, how much damage you personally do, etc. The only thing that changes your personal Elo is your team's success or failure in matches. And even then, only when you beat an ''on average better'' team, or lose to an "on average worse" team.


The matchmaker also ensures there are the same number of pre-made lances on each team, and will soon enforce three of each of the four Mech weight classes on each team (called "4x3").

Edited by Appogee, 09 June 2014 - 03:08 AM.


#3 Egomane

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • 8,163 posts

Posted 09 June 2014 - 02:43 AM

There is an Elo system in place, but the values are not visible.

Every pilot has four Elo values. One for light mechs, one for mediums, one for heavies and one for assaults.

Rookie pilots, while still on their first 25 matches have an Elo value that is artificially reduced by 200 points.

So yes, new and old player can and will be in the same match, if they are in the same Elo range, which can be rather wide, depending on the search time. The longer it takes to find a match, the higher the probability to have more extreme values playing with each other. If you drop near instantly, all the players will be on the same level, or at least very close to it, as yourself.

There are no ladders or other public ranking systems.

#4 Modo44

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bad Company
  • 3,559 posts

Posted 09 June 2014 - 02:48 AM

Player skill is tracked, currently separately for each mech weight class. The mechanism used in MWO is based on the Elo system (your skill values go up/down depending on wins/losses as compared to the expected match outcome). It is not great, but it generally works for players close to the average skill level. Elo values are not accessible to players at the moment.

#5 Zerberus

    Member

  • PipPipPipPipPipPipPipPipPip
  • Overlord
  • Overlord
  • 3,488 posts
  • LocationUnder the floorboards looking for the Owner`s Manual

Posted 09 June 2014 - 02:50 AM

View PostEgomane, on 09 June 2014 - 02:43 AM, said:

.... If you drop near instantly, all the players will be on the same level, or at least very close to it, as yourself.

Actually, the "instant" drops are generally the ones I feel are most mismatched. Likely due to others already searching for 5 minutes and then finally being able to complete the group when I / my group start our search.

The system "unfortunately" works both ways. :)

Edited by Zerberus, 09 June 2014 - 02:50 AM.


#6 Egomane

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • 8,163 posts

Posted 09 June 2014 - 02:55 AM

View PostZerberus, on 09 June 2014 - 02:50 AM, said:

Actually, the "instant" drops are generally the ones I feel are most mismatched. Likely due to others already searching for 5 minutes and then finally being able to complete the group when I / my group start our search.

The system "unfortunately" works both ways. :)

Well... this might be a case of confirmation bias. But as I didn't program the matchmaker, I'll simply say I believe you, but my own experience is telling me otherwise. ;)

#7 Zerberus

    Member

  • PipPipPipPipPipPipPipPipPip
  • Overlord
  • Overlord
  • 3,488 posts
  • LocationUnder the floorboards looking for the Owner`s Manual

Posted 09 June 2014 - 03:09 AM

^^ It might, but it`s actually kind of logical that is someone is waiting 5 minutes for a match, somneone is going to have an instant connect .. wheras the chances of instantly finding 20 people well matched to you are (at least in theory) significantly lower .

But as you say, it`s anecdotal evidence either way :)

#8 Appogee

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 10,966 posts
  • LocationOn planet Tukayyid, celebrating victory

Posted 09 June 2014 - 03:12 AM

View PostModo44, on 09 June 2014 - 02:48 AM, said:

Player skill is tracked,
But it's not... unless you believe that:

1. a player's personal skill is accurately reflected by the collective success/failure rate of the 11 random people he/she drops with from match to match; or

2. a player's personal skill is the same as that of the other players in the premade he/she consistently plays with.

Both assumptions would be illogical, based on the facts of how Elo is implemented in Mwo (as I documented above).

The only genuine assessment of individual skill that we have is our ranking in tournaments. Even those results are quite influenced by individual tournament rules, the preparedness/reticence of players to spend up big on consumables, player proximity to the game servers (due to hit detection issues caused by lag), how much time a player has to grind away to achieve a ''10 best" score in a match, and sync droppers who create hidden teams in individual tournaments for personal advantage.

MWO's implementation of Elo maybe gives a rough indication of personal skill over a lot of matches and a long period of time, on the basis that personal performance will have a general impact on team results over the long haul. However, given all the other factors it takes account of outside of the payer's control, at best it would be very approximate, perhaps only accurate at the quartile level.

Edited by Appogee, 09 June 2014 - 03:27 AM.


#9 Modo44

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bad Company
  • 3,559 posts

Posted 09 June 2014 - 03:58 AM

True, Elo works worse than for 1 on 1 games, but it does work. Ask any good player who made a new alt account. **** is crazy in the lower brackets. Every aspect of your capabilities as a player is taken into account. This includes things not seen in stats, like reacting to your team and the situation, or efficient PUG herding.

Also, take a chill pill. If MWO is giving you such an angry *****, maybe it is time to relax elsewhere.

Edited by Modo44, 09 June 2014 - 04:00 AM.


#10 Crankshaft87

    Rookie

  • Knight Errant
  • 3 posts
  • LocationGermany, Bavaria, Ingolstadt

Posted 09 June 2014 - 04:49 AM

Okay, so i see we have a hidden ranking. Ive been playing MW since part 2 came out :) Before i discovered the awesomeness of MWO i was playing Starcraft 2 pretty much. My new question now is, when we have a elo, why not show it? Not that i want to know my points, but with increasing points you could just something do with like in Starcraft. Ranks, that show, that you have to be somewhere between 500-750 for example. Just like: Leagues. The rating can be hidden, thats okay, but an indicator like a rank or something would be really nice. It would push players to become better i guess.

Thank you very much for your very informative and fast replies ;)

#11 Egomane

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • 8,163 posts

Posted 09 June 2014 - 04:55 AM

It's not shown, primarily to prevent units from selecting members by Elo. Even if the value is only visible to the user it belongs to, they could demand screenshots or you are not allowed to join. Such elitism isn't really a good thing, because the new and not so good players will have a harder time to find a unit for themself.

It would also cause problems when arguing about various concepts of the game, when the view of player with the lower or hidden Elo is dismissed because of his supposedly lower understanding of game mechanics.

#12 Modo44

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bad Company
  • 3,559 posts

Posted 09 June 2014 - 04:58 AM

Displaying Elo would make a lot of forum warrios shut up for good. Nobody wants their e-peen cut off.

#13 White Bear 84

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 3,857 posts

Posted 09 June 2014 - 04:59 AM

View PostEgomane, on 09 June 2014 - 04:55 AM, said:


It would also cause problems when arguing about various concepts of the game, when the view of player with the lower or hidden Elo is dismissed because of his supposedly lower understanding of game mechanics.


Not to mention in game flaming and the sort. Too many players out there that will watch a pilot get killed and give them grief over it because they made a mistake - having ELO displayed would only add fuel to that fire...

#14 Modo44

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bad Company
  • 3,559 posts

Posted 09 June 2014 - 05:06 AM

You know what happens when everybody thinks they are the ****? The actually good players get flamed like so and there is no way for "Let us discuss average Elo meta" threads because nobody knows who is actually playing in which bracket.

#15 Appogee

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 10,966 posts
  • LocationOn planet Tukayyid, celebrating victory

Posted 09 June 2014 - 05:11 AM

View PostModo44, on 09 June 2014 - 03:58 AM, said:

Also, take a chill pill. If MWO is giving you such an angry *****, maybe it is time to relax elsewhere.

I'm not angry. I can't even imagine why you would suppose that, when all I write was a mere factual correction to what you posted, in the interests of the OP having clarity on the question he asked.

View PostModo44, on 09 June 2014 - 03:58 AM, said:

Every aspect of your capabilities as a player is taken into account. This includes things not seen in stats, like reacting to your team and the situation, or efficient PUG herding.

I'm not sure what you mean here. The only thing that's actually ''taken into account'' is whether your PUG team had a lower average Elo than your enemy, and whether you won or lost. That's it.

Sure, our abilities to personally carry a team of low Elo players to victory can have a contributing influence on the outcome. But it's mostly not the deciding factor. Teams ignore your coaching, or otherwise prove incapable of winning. Besides, half your matches you start out with a higher Elo, and even a win results in no change to your personal Elo.

To grind up your Elo from starting point (which I have done with a second account by the way) you need to drop with a team that proves to be more competent that your opponents, but with a lower average Elo. That is the only way to improve your Elo.

Edited by Appogee, 09 June 2014 - 05:14 AM.


#16 Modo44

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bad Company
  • 3,559 posts

Posted 09 June 2014 - 05:13 AM

As many others, you fail to understand that winning translates to all skills combined. Every matchmaker that tries to track separate values instead of Elo fails even more spectacularly than this one, because is it guaranteed to not account for all possible factors.

Edited by Modo44, 09 June 2014 - 05:14 AM.


#17 Appogee

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 10,966 posts
  • LocationOn planet Tukayyid, celebrating victory

Posted 09 June 2014 - 05:24 AM

View PostModo44, on 09 June 2014 - 05:13 AM, said:

As many others, you fail to understand that winning translates to all skills combined. Every matchmaker that tries to track separate values instead of Elo fails even more spectacularly than this one, because is it guaranteed to not account for all possible factors.

No-one in this thread has either discussed or proposed alternative matchmaking approaches. In this thread, we're trying to help the OP understand how individual ranking actually occurs and is used in matchmaking in MWO.

If you can factually describe how an individual's "all skills combined'' are taken into account in the actual system currently used in MWO - and how this has a bigger effect on individual Elo that the relative capabilities of the other 11 players on their team - then that would helpfully advance the thread.

Quid Pro Quo, if you need me to, I can give you specific examples of how the system currently used in MWO leads to individuals having Elo ratings which are either too high relative to their skill, or too low. In both cases, their ''all skills combined'' have less to do with their actual Elo rating, and the primary determinant is the capabilities of the other players they team with.

Edited by Appogee, 09 June 2014 - 05:30 AM.


#18 Itsalrightwithme

    Member

  • PipPipPipPipPipPip
  • Survivor
  • 391 posts
  • LocationCambridge, MA, USA

Posted 09 June 2014 - 06:04 AM

View PostModo44, on 09 June 2014 - 05:13 AM, said:

As many others, you fail to understand that winning translates to all skills combined. Every matchmaker that tries to track separate values instead of Elo fails even more spectacularly than this one, because is it guaranteed to not account for all possible factors.


I think Appogee is speaking of players who get pulled up or down the elo ladder through playing mostly with a premade group. This is reality. Try grouping up with a very new player (or a not so good one) such that your group elo is lowered to the elo level of a baby hippo. If you win a lot of matches, and you can win a lot of matches, the new player's elo is dragged up while yours may not be changed a lot. The next time the new player does a solo drop .... well, it may not be pleasant.

#19 Dolph Hoskins

    Member

  • PipPipPipPipPipPip
  • The Territorial
  • The Territorial
  • 499 posts
  • LocationThe Machine

Posted 09 June 2014 - 06:28 AM

Whatever ranking system they decide to develop I hope they keep the personal statistics hidden, voluntarily, and selectively shown.

If I am wrong about how it currently is please forgive me as I am quite new, but I can say based off of past experience with games that as soon as personal stats are flagged about everywhere all gameplay becomes about what numbers can "I" get. This is a squad level game and I feel most everything about it's public gauging should lean towards overall team outcomes.

Just my initial thoughts on the matter. It would be neat to do a kind of profile ID code you can create kind of like EVE online does so it is possible to share personal stats in an official way for those it concerns, like clans, private showboating and whatnot.

#20 Itsalrightwithme

    Member

  • PipPipPipPipPipPip
  • Survivor
  • 391 posts
  • LocationCambridge, MA, USA

Posted 09 June 2014 - 06:54 AM

There is already a lot of showboating and vanity, unfortunately. Players hide at the end of a match to save their precious KDR. Some disconnect when they think they are going to die, thinking the death won't count. For some time this was true, but they changed it waaay back in beta.

Some teams ask for a screenshot of one's stats, but the better teams know what's most important is how the pilot contributes to the team. So most just go by reputation and recommendation, since it's a small community after all. Anybody who did competitive league play prior to lobbies know of mis-syncs, and they tend to pad one's deaths.

Ultimately, stats do not differentiate between performance while pugging versus while in a 4-man versus while in 12-man.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users