31 replies to this topic

[Diablobo]

The Elo rating system was designed to rate chess matches. It has since been used in other various games, but they all have the same common characteristic:

THEY ALL HAVE THE SAME STARTING EQUIPMENT.

This is obviously not the case with MWO. One team might have more assaults than the other. One team might have more lights in a conquest game.

The Elo system would not work in chess matches where one player gets an extra rook or bishop, so how is it going to work in MWO when one team has more/better starting equipment? I'll give you a hint: it's not.

The MWO matchmaker needs to take tonnage and/or BV into account.

Edited by Diablobo, 19 July 2013 - 12:55 AM.

[The Cheese]

Diablobo, on 19 July 2013 - 12:54 AM, said:

The MWO matchmaker needs to take tonnage and/or BV into account.

And after it's done that, it can just throw people into a match together and have the vets get narky about dropping with newbies and the newbies getting narky because they get rolled in seconds.

And then we need a matchmaker based around something like ELO.

What we got aint perfect, but it's potentially better than not having any skill based matchmaker at all. I say "potentially", because it's clear as day that our matchmaker needs work.

[Tennex]

or any team game for that matter.

but its the best system anyone has come up with to date... so..

[Borengar629]

In fact the only real solution would be to take tonnage and ELO into account. But in order to do that I think you need a hell lot of active players online at all time or the waiting will take like forever.
But most of all I have the feeling that ELO is fluctuating very harsh. If you play for a couple of hours I can't shake the feeling that I crawl up the ladder very fast in the beginning, then being just annihilated for a few matches just to end up with a lot of douches soon afterwards.
I think the rate with which ELO can change should be slowed down a little to begin with.

[Hauser]

Diablobo, on 19 July 2013 - 12:54 AM, said:

The Elo system would not work in chess matches where one player gets an extra rook or bishop, so how is it going to work in MWO when one team has more/better starting equipment? I'll give you a hint: it's not.

Elo is only influenced by winning or losing unexpectedly. So it measures your overall performance. Since that performance is influenced by the gear you bring it will work fine. Players with extra rooks or bishops would get higher Elo.

For MWO you also have a different Elo value for lights, mediums, heavies and assaults.

Diablobo, on 19 July 2013 - 12:54 AM, said:

The MWO matchmaker needs to take tonnage and/or BV into account.

http://mwomercs.com/...-making-update/

Quote

Is the match maker taking weight class into account?

Yes the version currently live on production does take weight class into account. The way the match maker works it tries to group players within certain thresholds and grows those thresholds over time to ensure a match is made within currently a 2 minute time period.

These curves will likely see some tuning for both skill and weight as we monitor the data coming in and work to ensure more and more matches are created within good thresholds. This tuning will happen over the course of upcoming patches.

Edited by Hauser, 19 July 2013 - 01:44 AM.

[Kahoumono]

Just take matches played for that mech and weight into account and separate the pugs and the groups. At the moment someone with a similar elo in a group will almost certainly do better than the pug because of coordination. Simply taking matches played in a mech takes away and complicated calculations. If you are still garbage in a mech you have played 100+ matches in then you should find another hobby.

[Appogee]

But MechWarrior is just like a game of chess...

Incredibly strategic, requiring analysis of multiple scenarios, planning ahead many moves in advance...

Umm, nah. You're right. Elo is making the game worse, not better.

[FunkyFritter]

Diablobo, on 19 July 2013 - 12:54 AM, said:

The Elo rating system was designed to rate chess matches. It has since been used in other various games, but they all have the same common characteristic:

THEY ALL HAVE THE SAME STARTING EQUIPMENT.

League of Legends.

[TexAce]

ELO is fine, the only problem is that the matchmaker takes ANY elo it can find to get a group ELO and match this one against another group.
What it should do is just put people together with nearly identical elos.
Level 5's only fight Level 4-6 and so on.
At the moment all kind of ELOs represented in a group which is the culprit.

Problem is: we need more players to get this out the way.

Edited by TexAss, 19 July 2013 - 02:29 AM.

[eRoiK]

Problem is that i'm actually around 1 W/L ratio, all my games are full of "Trial Atlas" new player that don't even know how to target ...
Atlas RS trial is the only playable build on trial but please... when all assault slot are taken by new players VS 2000+ games player is just dumb.
That make the game worse.

[Matroid]

Diablobo, on 19 July 2013 - 12:54 AM, said:

The MWO matchmaker needs to take tonnage and/or BV into account.

I agree

Hauser, on 19 July 2013 - 01:44 AM, said:

http://mwomercs.com/...-making-update/

(1) the post was so old (Feb 2013) that it isn't useful anymore. it's probably from when ELO was launched
(2) they say it takes "weight class" into account, nothing about tonnage. they don't give a corresponding value... it could be 0.00001% and nearly useless
(3) i've seen matches with 7 assaults vs a team with 2 assaults = they prioritize other things >> tonnage. i'm guessing "finding a match within a window of 2 minutes" is a bigger variable in the algorithm


The best system they had so far was the Class Matchmaking system, as this only kept people from playing the lower-end of the weight classes (Awesomes, Dragons, Commando). people don't play those mechs much anyways (due to hit-boxes)


A better system than ELO will take into account:
- Pilot skill (calculated from K/D, W/L, total games, etc)
- Mech (tonnage, body shape, hard-points)
- Load-out (weapons, engine type, upgrades, equipment)

Then [Value = Pilot Skill * Mech * Load-out]


This will prevent new players with a Dragon-1C full of flamers (lowest value) from fighting skilled players with an Atlas D-DC equiped with ECM / Gauss / SRM6's (highest value)

Edited by Hadros, 19 July 2013 - 04:32 AM.

[Snowhawk]

Diablobo, on 19 July 2013 - 12:54 AM, said:

The MWO matchmaker needs to take tonnage and/or BV into account.

I agree.... but now try to explain this to the devs...

[Kunae]

The Cheese, on 19 July 2013 - 01:15 AM, said:

And after it's done that, it can just throw people into a match together and have the vets get narky about dropping with newbies and the newbies getting narky because they get rolled in seconds.

It's doing that already, but typically with a huge weight disparity.

The Cheese, on 19 July 2013 - 01:15 AM, said:

And then we need a matchmaker based around something like ELO.

What we got aint perfect, but it's potentially better than not having any skill based matchmaker at all. I say "potentially", because it's clear as day that our matchmaker needs work.

Weight-class matching, BV matching, or weight matching should be the primary determinant. After that, you can figure in some "skill-based MM". Please note, this should only be for pug matches. Other methods should be available for set matches.

[jakucha]

That's a nice video talking about ELO. TL;DR, ELO has since been modified for team games.

We're also getting tonnage limits, which will help a bit.

Edited by jakucha, 19 July 2013 - 06:56 AM.

[Syllogy]

Diablobo, on 19 July 2013 - 12:54 AM, said:

The Elo rating system was designed to rate chess matches. It has since been used in other various games, but they all have the same common characteristic:

THEY ALL HAVE THE SAME STARTING EQUIPMENT.

This is obviously not the case with MWO. One team might have more assaults than the other. One team might have more lights in a conquest game.

The Elo system would not work in chess matches where one player gets an extra rook or bishop, so how is it going to work in MWO when one team has more/better starting equipment? I'll give you a hint: it's not.

The MWO matchmaker needs to take tonnage and/or BV into account.

Counterstrike uses elo.
NCAA uses elo.
Major League Baseball uses elo.
Guild Wars uses elo.
World of Warcraft uses elo.
Magic the Gathering tournaments use elo.
FIFA uses elo.

... Not just for chess, haus.

[Bagheera]

TexAss, on 19 July 2013 - 02:29 AM, said:

Problem is: we need more players to get this out the way.

This.

More players across the Elo spectrum will balance out who is teamed with whom. Problem is right now that there are not enough players available for a match at the same time with similar enough Elo to make the system "feel" like it is working.

Larger set from which to select and these problems start to balance out.

[Amsro]

Appogee, on 19 July 2013 - 02:01 AM, said:

But MechWarrior is just like a game of chess...

Incredibly strategic, requiring analysis of multiple scenarios, planning ahead many moves in advance...

Umm, nah. You're right. Elo is making the game worse, not better.

Elo has no effect on making the game better or worse, matchmaker is still broken just like it was 6 months ago.

The only place Elo should be tracking is competitive 8 man (future 12 man). At least this way you can have some degree of control over the ENTIRE team. Like chess you control all of a team.

Pug drops,4 man drops you can only get %50 of your chess pieces in a match the rest is controlled by .. random chess players (maybe never even played chess/mechwarrior)

If it isn't working because of our small community... then why are we trying to use it in a small community?

[Gallowglas]

Snowhawk, on 19 July 2013 - 04:37 AM, said:

I agree.... but now try to explain this to the devs...

Quote

MeatForBrains: Are there future plans to match by tonnage as opposed to weight class? What are the pros and cons of this approach?

A: We endeavor to allow players to play the game however they chose. Putting artificial limits always tends to create other problems. Sometime it’s necessary to promote a more diverse ecosystem. In the case of tonnage vs weight class, tonnage is better for players, as they can field whatever they chose, and weight classes is better to guarantee match quality/equality. We’re working on striking a balance between the two.

Stjobe: What are your plans for getting more lights and mediums onto the field?
The current game is absolutely dominated by heavies and assaults, to the point where I'm often the lightest thing on the field - both teams included - when I drop in a Blackjack. I'd love to hear your take on this, and your plans on rectifying the situation (if you have any).
If I may make a suggestion as well as pose a question: I suggest you take a look at your reward structure; it's almost completely focussed on what heavies and assaults are good at, and almost completely neglects to reward what lights and mediums are good at - and the rewards we do have for things like spotting and capping are a pittance compared to what you get for dealing damage and killing.

A: You pretty much nailed it on the head. Rewards, along with expanded roles for lights and mediums, along with some limitations on drop weights, will all contribute to a more diversified battlefield.

..but now try to explain this to the forum-goers who don't read dev posts.

Edited by Gallowglas, 19 July 2013 - 07:53 AM.

[Appogee]

Amsro, on 19 July 2013 - 07:42 AM, said:

If it isn't working because of our small community... then why are we trying to use it in a small community?

That, indeed, is the question!

[Diablobo]

FunkyFritter, on 19 July 2013 - 02:07 AM, said:

League of Legends.

How many players does LoL currently have playing and how are they different? Do they have level 10 players (stock medium mechs) going up against level 30 players (custom assaults)? I think not.

Edited by Diablobo, 19 July 2013 - 09:11 AM.