73 replies to this topic

[gjnii]

How to claw out of elo purgatory:
1) prioritize ANYONE that looks like they're the top threat. Extra kills on easy targets may pad your K-D but Picking off the top 2 mechs on the enemy team can make a much bigger difference than killing the bottom 4. let the noobs flail around with each other, you can just pick them off later.
2) play to win. its tempting to get caught up in winning your fights but sometimes you can sacrifice match rating for a win. sometimes its as simple as standing your ground and dying, other times its more complicated, things like deliberately kiting the enemy team away while engaging rather than taking the kill shots you know you can get.
OR
Grab a fresh assault, something you know you can win with, and just pull it up using a different chassis.
I for instance have a knack for losing in my stalkers, but a knack for winning in my atlas's...
If I play my stalkers too long I start seeing "greens" aka default paint mechs in large numbers.
Thats usually a give away thats its time for me to roll out in an Atlas and claw my elo back up.
I'm not sure what mech would be ideal for winning, but I'd imagine a mobile assault with range/skirmish capabilities would probably be ideal. you want something where you can make sure you're there to save the day wherever and whenever they need it.
Or grab an Atlas and just brute force it.
The RS with
4ML AC20 ASRM6/ASRM4 with a 350 engine will let you close and annihilate almost anything
60 point alpha, and capable of sustained fire with either the 4ML for fast moving targets OR the AC/SRM's for anything that still wants to brawl after the alpha. your team can almost do anything it wants at that point, you can usually just hammer your way to a win.
Or a DDC if you'd prefer.
Its a small investment compared to the time spent in purgatory... and who knows maybe you'll find another chassis you like enough to master.

EDIT: Also, play conquest, you're much more likely to see random wins just due to team composition, and through careful positioning you can help your team a lot more. In assault if you lose the fight you lose the match, in conquest its not so sure a thing.

Edited by gjnii, 18 July 2013 - 01:44 PM.

[Shadey99]

Even the 12v12 test server sees often flip flopping games were you get stuck with the people who somehow found the test client, but can't point their guns in the right direction... So I can go from getting a high score in a crazy fight...



to the bottom score as I'm sniped from beyond visual range within 30 seconds of game start.



The 2nd game had both really good players on the other side (like that sniper) and many from my team had problems figuring out which way to point their guns.

Edited by Shadey99, 18 July 2013 - 05:38 PM.

[Urdnot Mau]

Lots of losses in a row. I've been there, but never for so much time. When i'm pugging i get so frustrated because my team mates suck so hard (much like the ones you described) it just makes me want to stop playing. At certain hours i can't play with my corp mates cause they aren't online, so i'm forced to PUG and basically i'm either the last one alive of caught alone because half my team is scattered across the map. I would imagine that after grinding millions of c-bills via trial mechs to buy a mech would be enough to make people realized that roaming alone is a bad thing.

[Karenai]

Hauser, on 18 July 2013 - 12:04 PM, said:

I'm afraid you both are mistaken.

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

"It's closest to a target value" aka you have a target value for the team. The system wants a game of 8 1300 rating players or a team of 10400. But you do not have 8 1300 players. So the system puts 5 1300 players with one 1950 and two 975 players into the team (or the nearest equivalent of the sum). Would it go by average rating, you could end up with 8 1400 players in a 1300 game, or 8 1200 players in a 1300 game. A game of 1200 vs 1400 is what the developers think of bad matchmaking, which it is. There is no perfect way of doing elo for team sports, when the teams are random.

[Kunae]

Karenai, on 19 July 2013 - 08:24 AM, said:

"It's closest to a target value" aka you have a target value for the team. The system wants a game of 8 1300 rating players or a team of 10400. But you do not have 8 1300 players. So the system puts 5 1300 players with one 1950 and two 975 players into the team (or the nearest equivalent of the sum). Would it go by average rating, you could end up with 8 1400 players in a 1300 game, or 8 1200 players in a 1300 game. A game of 1200 vs 1400 is what the developers think of bad matchmaking, which it is. There is no perfect way of doing elo for team sports, when the teams are random.

It does... not... work... that,.. way.

It goes and picks a value(which value we've never been told, but I assume it bases it off the first person to click "launch". It then looks for people, for each side, close to that value, and adds them in as it finds them until it has 8 per side.

There is no aggregate total that it tries to make. Thus, if the initial value was 1300, and it has expanded the range out by 300 points, it's entirely possible to have a team of 1 1600 and 7 1000's vs a team of 4 1600's, 3 1400's and a 1500.

It does not add up the team.

[Fooooo]

Kunae, on 19 July 2013 - 08:54 AM, said:

It does... not... work... that,.. way.

It goes and picks a value(which value we've never been told, but I assume it bases it off the first person to click "launch". It then looks for people, for each side, close to that value, and adds them in as it finds them until it has 8 per side.

There is no aggregate total that it tries to make. Thus, if the initial value was 1300, and it has expanded the range out by 300 points, it's entirely possible to have a team of 1 1600 and 7 1000's vs a team of 4 1600's, 3 1400's and a 1500.

It does not add up the team.

As far as I understand it you are correct.

Picture it like this.

The server say creates 5 matches.

Game 1 = 2500
Game 2 = 2000
Game 3 = 1800
Game 4 = 1500
Game 5 = 1000

It then looks at the pool of players who are in the queue, (players who have pressed launch) and adds players who are close to matching that rating for each match. The longer it takes to fill the match, the more it will relax the criteria for that match.

So for game 1 you might find that.

1st player selected = 2500 elo (time taken 10 sec)
2nd player selected = 2400 elo (time taken 20 sec)

waiting.................
waiting.................
relax critera.

3rd player selected = 2000 elo (time taken 1min)
4th player selected = 2400 elo (time taken 1min30sec)(this player pressed launch after the match already relaxed criteria, otherwise there would have been 3 players selected b4 the match relaxed the first time. )

waiting.............
relax criteria to lowest.

5th player selected = 1500 elo
6th = 1400
7th = 2000
8th = 2500
9th = 1000
etc etc etc......


This is one of the reason I think you can get instant matches 1 time, then 2min waits another. At first I thought the instant matches would mean there must of been plenty around my elo to fill a match.

However, if a match started by the server has been waiting 2mins for a player, and I happen to click launch and fill the criteria, I will get an instant match nomatter what really. It could be a higher ranked match that relaxed, or one at my rank that only needed 1 more and was about to relax, but I hit launch just in time for.

So an instant match (should) generally be either one of those situations.

I have no idea if when a match relaxes if it goes both ways, as in UP and down, or just down. If it does go up as well as down it would mean a game that started looking for 1000elo players, could end up throwing in 2500elo players to that match.

Edited by Fooooo, 19 July 2013 - 09:49 AM.

[Kunae]

Fooooo, on 19 July 2013 - 09:41 AM, said:

As far as I understand it you are correct.

Picture it like this.

The server say creates 5 matches.

Game 1 = 2500
Game 2 = 2000
Game 3 = 1800
Game 4 = 1500
Game 5 = 1000

It then looks at the pool of players who are in the queue, (players who have pressed launch) and adds players who are close to matching that rating for each match. The longer it takes to fill the match, the more it will relax the criteria for that match.

So for game 1 you might find that.

1st player selected = 2500 elo (time taken 10 sec)
2nd player selected = 2400 elo (time taken 20 sec)

waiting.................
waiting.................
relax critera.

3rd player selected = 2000 elo (time taken 1min)
4th player selected = 2400 elo (time taken 1min30sec)(this player pressed launch after the match already relaxed criteria, otherwise there would have been 3 players selected b4 the match relaxed the first time. )

waiting.............
relax criteria to lowest.

5th player selected = 1500 elo
6th = 1400
7th = 2000
8th = 2500
9th = 1000
etc etc etc......


This is one of the reason I think you can get instant matches 1 time, then 2min waits another. At first I thought the instant matches would mean there must of been plenty around my elo to fill a match.

However, if a match started by the server has been waiting 2mins for a player, and I happen to click launch and fill the criteria, I will get an instant match nomatter what really. It could be a higher ranked match that relaxed, or one at my rank that only needed 1 more and was about to relax, but I hit launch just in time for.

So an instant match (should) generally be either one of those situations.

I have no idea if when a match relaxes if it goes both ways, as in UP and down, or just down. If it does go up as well as down it would mean a game that started looking for 1000elo players, could end up throwing in 2500elo players to that match.

Yep, you seem to have it.

It does relax both up and down. I dread the matches where I am waiting 3 minutes... one side or the other seems to get stacked with window-lickers, in those cases.

[Agent 0 Fortune]

This is why I actively avoid taking FOTM builds. Sure they are great for padding your stats and sky rocketing your ELO, but when the scales of balance shift, you are in for a long pain train until you ELO rebalances.

[c0mbatphil]

Go ahead and add me also. I pug and I play with some really fun guys that also have weird hours.

[mack sabbath]

Funny thing.

Today I built a Spider 5V, one of the worst in the game. I gave it my Hunchbacks 250 Std and 2 med lasers.

I have not lost a match (all Conquest) and have yet to die....With a firepower of 10, I can do 100+ damage with 7 assists. No deaths, no losses, wow?


My poor, poor Misery!

Edited by Die Primate Die, 19 July 2013 - 10:55 AM.

[Lugh]

Die Primate Die, on 19 July 2013 - 10:55 AM, said:

Funny thing.

Today I built a Spider 5V, one of the worst in the game. I gave it my Hunchbacks 250 Std and 2 med lasers.

I have not lost a match (all Conquest) and have yet to die....With a firepower of 10, I can do 100+ damage with 7 assists. No deaths, no losses, wow?


My poor, poor Misery!

Broken mech is broken,

Hit boxes on this guy must be about 1 pixel each as 80% of all damage dished on them is NOT registered. Easy to have success on a mech like that.

I stopped playing my Invincible Dragon during the Hahahah can't hit me stage and the Hahah Dragon Bowling stages of closed beta because it was too damn easy to wrack up kills that way.

It did NOT make me a better pilot. Getting into other mechs at that time did.

[mack sabbath]

Lugh, on 19 July 2013 - 11:04 AM, said:

Broken mech is broken,

Hit boxes on this guy must be about 1 pixel each as 80% of all damage dished on them is NOT registered. Easy to have success on a mech like that.

I stopped playing my Invincible Dragon during the Hahahah can't hit me stage and the Hahah Dragon Bowling stages of closed beta because it was too damn easy to wrack up kills that way.

It did NOT make me a better pilot. Getting into other mechs at that time did.

Yeah, well I remember killing Spiders occasionally, so they aren't Gods, especially this variant. And I have NO remorse for staying alive, doing representative damage and most of all, ending a losing streak that defies all sanity.

[Jman5]

Xie Belvoule, on 18 July 2013 - 10:31 AM, said:

That's a great video. I never played LoL, so when I started hearing about "Elo Hell" I got really confused. The entire concept makes zero sense. If you're better than those in your Elo bracket, then your effort will tip your team to win more than lose in the long run. Winning more than losing makes your Elo go up in the long run. However, if your Elo is not going up then either you're about as skilled as the average person in your bracket, or you're worse.

[ArmandTulsen]

Elo definitely is ridiculous.

Every PUG I play it's literally a flip every match between playing against teams and getting steamrolled or playing against complete and utter noobs who get steamrolled by us.

Very rarely there's the occasional streak where you you win several in a row or lose several in a row.