37 replies to this topic

[Misadventure]

I am an average player. I do have the occasional good match but overall I am just average.

It seems like my opponents are getting unreal powerful, were talking the type of people who could fire a small pulse laser 4K and behead my entire team in one shot kind of good. (ok not really but it feels like that most games.)

Premades, well I know there not cheating but you'll never convince me of that in PUGs.

So naturally being the never before seen Enterprise crew member chosen for the away mission, you can imagine what my games are like right now. Every match is transporter malfunction this, and lava monster that, and opps the klingon shot me while going for kirk. You know the deal.

My issue is I am on the most unimaginable losing streak ever, even in matches when my team is good it’s rarely a win unless desperation capping is involved. Every team member is important but 1 member can almost never sway a battle, well 1 member not in a spider anyway. I still manage 400ish points and most of the time a kill or 2 a game so I am not going down without weakening my enemy.

So why, after so many losses am I still getting paired with the demigod ELO players, thankfully not the gawd ELO players as I would have to quit! When will my ELO start dropping down so I can play against people I should be pair with?

For that matter WTF is ELO based off team wins/losses instead of mostly individual skill related stuff. Should ELO not be based off damage, scouting, capping, killing, hits vs misses, basically all the individual things that are tracked for the end of match and online stats with just a small amount of input from the team win/loss data.

Sorry needed to vent, it’s been a horrible wipe weekend and seeing the entire team get wiped 3 matches in a row tonight, (only 1 match was the team stupid btw), is starting to wear on me, I am so tired of how MWO is matchmaking I am depressed enough not to want to play any games tonight.

[Roland]

I honestly will never understand folks who complain about the competition being better than them.

[Roadbeer]

Love the analogys OP

Consider it a bad week. I've had weeks where I could count my losses on one hand, and others where I felt like the absolute worst player on the planet.

Rest of the time, I'm about even sooo....

[Koniving]

Sounds like someone could use another person to drop with. I'm in game now if you're interested?

[Nebuchadnezzar2]

I believe elo is doing it's job to get you somewhere close to 50:50 w/l in the long run. You'll have losing streak and some other time you'll have winning streak.
If you are on the negative side of 1 w/l ratio you should evaluate your playstyle, your impact to the game. I'm not talking about your damage permatch since you can have good damage per match but your playstyle is actually burdening the team.

Try to give positive contribution to the team, ease the burden of your teammate, be a team player, i believe your w/l will somewhat be positive and of course try to get competitive group will help since it will upset the 50:50 balance

[PEEFsmash]

Things get hard when Elo gets high. That's just the way it is. Embrace it.

[xengk]

with stats like this, either the ELO/MM is working or Im a some kinda fluke.
the WL is about 0.93 though.

[Selfish]

xengk, on 03 November 2013 - 09:10 PM, said:

with stats like this, either the ELO/MM is working or Im a some kinda fluke.
the WL is about 0.93 though.

Elo doesn't match based on anything other than W/L predictions and results. While improbable, you could have one of the worst KDRs of the bunch and still be an upper Elo player.

[Screech]

I am pretty sure if your Elo remains rising after 4 hours you should go to a doctor.

[Jonathan Paine]

From what the developers have shared regarding Elo, only your win/loss matters. If your win loss is higher than 1.00 and stays there, you should eventually drop against demigods or whatnot. If your Elo is <1.00, you should eventually see less demigods. Caveat being a badly programmed matchmaker and server population.

[Lexx]

I don't think that ELO looks at your win/loss ratio at all. I have a feeling it's mostly based on your match score.

Even if you are losing all your games, but you still get kills and score decent damage, then you're going to have good match scores. Those high scores will keep your ELO up there even if you are losing matches.

Also the ELO doesn't just look at you, but it's an average of your whole team. You might be the good player that it puts in with some mediocre and bad players to even things out. Also, ELO doesn't look at ECM and premade teams. I have seen some really lopsided matches. Like matches where one has no ECM and a bunch of LRM boats, against a team with 3 Atlas DDCs and an ECM Spider. Those aren't fair matches at all.

When we get community warfare with lobbies and tonnage limits I think it will get a lot better. We'll be able to see who we are dropping with and against ahead of time. We will also be able to set up voice coms with our whole team. We will even know the map we're fighting on, so we can go in much more prepared.

[Arcturious]

Some of the issue is the match maker won't move your Elo much if you lose a game it expected you to lose.

When it creates the matches, it will try and match you with similar bracket players. However, it also guesses which team will win.

Your Elo will only change if it's guess is wrong. So you could just be having a run of matches it thinks you were going to lose anyway, in which case your Elo score won't move much at all.

[Abivard]

ELO was designed so PGI could always troll it's player base.

This is the only logical answer.

[meteorol]

Selfish, on 03 November 2013 - 09:17 PM, said:

Elo doesn't match based on anything other than W/L predictions and results. While improbable, you could have one of the worst KDRs of the bunch and still be an upper Elo player.

Does this basically mean my elo is a result of pure randomness? Because that is how it feels.
My w/l is positive (around 1.15-1.2) and has always been. The k/d of the mechs i used in the last 300 matches (all assault mechs) is 3.26, 3.45 and 4.71.
My average k/d is close to 3 and raising every day.
Still, it doesn't seem like my enemies or teammates get better. Ever. I feel like my elo score didn't get higher since ages.

Probably i'm often matched against teams with worse average team elo, and therefore my elo doesn't get higher if i win.
Since you can only do so much in a 12vs12 when playing solo (i'm not playing in premades very often) it almost feels like my own performance has zero impact on my elo score.
This is really frustrating. I'm tired of seeing guys who can barely hit a single shot or run into buildings half of the match.

[Hauser]

Jonathan Paine, on 03 November 2013 - 09:18 PM, said:

From what the developers have shared regarding Elo, only your win/loss matters. If your win loss is higher than 1.00 and stays there, you should eventually drop against demigods or whatnot. If your Elo is <1.00, you should eventually see less demigods. Caveat being a badly programmed matchmaker and server population.

Just a nit pick here.

Elo doesn't look at your win loss ratio. At all.

Rather it looks at your teams Elo score and that of the other team. If yours is lower you are expected to lose, if it is higher you are expected to win. If you don't do as expected your Elo score is adjusted. How much it is adjusted depends on how big the upset was. The more unexpected the outcome, the bigger the change.

If you have a positive w/l it doesn't mean you're a good player. It means that you've often been matched against players that were worse then you. The reason for this can be that overall you're a good player, or it can be that there were no other players of your Elo rating around at the times you play.

The quality of players vary through out the week. Just an example. When playing Saturday morning in Europe you'll mostly have pugs and allot of new players. LRMs actually tend to work at these hours. Sunday in the European evening, American afternoon, all the organized units get online. You can see that team work improves, matches become more dynamic, players generally are better (game gets more fun too).

So in simple terms, your w/l is time dependent.

Edited by Hauser, 04 November 2013 - 12:55 AM.

[Hauser]

Lexx, on 03 November 2013 - 10:58 PM, said:

Also the ELO doesn't just look at you, but it's an average of your whole team. You might be the good player that it puts in with some mediocre and bad players to even things out.

Another nit pick. After assembling both teams the match maker does indeed average the Elo of both teams. However when assembling the match maker uses a target value. It will start with players near that value and as time goes by the match maker will get less picky.

[Nebuchadnezzar2]

Hauser, on 04 November 2013 - 12:50 AM, said:

Just a nit pick here.

Elo doesn't look at your win loss ratio. At all.

Rather it looks at your teams Elo score and that of the other team. If yours is lower you are expected to lose, if it is higher you are expected to win. If you don't do as expected your Elo score is adjusted. How much it is adjusted depends on how big the upset was. The more unexpected the outcome, the bigger the change.

This is reasonable but somehow does not explain why the majority of people has w/l ratio close to 1

EDIT: corrected myself, after experimenting with thousands of random number, if the chances of being on the side expected to win or side expected to lose is totally random (50:50) the resulting w/l ratio is also close to 1

So yeah what you say is very believeable

Edited by Nebuchadnezzar2, 04 November 2013 - 01:38 AM.

[anubis969]

Nebuchadnezzar2, on 04 November 2013 - 01:13 AM, said:

This is reasonable but somehow does not explain why the majority of people has w/l ratio close to 1

The point of the matchmaker, and of Elo, is to give both teams as close to 50% chance of winning as possible. As a result, once someone has enough matches under their belt for their Elo to stabilise their W/L will in theory settle at around 1.

However, there are things that can throw off your W/L such as playing more than one 'mech of a weight class, playing at off peak hours, playing both solo and premade with the same account, etc..

[christophermx4]

anubis969, on 04 November 2013 - 01:30 AM, said:

However, there are things that can throw off your W/L such as playing more than one 'mech of a weight class, playing at off peak hours, playing both solo and premade with the same account, etc..

Essentially, this.
I usually drop in premades and win. My friends in my unit are decent and we can rack up a string of wins. That said, when I drop by myself it seems I'm facing the finest Mech Warriors this side of the galaxy. Before ELO was introduced, I had a 5.00 KDR - since ELO, I've been trending down to a 2.5. The game does seem more exciting now though - I can't say it's fun being annihilated, but the increased challenge is making me play more often.

[Hauser]

anubis969, on 04 November 2013 - 01:30 AM, said:

The point of the matchmaker, and of Elo, is to give both teams as close to 50% chance of winning as possible. As a result, once someone has enough matches under their belt for their Elo to stabilise their W/L will in theory settle at around 1.

To a point yes. The match maker isn't trying to make two equal teams. It's trying to make two teams in a certain Elo range.

Two equal teams might be:

A: 1350,1300,1275
B: 1325,1325,1275.

But if the match maker is looking for players around 1300 with a delta of 50 it will pack them together in a way that also balances the weight classes.

So you may end up with:

A: 1300 (100), 1275 (50), 1275 (50).
B: 1350 (100), 1325 (65), 1325 (35).

Which is balanced weight wise but Elo wise it is a rather lopsided match.

This is (I would guess) also the reason the next iteration of match making will use a tonnage limit for both teams and let them then decide what they want to bring. It allows the match maker to remove tonnage from the equation and just focus on Elo.

Now I'd link a post here but I can't find it... :S

Edited by Hauser, 04 November 2013 - 05:07 AM.