183 replies to this topic

[Joseph Mallan]

Bogus, on 31 March 2013 - 05:59 PM, said:

So basically what the OP is saying is that ELO actually does work and it's making it harder for coordinated 4-mans to roflstomp pugs? Hm, interesting, most comments I've seen are that ELO isn't pairing people up right.

I started 3 Alt accounts to test Elo. I have played them completely as a PUG. My win loss records on all my accounts are roughly 60%(+1%). If Elo is supposed to make us .500 average, it is not quite working right yet.

[Yokaiko]

Mine are 1.22 on conquest and 1.77 on assault.

A lot of that is conquest is more fluid and the lack of communication hurts in pugs.

[Corwin Vickers]

Joseph Mallan, on 01 April 2013 - 04:20 AM, said:

I started 3 Alt accounts to test Elo. I have played them completely as a PUG. My win loss records on all my accounts are roughly 60%(+1%). If Elo is supposed to make us .500 average, it is not quite working right yet.

Well think about it, if you just started out and they give you a rating of 1000. All the other players around you have 1000 Elo. However, on your normal account you play at a rating of 1500. You're going to keep winning and gaining points till your rating is 1500, at that point you will start your 50/50 spread.

The way it works is if you win that first match at 1000 you take some of the points from the guys on the other team, say 20 points (depends on how much of a handicap you had vs the other team). Now you're at 1020 and they are at 980. You should be matched up with players who are around 1020 next time you play and they will be matched against players slightly lower than before.

Also take into account that it is a Target Point Average or however they described it. The match maker says I'm making a game where each side has 1000 average points per player. That doesn't mean all 8 players on each side have a 1000 rating, it means that some might have 500 but others might have 1500. This is the part that most people complain about.

To fix this the match maker would have to keep people in queue longer than it does now. It would have to wait for 16 players who all have around 1500 Elo. With the way people queue and the population at the time and all the people stuck in games already it could be quite a wait till the top echelon people get matches. In the middle of the bell curve it wouldn't be too bad though since most people will be in the middle. Larger pool means faster queues.

Then you add in 3rd person people and people who turn off certain game modes and you rapidly lose players to fill these matches with a close match.

[Sprouticus]

Caleb Lee, on 31 March 2013 - 06:10 PM, said:

No... I'm saying that I am constantly paired with the following types of players as the other 4 on our teams:

1) Little to no damage, can't hit what they do shoot at.
2) Run in as singletons.
3) No cohesiveness when they are a team, i.e. Pugs
4) Tell you to F*** off if you give any directions whatsoever.
5) No patience or situational awareness.
6) STOCK mechs.
7) Run out of ammo with cheese builds like 6 AC2 Jaegermechs.

Yeah, it's been fun. Really... Wasn't this bad prior to ELO.

I also often get scrubs in my other 4, but my experiences are different than yours.

1) If I give them directions, they listen about 80% of the time. Taking commander seems to help.
2) If they don't listen we usually just follow them instead, even if it is a bad tactical move bcause staying together is often better than dying alone.
3) I will agree that the stock mechs are annoying. I will also say that they often behave in ways that don't make sense.


I am NOT a fan of how ELO averages out high (or even above average) ELO groups. It is an artifact of the fact that they have a system where they pick the ELO and then try to find players to fit that ELO. That would be fine for a single player queue, but for team play it is painful.

But it is not nearly as bad as you are indicating.

Edited by Sprouticus, 01 April 2013 - 05:34 AM.

[Joseph Mallan]

Corwin Vickers, on 01 April 2013 - 05:29 AM, said:

Well think about it, if you just started out and they give you a rating of 1000. All the other players around you have 1000 Elo. However, on your normal account you play at a rating of 1500. You're going to keep winning and gaining points till your rating is 1500, at that point you will start your 50/50 spread.

The way it works is if you win that first match at 1000 you take some of the points from the guys on the other team, say 20 points (depends on how much of a handicap you had vs the other team). Now you're at 1020 and they are at 980. You should be matched up with players who are around 1020 next time you play and they will be matched against players slightly lower than before.nd people who turn off certain game modes and you rapidly lose players to fill these matches with a close match.

If only this were true. My primary account (Joe) is still just shy of 61% W/L. The team winning or losing has nothing to do with How well I play. They win or lose even if I get 3 Kills and 4 assists or die in a hideous 5 on 1 surprise gang dakka with 30 damage dealt. Winning is a Team statistic and as a PUG my Team is different everytime.

[Gargel]

Arkikim, on 01 April 2013 - 01:16 AM, said:

ELO is simple, aggregate score balance and a predictor that says you stand this much of a chance of losing the match. It can still match you up against a team that has a 90% chance of winning the match, they just get less ELO from the win, which is about as stupid a system as I could imagine, considering they are punishing a handful of players by putting them on a horrible team instead of putting two horrible teams against one another. And it is entirely based off of wins/losses.

Yeah, it is as I thought then. It's a very bad way of balancing teams. It should be more like starcraft 2 ladder or similar. You start at some bracket and work your way up if you're better than others in the bracket. People in the same bracket play against each other. Your teammates and your enemies should be of the same skill level. Forcing losses or wins as it is now is ridiculous.

[Onmyoudo]

Joseph Mallan, on 01 April 2013 - 05:40 AM, said:

They win or lose even if I get 3 Kills and 4 assists or die in a hideous 5 on 1 surprise gang dakka with 30 damage dealt.

While I also 100% agree with your post, this made me laugh out loud.

[Corwin Vickers]

Have you watched this yet?







Essentially, on average, how well you do does affect how well your team does. The first video has a nice graph that shows the relationship.

Edited by Corwin Vickers, 01 April 2013 - 05:55 AM.

[Joseph Mallan]

Onmyoudo, on 01 April 2013 - 05:51 AM, said:

While I also 100% agree with your post, this made me laugh out loud.

Yeah... I know, it's funnier reading it than poking your head up over a ridge/rounding a corner and seeing 4 assaults and 2 Heavies drooling!

[Caleb Lee]

Gargel, on 01 April 2013 - 01:00 AM, said:

This I wondered throughout the thread as well. From what I can tell, the assumption made is that because he is dropping in a 4 man team, it is more difficult for the rating system to pair his team with equivalent opponents since, if we trust OPs numbers, they have an incredibly inflated "ELO"-rating, and this FORCES bad teammates and good opponents to reach a 50/50 win-loss.

ELO is per weight class. However if I understand it correctly, weight is not taken into account all that well when pairing players (If at all, i'm not entirely sure about this).

Elo is per weight class but doesn't do class matching at all. I.E. if you take a light, it doesn't match you up with a player in a light of equivalent skill.

Worst case scenario. You take a light, enemy counterpart is in a Stalker or Atlas with a 70 tonnage spread. Meanwhile, his equivalent Atlas on your team does 0-100 damage if that and if there even is a light on the enemy team he is easily taken out.

Class matching was all there was, there was never tonnage matching and now there's not even that nor does it pair you up with a similar Elo in the same class.

Exoth3rmic, on 01 April 2013 - 01:18 AM, said:

I'll be more than happy to comment on this thread once I see your play in action, or we can discuss your concerns privately.

It takes less than 5 mins to get a stream up and running and with tiny PC overheads these days so it won't impact your play if you're running at a quality that simply allows me to see your actual gameplay.

Here's how in 3 easy steps:

1 )Make an account on twitch (http://www.twitch.tv/),
2) Download open broadcaster @ http://obsproject.com/download
3) Follow the simple instructions here courtesy of PC Gamer

I look forward to seeing you in action!

Edit: Twitch saves your stream as a VOD, so we can watch it any time - you don't have to be there for the live stream.

Sorry, but this will probably never happen. For a variety of reasons:

1) Don't care enough to bother right now, others have also made the same case in other threads.
2) Have a life and little time: I.E. I'm the CEO of a small IT consulting business, am married, friends, play sports and other social outings, and in what spare time I get I try to play games as they are a hobby. I don't enjoy doing any more 'paperwork' or overhead then I absolutely have to.

Cheers.

Edited by Caleb Lee, 01 April 2013 - 12:20 PM.

[Caleb Lee]

Gargel, on 01 April 2013 - 05:49 AM, said:

Yeah, it is as I thought then. It's a very bad way of balancing teams. It should be more like starcraft 2 ladder or similar. You start at some bracket and work your way up if you're better than others in the bracket. People in the same bracket play against each other. Your teammates and your enemies should be of the same skill level. Forcing losses or wins as it is now is ridiculous.

Umm... no. My experience with Starcraft 2 was this:

1) I played single player all the way through, then I beat it on the hardest difficulty. Believe it or not, I'd never played the original Starcraft, was more into FPS back then and Mechwarrior 3/4 and Total War etc... than SC.
2) Then I completed a lot of the achievements and when I felt I'd memorized most of the builds and keys to be somewhat competitive, I started playing in a few groups online 3v3 etc... made some friends and had some fun.
3) Then in between games, I started playing in the solo ladders. I won 4 out of 5 of my first games during the 'trial' Elo period I guess and somehow ended up at either the platinum/diamond level brackets. Granted, towards the bottom but still...

My next game experiences were loss after loss with the occasional brilliant win where maybe an unorthodox strategy caught my opponent by surprise and or they just had a really off game.

I finally gave up playing at that level. I didn't LIKE the game that much and by then was getting tired of it with no desire to push to the next level.

So please, save me from Blizzard's attempt as well... that was NOT fun and it way overestimated my abilities.

Joseph Mallan, on 01 April 2013 - 05:54 AM, said:

Yeah... I know, it's funnier reading it than poking your head up over a ridge/rounding a corner and seeing 4 assaults and 2 Heavies drooling!

Aye, it happens to all of us. Usually I'm trying to hold the line with 3 other friends while the enemy is pushing and then I realize it's just us and I'm totally screwed as I come under the fire of the entire enemy team cause our other 4 decided to go cap without warning us, try some half brained flanking tactic with 0 communication, etc...

Edited by Caleb Lee, 01 April 2013 - 10:44 AM.

[Exoth3rmic]

Caleb Lee, on 01 April 2013 - 10:35 AM, said:

Redacted my reply in favour of a PM.

Edited by Exoth3rmic, 01 April 2013 - 01:23 PM.

[Caleb Lee]

Exoth3rmic, on 01 April 2013 - 11:00 AM, said:

Redacted my reply in favour of a PM.

Yeah, it didn't come across quite the way intended. Will edit post.

[Gargel]

Caleb Lee, on 01 April 2013 - 10:46 AM, said:

Umm... no. My experience with Starcraft 2 was this...

I disagree with you, the system has worked well for me and friends of mine. You get demoted if you lose too many matches. Of course it takes a fair number of losses or wins to get your rating to the level of your skill. I don't know how many matches you played but it's simply impossible to stay at platinum or any other bracket if you lose enough times. No way around it.

My point was that such a system would work better than the current one since you would be placed against opponents of similar skill instead of being forced a win/loss ratio. I don't really understand how they could think that was a good idea from the start. If done correctly with good skill determinants (this is of course not as simple in a team based game), teams of equal skill would even out at 50/50 win/loss anyway.

[Kristov Kerensky]

OP posts that Elo is fubar and it's causing his WLR and KDR to drop, he also complains that he can no longer do 2x4 sync drops or drop as an 8 man in the PUG queque.

Troll post is obvious.

[Caleb Lee]

Then again, you can have some great matches with great teammates. Just had 4 games in a row with good teammates, only me and a friend on comms in a 2 man. Looks like overall better average as all the fights have been hard fought and with only 2-4 left alive on winning team.

Overall though... been pretty crappy but may be onto something here. Guess I need to just group with one below average Elo and seem to get fair teams.

If that's the case, then it really bites cause I'll still want to play with my friends.

[warner2]

Caleb Lee, on 31 March 2013 - 07:39 PM, said:

Maybe I can get them to do 8v8 now after my experience the other night.

You really should. My experience is that there are teams of varying abilities using varied tonnages playing 8 v 8. It's best for your 8 man team to be fairly high in tonnage (e.g. 3 assault 3 heavy 2 scouts or whatever) so that you're not over penalized, but there is no-one forcing anyone to take 8 Atlases, and when you play 8 mans you know your team mates, you know you have no n00bs or non-team players, and you can play the game how it was meant to be played, in an organised fashion.

[Caleb Lee]

Kristov Kerensky, on 01 April 2013 - 12:24 PM, said:

OP posts that Elo is fubar and it's causing his WLR and KDR to drop, he also complains that he can no longer do 2x4 sync drops or drop as an 8 man in the PUG queque.

Troll post is obvious.

I did NOT complain about being able to do 2x4 man synch drop. Please learn to read. I complained that others were doing it.

I'm also glad the days when we formed 8 mans and rolled pugs are gone. However, I miss playing with 7 of my friends on a regular basis.

As 8v8 seem to be getting better and were fun again, maybe I can do more of those now.

There are very real issues with Elo as many others have stated. Your post looks a LOT more trollish than anything I ever posted.

[Yokaiko]

Kristov Kerensky, on 01 April 2013 - 12:24 PM, said:

OP posts that Elo is fubar and it's causing his WLR and KDR to drop, he also complains that he can no longer do 2x4 sync drops or drop as an 8 man in the PUG queque.

Troll post is obvious.

My W/L went from 2.8 some to like 1.5 and my KDR went from 2.8 to 3.3 at one point.

......then I decided to grind jenners and spiders the rest of the way, so I'm back down to 2.7 lol


Oh and I stopped pugging completely, its unbearable.

Edited by Yokaiko, 01 April 2013 - 12:39 PM.

[NinetyProof]

Caleb Lee, on 31 March 2013 - 06:39 PM, said:

It IS though is the problem... my Win/Loss is now like 2.5 and my K/D is down to 4.5.

Why are they implementing a SOCIALIST program that handicaps us. I'm losing far more than I'm winning now as a result of what I've posted above.

They are FORCING a losing situation and it really, really sucks.

You still don't get it ... if ELO is working right, your Win/Loss would be 1 ... and your K/D would be closer to 1 (depending upon the weight class your playing). The fact that your still over 1/1 win/loss ratio means your still not going up against equally skilled players. You should rejoice! It's means your way better then everybody else ... at least everybody else that queued in that 10 second window.

Sorry that you don't get to faceroll n00bs anymore and the system is actually matching you up in situations where you actually have to WORK TO WIN ... instead of just pounding the IWIN button.

Edited by NinetyProof, 01 April 2013 - 12:41 PM.