277 replies to this topic

[BFett]

PhoenixFire55, on 15 May 2013 - 12:28 AM, said:

These BV's are for stock configs, thus irrelevant for non-trial mechs.

OK, these values were only for the example and could be updated as weapons and equipment changed on the mech.

I just wanted people to see what I was thinking. If you have a better or different idea feel free to post it.

[PhoenixFire55]

BFett, on 15 May 2013 - 05:47 AM, said:

I just wanted people to see what I was thinking. If you have a better or different idea feel free to post it.

Don't have anything against your idea, just saying. In general that is a good idea to match not only peoples 'skill' but 'quality' of their mechs as well, but people who played TT game or MegaMek etc will agree with me that BV system never really worked well. There are just too many factors that should be included in BV that aren't there. Apart from that there is always a question of how exactly do you calculate (player value & mech value) score - Elo+BV or Elo*BV etc.

I've posted several suggestions about matchmaking ever since phase.1 in different topics, sadly they don't care.

[PhoenixFire55]

Thontor, on 15 May 2013 - 07:56 AM, said:

My problem with any BV type system, placing value on certain weapons, is that some people are wetter with some weapons, and other people are bad with those same weapons but better at different weapons.

A system assigning value to stuff like that would have to give a unique value for each weapon system to each person, based on their skill with that weapon system. Accuracy, damage per hit for lasers, etc.

I think that's the only way a system of assigning value to mechs would work. But I also think that it would not be worth the effort.

Well we all do have those stats already ... player X hits with weapon Y with Z% rate. You can check it in your profile (You can go further and make it player X hits with weapon Y with Z% rate in chassies A on map B and so on...). It just takes willingness to do it from PGI but they keep pedalling their broken Elo matchmaker and keep denying that it is broken. A screenie I posted just above is just an example of it being broken, I've had many games like that. This one is just unique coz none of the people on one team breached 100 damage and match ended within 3 mins.

Edited by PhoenixFire55, 15 May 2013 - 08:07 AM.

[PhoenixFire55]

Thontor, on 15 May 2013 - 08:31 AM, said:

No matchmaking system is perfect. What we have now works fine, coming up with a whole convoluted new system is just not worth the effort for what would likely be very little difference.

PhoenixFire55, on 14 May 2013 - 03:23 AM, said:

Elo matchmaker ... really close game for you ...

Is THIS fine?? Less then 3 mins. Nobody does more then 100 damage.

Edited by PhoenixFire55, 15 May 2013 - 08:41 AM.

[PhoenixFire55]

Thontor, on 15 May 2013 - 08:47 AM, said:

The current system is fine, but still needs tuning. But I already said that in the part of the post you didn't quote.

Few outliers is when you get stampede like that once in 50 games not every second game. The reason for those stampedes is Elo itself and no fine-tuning is gonna fix it. Besides PGI says that their matchmaker is working as intended so I wouldn't expect any changes.

[PhoenixFire55]

Thontor, on 15 May 2013 - 10:20 AM, said:

"Stampedes" are far more rare than that.

You want me to send you screenshots with times and dates so that you see it happens every 2nd game for me?

Thontor, on 15 May 2013 - 10:20 AM, said:

The reason "stampedes" happen is because in MWO, number of forces plays a huge part. Once the first domino falls, it's very hard to prevent the rest from tumbling down behind it.

Nothing to do with domino effect. When one team does 500-600 damage combined and another team does 3000 damage its always gonna end up one way no matter if 500 dmg is focused or not and no matter what forces play their part. And this happens even when team tonnages and mech types are in perfect harmony. Needless to say that a lot of stampedes happen due to ret@rded team compositions in terms of mech types and weapons those mechs have.

Right now a guy can 'tank' his Elo playing with 3 great players thus boosting his win/loss ratio. Even if hes not so good himself he'll have high Elo. Games will always be unbalanced while Elo is based off wins and losses.

[zhajin]

There are far to many other factors in game for us to really know if elo is having any effect. I would even go so far as to say that might be the case for PGI as well, even with all their metrics gathering. looking at pretty graphs is great and all but statistics often lie. ELO is only one small part of the puzzle.

I will go out on a limb and say that groups, mech builds, and player experience in that particular build, all factor in at least as much as ELO. Now multiply that by 8 (and soon 12). Then you have to factor in all the things that players do not have control over. map factors (starting base, mech build on particular maps), disconnects/bugs, weight balance, speed balance (4 lights on 1 team vs 1 on the other type stuff). and dozens of other factors i can not even think of at the moment.

When you add it all up ELO is small part of the match making. And true balance will never really be achievable. Which I think is ok because it make the game a bit more real. frustrating at time yes, but so is life...

Edited by zhajin, 15 May 2013 - 11:55 AM.

[Hauser]

I've got 4 concerns with Elo. E.g. things that influence the quality of match making.

New players
New players are put in with an average Elo. They will descend down but on their way down they'll run into allot of people with average Elo that are annoyed by the fact they'll lose what would otherwise have been a close game due to a new player on their team.

Old players, new mechs
A new mech is nowhere near as good as one with all the upgrades. By starting a new mech you'll handicap yourself and your team.

Premades
People on Teamspeak have an advantage over normal players. They don't have this advantage when they play solo. In the long run Elo will find an appropriate rank for players who do either. If they do both however their Elo value is always wrong. As such when players join a premade their Elo should go up a bit. PGI should have enough data to see how much more likely a premade is to win and adjust the Elo of a premade accordingly.

[PhoenixFire55]

Thontor, on 15 May 2013 - 11:35 AM, said:

As I said, "stampedes" happen both ways even when it's the same exact people in the same exact mechs.. Stampedes are the nature of the game.

And yes, games will always be unbalanced.... Whatever system they use. There is no perfect matchmaking system... There is no such thing as always having balanced games.

Well then the whole idea of Elo matchmaker is complete BS. Because only reason why PGI began doing it is because of poor whining kids who cryed about getting stomped by overpowered 'premades' and teamwork. They wanted evenly matched games that would be 'interesting' for them and would end up like 8:7, 8:6 etc. In reality 'premades' were to be blamed for like 5% of stampedes while remaining 95% were stampedes of PUGs by PUGs. This is exactly same thats happening right now, PUGs are still stampeding other PUGs, and stampeded PUGs still cry about 'premades' sync-dropping or whatever. I have been accused of being a sync-dropped premade twice just yesterday while in reality I was playing on my own all day.

If matchmaker is not making any even matches and if its not meant to do so as you claim then this matchmaker is no better then random matchmaker (that matched weight classes only) that we had in closed beta. From my experience (and I obviousely can only judge by my experience and experience of people I'm sometimes playing with) Elo phase.4 matchmaker is producing way more unbalanced and uneven matches then we had prior to phase.1.

You are saying that Elo is better then no Elo, but please explain just how is it better then phase.0.

[PhoenixFire55]

Hauser, on 16 May 2013 - 01:06 AM, said:

New players
New players are put in with an average Elo. They will descend down but on their way down they'll run into allot of people with average Elo that are annoyed by the fact they'll lose what would otherwise have been a close game due to a new player on their team.

This should indeed be no.1 concern for PGI right now. Its a shame they don't realize it.

Hauser, on 16 May 2013 - 01:06 AM, said:

Old players, new mechs
A new mech is nowhere near as good as one with all the upgrades. By starting a new mech you'll handicap yourself and your team.

Agreed. But its easy to fix, just make Elo score for each chassis variant instead of just Elo for lights, mediums etc. IMO it'll also fix or at least improve mech vs mech disbalanced like for example between Raven 3L and Raven 4X. A player in 4X just won't be able to achieve same Elo score as he'll have in 3L, thus a 'weaker' mech will always have lower Elo score.

Hauser, on 16 May 2013 - 01:06 AM, said:

Premades
People on Teamspeak have an advantage over normal players. They don't have this advantage when they play solo. In the long run Elo will find an appropriate rank for players who do either. If they do both however their Elo value is always wrong. As such when players join a premade their Elo should go up a bit. PGI should have enough data to see how much more likely a premade is to win and adjust the Elo of a premade accordingly.

Not everyone who is in group is also using TS or other voice comms. I've played in groups of 4 that just randomly joined up together without any comms but in-game chat boxes. These people aren't any different from typical 1-man PUGs. And there is no way to track if people use comms or not.

Edited by PhoenixFire55, 16 May 2013 - 02:06 AM.

[Hauser]

PhoenixFire55, on 16 May 2013 - 02:05 AM, said:

Agreed. But its easy to fix, just make Elo score for each chassis variant instead of just Elo for lights, mediums etc. IMO it'll also fix or at least improve mech vs mech disbalanced like for example between Raven 3L and Raven 4X. A player in 4X just won't be able to achieve same Elo score as he'll have in 3L, thus a 'weaker' mech will always have lower Elo score.

That may cause problems where the majority of people don't have enough games on a variant to give them the proper elo value. Bit hard to discuss that without statistics.

PhoenixFire55, on 16 May 2013 - 02:05 AM, said:

Not everyone who is in group is also using TS or other voice comms. I've played in groups of 4 that just randomly joined up together without any comms but in-game chat boxes. These people aren't any different from typical 1-man PUGs. And there is no way to track if people use comms or not.

Wait, that actually happens?

But for the match maker the difference doesn't really matter. We're looking at averages anyway. PGI should have access to data to draw up some statistics that show premades win X% of the time while the match maker predicted (X-Y)%.

Edited by Hauser, 16 May 2013 - 04:57 AM.

[PhoenixFire55]

Hauser, on 16 May 2013 - 04:57 AM, said:

Wait, that actually happens?

It actually does. Me and a buddy were on TS the other day and we picked 2 more guys who didn't use TS to fill our group. I've been asked myself to fill in a group of 4 several times by people I've never met before. They didn't use TS or C3 either.

[Hauser]

You know? It'd be nice if we'd had some statistics to discuss this all. Aside from the occasional anecdote and screenshot we've got absolutely nothing solid to discuss right now.

Edited by Hauser, 16 May 2013 - 05:57 AM.

[PhoenixFire55]

Thontor, on 16 May 2013 - 06:25 AM, said:

This is about the best we've gotten. Garth hearing from an engineer who has seen the statistics first hand.

I tend to think that the 30% decline is due to newbies who started in main Elo package dropping down hard in Elo score thus affecting 'middle zone' matches less.

Still, for me its 33% 8:0-8:2, 33% 0:8-2:8 and 33% in-between those. Some people are saying that they lose 90% of their games, and I really have no reason not to believe them.

[Major Derps]

I really don't feel ELO works at all for this game, that's my opinion. Perhaps if newer players were dropped into an elo of 0, instead of the median, it wouldn't be so bad; but right now I am getting frustrated with having to work my *** off to get a win (let alone 100+ damage), because my team mates are still playing like it's COD.
To me, it feels like you are locked into the median, by way of chance. A couple of good players can be dropped with a group of new players, or bad players piggy backing off team mates; or be on the team of a disconnect, or disconnect themselves. Now as much as we all like to think we can, 1 player can't dictate whether their team wins or not, only influence to a degree. This is a game where the greater number wins 90% of the time, and skill always comes second to that. IMO, this needs a whole new idea for ranking; something like the way you run your tournaments would be perfect.
Simply awarding a player more for killing a higher ranked player doesn't cut it. 1v1 fights are too few and far between, and an outnumbered (even by 1) elite player can be defeated by a noob (I know, because I've done it myself).

On the changes to weight balancing: Honestly, I haven't noticed too much of a difference; though what difference I have noticed, is definitely a positive one.

In conclusion: Whilst I will concur that the game is more enjoyable than it was without it, I don't think that ELO is the system we need. A ranking system based on the way your tournaments are scored, would be ideal.

Edited by Mokey Mot, 17 May 2013 - 10:38 PM.

[Hauser]

You can always hop onto Teamspeak. The community is generally mature and pickup games don't take too long to setup either. Won't fix match making but it will help with your problems.

http://mwomercs.com/...mspeak-servers/

Edited by Hauser, 19 May 2013 - 02:59 AM.

[BFett]

The general problem I have with ELO is that currently it's not based off of how well you as a player do, it's based off of whether or not your team wins or loses.

If this remains true, a player could take out 7 enemy players in assault mode and still have their ELO lowered due to losing by base capture. What this tells me is that the good players actually get stuck with the poor players because ELO only cares about if that player wins or loses a match.

[Hauser]

Match Making Update

- All new players starting the game will have 2 Elo scores assigned to them.
- New Player Elo = 1100 (N-Elo)
- Standard Player Elo = 1300 (S-Elo)
- Any new player (has 25 games or less in their history) will be using their N-Elo score to be matched with other players.
- The outcome of the game will affect both their N-Elo score and their S-Elo score.
- When the player has completed their 25th game, the N-Elo score is eliminated and the player from that point on will use their S-Elo score for match making.
- This change makes the initial experiences of a new player much easier and prevents the player from playing players with numerous matches under their belts.


-----

I like the provisional rating for players with less then 25 matches. I haven't seen any trial mechs in my games these days.

[GrizzlyViking]

Matchmaking still does not work reasonably for about 50%+ of matches. Too many lopsided outcomes too frequently in PUG matches. I have experience in both. When one team has more grouped players than the other team they almost always defeat the team that has more solo players. My recommendation is to separate PUG solo play completely from all groups. I think this will solve the balance issues that we are seeing now.

Edited by GrizzlyViking, 26 May 2013 - 10:00 AM.

[HarmAssassin]

Matchmaking 4.... Don't know what game they're talking about with 80+ % of matches falling within tolerable weight matchups... but it sure isn't MWO.

I'd say 80 % of matches I play in, are horribly lopsided weight-wise. Can't count how many times in one night we'll see one team's heaviest mech being a 65 tonner, while the other team rolls in with a half dozen assault mechs (and that's assuming my team even has 8 mechs to start with, which is pretty rare on most days).