Jump to content

- - - - -

Elo Threshold Adjustment - Poll


168 replies to this topic

Poll: Elo Threshold Adjustment - Poll (385 member(s) have cast votes)

Have you noticed any change in Matchmaking Wait Times today (January 23rd)?

  1. Voted No Change (160 votes [41.56%] - View)

    Percentage of vote: 41.56%

  2. Longer Wait Times (100 votes [25.97%] - View)

    Percentage of vote: 25.97%

  3. Shorter Wait Times (125 votes [32.47%] - View)

    Percentage of vote: 32.47%

Vote

#81 Li Song

    Member

  • PipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 225 posts
  • LocationSweden

Posted 24 January 2014 - 12:27 PM

View Postlowe0, on 24 January 2014 - 10:28 AM, said:

I like this idea, with one additional change: bias the range away from the center of the X-axis. So, for players with above average scores, the system will prefer to match against players with even higher ELO, while for those with below average scores (me) the system will prefer opponents with even lower ELO. This would have a flattening effect on the curve over time, driving players away from the edges and center towards the middle-low and middle-high sections.


This is what my post http://mwomercs.com/...ost__p__3097865 proposes, but I have done the statistics/math for PGI already, they just need to implement. :unsure:

#82 Kraven Kor

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 5,434 posts

Posted 24 January 2014 - 12:31 PM

Really needs something better than ELO.

Fix Match Score, at some point, and replace ELO with average match score or something. Have it take into account as many factors beyond damage as possible - accuracy, damage dealt compared to expected damage by total # of shots, damage per kill (with lower damage = better); as many factors as possible and reasonable.

Then get "battle value" in and balance matches by Pilot Ranking (NOT ELO, but using Magic Secret Rank Formula) and Battle Value.

#83 AdamBaines

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 1,384 posts

Posted 24 January 2014 - 12:35 PM

View Postanubis969, on 24 January 2014 - 07:55 AM, said:



Is this the most recent? Thare have been no changes in determining Elo and matches that they have published? Thanks again.

#84 Ace Selin

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,534 posts
  • LocationAustralia

Posted 24 January 2014 - 02:38 PM

355 ton difference between the 2 teams, game I just played, things are still horrible in the matchmaking dept

#85 Khobai

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 23,969 posts

Posted 24 January 2014 - 02:59 PM

Shorter wait times.

But weight class disparity is still a huge problem. Im personally against tonnage limits, since tonnage limits set an extremely bad precedent that more tonnage is always better, which isnt true at all. For example, a Jagermech is more than a match for most Atlases despite being 35 tons lighter. I do however believe both teams should have the same number of mechs from each weight class.

Rather than tonnage limits we should have weight class matching. Mechs should be divided into five different weight classes and each team should get 2 mechs from each weight class in addition to 2 mechs from random weight classes (one team might get a tiny mech and an assault while the other team gets two mediums, but it would try to balance the randoms by their combined weight).

That system is way more fair since it doesnt make the incorrect assumption that more tonnage is always better.

Edited by Khobai, 24 January 2014 - 03:10 PM.


#86 Appogee

    Member

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

Posted 24 January 2014 - 03:02 PM

View PostMischiefSC, on 24 January 2014 - 12:04 PM, said:

Just in case people are reading this -

Can you look at matching to a range instead of high/low to target? By matching high/low to target you effectively hang rookies off of every veteran to carry in many matches. If the vet isn't in his top mech, dropping with his friends on VOIP he already isn't carrying his max weight in Elo to begin with.

A better match experience might be just matching to a range. This may also give more wiggle room to weight-matching, at least for mid-range matches. For a game at ~2000 Elo for example just weight-match players as close to 2,000 on both teams as possible. The critical piece is that you leave one or two slots on each team open for those outlier high Elo players who don't have full weight balanced matches available at their range. Try to match one or two on each team but still, one high-Elo player has less impact on the results of a 12v12 match than 3 low Elo players.

Even if the variance between total Elo on both teams is 200, 300, even 500 points so long as it's just a byproduct of 20 or 50 points variance between 10 players it's not as significant as having 2 or 3 or 4 players with scores 100, 200, 300 points below the other teams average.

It might give a better experience and result in more balanced actual matches.


^ This, this, a thousand times, this!

#87 SgtKinCaiD

    Member

  • PipPipPipPipPipPipPipPip
  • Overlord
  • Overlord
  • 1,096 posts
  • LocationBordeaux

Posted 24 January 2014 - 03:35 PM

I tried to level my BH tonight. I stopped after 10 matchs. The quality of match is still in the gutter.
Just a few example of tonight :

Posted Image

Posted Image

Posted Image

Posted Image

Posted Image

#88 Paul Inouye

    Lead Designer

  • Developer
  • Developer
  • 2,815 posts
  • LocationVancouver, BC

Posted 24 January 2014 - 03:38 PM

Thanks for the poll responses folks. It's showing us that the change is doing what was expected. While wait times have gone up for the extremely high and extremely low Elo players, the number of failed matches has not increased significantly. This is meaning that quality matches are kicking off before the max Elo range is being reached.

And off topic: Again, Elo has nothing to do with tonnage balancing. We are still seeing teams that are over tonnage over others. This is where the tonnage matching of the CW Launch Module comes into play.

#89 Bors Mistral

    Member

  • PipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 313 posts
  • LocationMontreal

Posted 24 January 2014 - 03:44 PM

I haven't noticed much of a change in the time it takes to find a match.

What I do seem to get a lot more of is being put on a team that's a player or two short at the start of the game.

#90 anubis969

    Member

  • PipPipPipPipPip
  • 130 posts

Posted 24 January 2014 - 05:38 PM

View PostAdamBaines, on 24 January 2014 - 12:35 PM, said:


Is this the most recent? Thare have been no changes in determining Elo and matches that they have published? Thanks again.

That is the most recent detailed explanation of how Elo is calculated that I am aware of. Since Paul made that post they have fiddled around with the Elo and weight thresholds, as well as changing how Elo works for new players. New players now start with an Elo score of 1100 which is fixed for their first 25 games and then it gets reset to 1300.

But that is it for changes that I'm aware of. The core mechanics of how Elo is calculated remains the same to my knowledge.

However if someone more knowledgeable on the subject *cough*Paul*cough* wishes to correct me... *wink wink nudge nudge*

:unsure:

Talking of which, as an update Paul; now that things have settled a bit my matches tonight (24/01/2014) during British prime time were much better balanced than last night (23/01/2014). They weren't as close as they used to be before the previous threshold change but they were close enough to be fun. My wait times are still longer than they were after the previous change but lower overall.

Also my apologies Paul if my comment yesterday was a bit curt but I was rather frustrated after having my match quality go down the pan since the previous change to the Elo threshold. However that is no excuse for rudeness. Hopefully in time the new threshold will prove to be an acceptable middle ground.

Edited by anubis969, 24 January 2014 - 05:39 PM.


#91 Mechsniper

    Member

  • PipPipPipPipPipPip
  • The Formidable
  • The Formidable
  • 457 posts
  • LocationUnited States

Posted 24 January 2014 - 06:38 PM

Tonnage is not your only problem, we need exp based drops, ELO was a horrible idea and is a frequent reason I contemplate finding another game to play that is not borked by 8/10 drops leaving me with what I swear must be the monkeys at the zoo playing on my team. Please drop it. It cannot work well in a TEAM based game with brand new guys dropping with 25 yr players.

#92 Agent 0 Fortune

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 3,403 posts

Posted 24 January 2014 - 11:35 PM

Time to get into match: same as before (fast)
Qualtity of match: same (shtty matching of Elo)
matches based on map mode (conquest, assault, skirmish): no differnce
matches based on tonnnage: running medium only, no differnce in enemy tonnage noticed.
matches based on Elo: stomp, stomp, stomp

conclusion: welcome to the new Matchamaker same as the old.

Note: all recent matches run as Medium mech (note: PGI you have claimed that elo is weight classed based, I predict that my skill level ie. elo level is mediocore in medium mechs, which is why it tried to production/test your eveniornment in mediums rather than Heavy (high elo), light (low elo), or assault (medium elo).

In short. the matchmaker has not changed at all for me. Same unbalanced matches, same time to create. Thanks for nothing.

Edited by Agent 0 Fortune, 24 January 2014 - 11:38 PM.


#93 Agent 0 Fortune

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 3,403 posts

Posted 25 January 2014 - 12:30 AM

I did discover that most of my games (set to random) landed on conquest. and by most I belive it was 90%+

#94 Appogee

    Member

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

Posted 25 January 2014 - 01:03 AM

View PostPaul Inouye, on 24 January 2014 - 03:38 PM, said:

Thanks for the poll responses folks. It's showing us that the change is doing what was expected.


Paul, PLEASE... you are not fixing the main problem.

The matchmaker currently creates teams with equal average Elos. But the average is misleading. Teams often comprise a couple of high Elo players and a bunch of non-competitive players.

If instead the game found 24 players of about the same Elo, and split them into two teams, the games would end up being much more evenly matched.

Edited by Appogee, 25 January 2014 - 01:05 AM.


#95 Agent 0 Fortune

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 3,403 posts

Posted 25 January 2014 - 01:12 AM

Wow a dissapointing evening.
FYI they all important launch times stayed the the same, and match quality can only be defined as " the same dirty toilet paper as we used yesterday and the day before".

I so much want this game to be good, I have a significant emotional (and financial) investment. But the gameplay can on be described as "punative".

#96 Mycrus

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Overlord
  • Overlord
  • 5,160 posts
  • Google+: Link
  • Twitter: Link
  • LocationFilipino @ Singapore

Posted 25 January 2014 - 06:44 AM

the quality of matches is basically this...

Posted Image

#97 Nicholas Carlyle

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Bad Company
  • 5,958 posts
  • LocationMiddletown, DE

Posted 25 January 2014 - 07:58 AM

How in god's name did they glean ANYTHING from this poll?

How the hell does this poll suddenly give them the information they needed, whereas EVERY poll that goes against what they want to do is skewed, or wrong, or voted on by people who aren't the target audience?

Come the hell on people, get your heads out of your rear ends.

#98 RolloI

    Member

  • PipPip
  • The 1 Percent
  • 41 posts
  • Location[ITA]

Posted 25 January 2014 - 07:59 AM

less time, but the game always unbalanced because the weight or ELO players -__-
change ineffective for the purpose

#99 VikingN1nja

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 607 posts
  • LocationIreland

Posted 25 January 2014 - 08:16 AM

View PostPaul Inouye, on 24 January 2014 - 03:38 PM, said:

Thanks for the poll responses folks. It's showing us that the change is doing what was expected. While wait times have gone up for the extremely high and extremely low Elo players, the number of failed matches has not increased significantly. This is meaning that quality matches are kicking off before the max Elo range is being reached.

And off topic: Again, Elo has nothing to do with tonnage balancing. We are still seeing teams that are over tonnage over others. This is where the tonnage matching of the CW Launch Module comes into play.



Off topic but....

http://mwomercs.com/...4-gauss-module/

#100 NRP

    Member

  • PipPipPipPipPipPipPipPipPip
  • Fire
  • Fire
  • 3,949 posts
  • LocationCalifornia

Posted 25 January 2014 - 10:53 AM

The purpose of this change (and the poll) was to assess whether peoples' average time to find a match changed. It had nothing to do with "unbalanced teams".

Anyway, the "unbalanced teams" issue is intriguing. Apparently, PGI is happy with the way the matchmaker is grouping people via ELO (Paul has used the term "quality matches" a few times now), yet I think a lot of players would say that the majority of matches certainly don't feel like "quality matches". I wonder why this is?

Some people feel like a tonnage imbalance is the reason, but I'm not so sure. We've had people who run 12 mans where the teams agree to strict tonnage limits say they still have blowout matches.

Some people feel like the current ELO implementation in the matchmaker may not actually be making "equal" teams, skill-wise, despite the two teams' average ELO falling within a strict tolerance. Personally, I think this might be closer to reason some (most?) matches are stomps. Then again, I've watched videos of the top pro teams going at it and a lot of those matches are 12-2 stomps as well. Now these dudes all have extremely high skills, are running optimized mechs, and are on voice comms running detailed strategies. And they still have what looks like roflstomp games. The pro games seem to have everything we want the matchmaker (and ELO) to do, so why would this be?

As I said, I find this issue intriguing. Might it just be that 12-0/12-2 roflstomps are the nature of this game, even when teams are perfectly balanced in terms of player skill and tonnage? Maybe this is the answer? Once a team loses a mech or two for whatever reason, either through stupid "Leroy Jenkins" charging or if they don't all move as a cohesive unit and encounter superior enemy numbers, maybe the whole damn thing snowballs out of control and leads to a stomp?





6 user(s) are reading this topic

0 members, 6 guests, 0 anonymous users