52 replies to this topic

[Appogee]

Note that the topic title is tongue in cheek*! So put down that flamer and let me tell you a story as curious as it is mysterious...

Last few days, I've been dropping in a lot of matches which go the following way...

1. Drop with a lot of assaults and heavies. So far so good.
2. Head off to battle together.
3. On initially engaging with the enemy, six of my team members get almost instantly annihilated by the enemy team. They are there one minute, dead the next.
4. Me and my remaining one other team member, now facing impossible odds of 8 on 2, get rolled over by the enemy team.

At the end of the match, me and the one other team member have scored maybe 30 in the brief time we were alive long enough to fight. Four of our team scored less than 10, and the other two less than 20. The enemy team has scores in the 40s to 60s, maybe a couple higher.

What do you reckon is going here...? Some kind of matchmaking quirk? It's happened frequently enough in the past few days for me to notice the pattern and ask this question.



* I'm not bad, but I'm certainly not awesome. I've come in around 80th while pugging in the recent tournaments, if that helps put my skill/experience in some kind of context.

Edited by Appogee, 17 June 2013 - 11:58 AM.

[Stoicblitzer]

here is my tongue-in-cheek reply: l2p and working as intended™.

i like how people argue that allowing 8 mans back into the pug queue will cause roflstomps....like they don't happen anymore.

<3

[JokerVictor]

Allow me to answer that by copying this from the Elo thread that's currently going.

....
The Elo system was developed for chess. An individual game. In that setting, the wins and losses are based solely on the individual player's skill and ability to adapt. Moving individual players toward a 1.0 W/L works extremely well in it's intended setting. Whether you beat a superior rated opponent or lose to an inferior rated one is entirely down to individual performance. The system itself does not have a hand in determining the outcome, though the system obviously does predict what the outcome is based on the player's ratings.

Now, let's apply that same system to a random team setting, like MWO. The goal of the system is now to move teams toward a 1.0 W/L, so by definition the system now has an active role in determining game outcome. Whether you win or lose in an Elo based team setting is almost entirely down to the matchmaker. If it determines you need to lose, you will most likely lose... and your individual skill will likely not matter at all. If the system predicts you'll lose, it's because the teams are stacked and not because you are individually ranked lower than your opponent and visa versa.

It's an awful way to balance skill. It screws literally everyone involved. The high skill players get saddled with teams of boat anchors and the low skilled players are forced to compete with much higher skilled players on an almost constant basis. Or, in this game's case, generally much better equipped players are playing against and with players with terrible equipment.

And, in this game, having even 2 extremely below average players on your team is an almost unbeatable crutch, if the other team doesn't have 2 correspondingly bad players. It completely destroys any sense of 'fairness' in losing or winning because the teams are generally built from such wide deviation averages.

Add in the fact that weight class matching is non-existent and you have the giant t*rd of a match maker we've currently got. Another great idea executed PGI, way to go.

[Appogee]

That's what I was thinking... maybe some kind of weird averaging of wildly divergent skill levels.

Good thought on the equipment levels, too. I will have to pay more attention to how many trial mechs I'm dropping with. If a lot of my team is fighting in underpowered overheating standard chassis variants, then it's no wonder they;re dying so quickly.

But I'd just prefer to be dropping with teams of about my level, and fighting against teams of about my level, rather than these one-sided ROFLstomps.

[Dracol]

Can't speak specifically about your matches, but I will share with you an observation I have seen often.

After initial contact with the enemy, the front most mech in the congo line of pugs will either stop and exchange fire or continue forward towards the enemy. The congo line continues forward, and each one dies in turn.

Those who have more experience will do two things.
First: the spread out along a firing line so that when the enemy is engaged, they have multiple targets to fire at instead of the entire team focus firing down each individually.

Second: After initial contact and receiving damage, the more experienced players will back up into cover and regroup with their team.

[Lostdragon]

I typically find that the first team to lose a player will generally lose the match, especially if the first kill is a heavy or assault. Once the numbers are against you it is really hard to make up for that.

[Zerberus]

I love all the "Elo is not useful for team games" hate.

Major league baseball, the NBA, and college football among others would all like a word with you people.

http://en.wikipedia....gs_beyond_chess

Yes, it wa ORIGINALLY designed for sole 1:1 situations (specifically chess)

So was teh telephone.

and like the telephone, it was adapted to fit new requirements (conference calling, video calling..).

The most hilarious of the complaints are the ones about being "stuck in Elo hell".



TLDR: Learn to analyze your own playing for mistakes, and noot just other`s, and always assume that when you lose it was (at least partially) your fault and not entirely everyone elses, becasue statistically it IS at least partially your fault. A large part of becoming a competitive player in ANY game /sport is learning to think like one instead of assuming that you are already as good as you can get. Becasue when you stop getting better, you cease to be good.

Edited by Zerberus, 17 June 2013 - 12:20 PM.

[Appogee]

Zerberus, on 17 June 2013 - 12:08 PM, said:

The most hilarious of the complaints are the ones about being "stuck in Elo hell".

That's the place where ''Don't Bring Me Down'' plays on endless rotation.

[JokerVictor]

Zerberus, on 17 June 2013 - 12:08 PM, said:

I love all the "Elo is not useful for team games" hate.

Major league baseball, the NBA, and college football among others would all like a word with you people.

http://en.wikipedia....gs_beyond_chess

Yes, it wa ORIGINALLY designed for sole 1:1 situations (specifically chess)

So was teh telephone.

and like the telephone, it was adapted to fit new requirements (conference calling, video calling..).

The most hilarious of the complaints are the ones about being "stuck in Elo hell".

Yes, the BCS is a great example of how awesome Elo ratings can be in team settings. Lulz.

Also, these are ORGANIZED TEAM EVENTS. Not random dookie matches arranged by a terrible algorithm.

Edited by JokerVictor, 17 June 2013 - 12:17 PM.

[Stealthsfury]

I will say that the system seems to work better during prime time hours. Some closer matches but not by much.

With that being said. The vast majority of my PUG drops seem to be all about lambs to the slaughter. I often recognize several names on the other team. While this in itself is not a bad thing it does give me some insight as to if I am facing a group. Besides the great coordination that presents itself with my team being utterly steam rolled or just our complete lack of anything planned. This is all the evidence I need.

I figured when I PUG dropped that there would be at least one group on our side to lead our side by their example. I have rarely ever seen anyone take commander in the game. I rarely see my side being able to move together with any goal in mind.

I don't want to take command and lead a group when I pug. I don't even lead when I do drops with my unit. My unit often takes command and communicates with our other PUG. So I either wait for my group to log on or I just sit there and let the ELO gods decide my fate.

It is this level of being powerless that I feel hurts people's enjoyment of the game.

Before ELO I felt I could be a real deciding factor. Now it is just get my kills, fend for myself and help my own grind before caring about a W/L.

Edited by Stealthsfury, 17 June 2013 - 12:29 PM.

[Lostdragon]

I am unfortunately not able to play with a group. I solo drop only and I feel like my matches fall pretty evenly into 3 categories: Good fights, we stomp them, and they stomp us. Often times when either sort of stomping is occurring I notice a higher level of coordination among some of the stompers. It is easy to spot groups that are on voice chat because they actually respond to and conduct flanking and ambushes in a coordinated fashion.

So teamwork is indeed OP

[Unbound Inferno]

JokerVictor, on 17 June 2013 - 11:59 AM, said:

Allow me to answer that by copying this from the Elo thread that's currently going.

....
The Elo system was developed for chess. An individual game. In that setting, the wins and losses are based solely on the individual player's skill and ability to adapt. Moving individual players toward a 1.0 W/L works extremely well in it's intended setting. Whether you beat a superior rated opponent or lose to an inferior rated one is entirely down to individual performance. The system itself does not have a hand in determining the outcome, though the system obviously does predict what the outcome is based on the player's ratings.

Now, let's apply that same system to a random team setting, like MWO. The goal of the system is now to move teams toward a 1.0 W/L, so by definition the system now has an active role in determining game outcome. Whether you win or lose in an Elo based team setting is almost entirely down to the matchmaker. If it determines you need to lose, you will most likely lose... and your individual skill will likely not matter at all. If the system predicts you'll lose, it's because the teams are stacked and not because you are individually ranked lower than your opponent and visa versa.

It's an awful way to balance skill. It screws literally everyone involved. The high skill players get saddled with teams of boat anchors and the low skilled players are forced to compete with much higher skilled players on an almost constant basis. Or, in this game's case, generally much better equipped players are playing against and with players with terrible equipment.

And, in this game, having even 2 extremely below average players on your team is an almost unbeatable crutch, if the other team doesn't have 2 correspondingly bad players. It completely destroys any sense of 'fairness' in losing or winning because the teams are generally built from such wide deviation averages.

Add in the fact that weight class matching is non-existent and you have the giant t*rd of a match maker we've currently got. Another great idea executed PGI, way to go.

Interesting analysis. If that holds true, then our ELO is comparable to our individual win/loss ratio on our mechs? Maybe balanced by our overal win/loss?

I really wish the devs would cough up what the hell they concoted here so we could grasp how its 'supposed to' be.

[Zerberus]

JokerVictor, on 17 June 2013 - 12:16 PM, said:

Yes, the BCS is a great example of how awesome Elo ratings can be in team settings. Lulz.

One counter-example has never invalidated any rule, but nice try. Show me a list of where Elo is obviously failing in practice (not debatably failing such as here and LOL) that is at least half as long as the number of organizations and games that swear by it and maybe you`ll have a leg to stand on.

Unbound Inferno, on 17 June 2013 - 12:41 PM, said:

Interesting analysis. If that holds true, then our ELO is comparable to our individual win/loss ratio on our mechs? Maybe balanced by our overal win/loss?

I really wish the devs would cough up what the hell they concoted here so we could grasp how its 'supposed to' be.

Ask and you shall receive. I present to you the Matchmaking Dev Threads, all on page 1 of the command chair.

http://mwomercs.com/...79-matchmaking/
http://mwomercs.com/...-making-update/
http://mwomercs.com/...ted-april-19th/

And for good measure, the wikipedia enty on Elo matchmaking (Warning, advanced math and science content, it should come as no surprise that most people without a college education cannot grasp the underlying concepts ):

http://en.wikipedia....o_rating_system

P.S.: I am NOT arguing that Elo is perfect for our use, nor am I even going so far as to theorize that it might be. I AM directly saying that 95%+ of teh pople her do not understand how it works or is supposed to work, or even care enough to try to understand it. Nor do I see anybody presenting any form of algoritm that could be considered a viable alternative.

It`s easy to complain and say "It sucks, make it better"... actually contributing something useful towards that improvement is, however, an entirely differnt ballgame, and to be frank not something most of us are even remotely capable of doing as many posts in thesse threads present hard evidence of.

Edited by Zerberus, 17 June 2013 - 01:20 PM.

[JokerVictor]

Zerberus, on 17 June 2013 - 01:08 PM, said:

And for good measure, the wikipedia enty on Elo matchmaking (Warning, advanced math and science content, it should come as no surprise that most people without a college education cannot grasp the underlying concepts ):

http://en.wikipedia....o_rating_system

P.S.: I am NOT arguing that Elo is perfect for our use, nor am I even going so far as to theorize that it might be. I AM directly saying that 95%+ of teh pople her do not understand how it works or is supposed to work, or even care enough to try to understand it. Nor do I see anybody presenting any form of algoritm that could be considered a viable alternative.

It`s easy to complain and say "It sucks, make it better"... actually contributing something useful towards that improvement is, however, an entirely differnt ballgame, and to be frank not something most of us are even remotely capable of doing as many posts in thesse threads present hard evidence of.

Guess what, I have a math and science background. Elo is terrible for random team matching. I have a feeling you are part of the 95% that has no clue how the math behind the system actually operates.

Quote

How does the match maker compose a teams Elo rating, is it average rating or closest to a target?

It's closest to a target value, so the match maker starts trying to make a match for an Elo of say 1300 and will pull in players to those teams closest to those values; however, as mentioned earlier within growing thresholds and those curves will be tuned. Currently it may be a bit 'sloppy' about how it's filling those buckets but over time it will be tuned to be much more precise.

We need to do this carefully over time as generally the cost of precision is time to find a match we want to slowly find a very nice balance between time to find a match and the number of matches that are correctly composed.

As the player populations go down, this gets worse... and worse.... and worse. 'Sloppy' indeed.

[Pater Mors]

Actually... While Elo was the original rating system for Chess, the better way to measure yourself in Chess these days is Glicko, not Elo and rating systems are slowly changing to incorporate a reading of both. It gives a much fairer reading of your skill than your Elo.

No actual bearing on this thread... I just play a lot of Chess.

http://en.wikipedia....o_rating_system

[Zerberus]

JokerVictor, on 17 June 2013 - 01:27 PM, said:

Guess what, I have a math and science background. Elo is terrible for random team matching. I have a feeling you are part of the 95% that has no clue how the math behind the system actually operates.

Don`t you have a list to work on? Or was one counter example all you could dig up?

Alternatively: where is your improved solution? Or do you prefer to just take the easy way out and complain despite your higher education being beneficial to your cause, further underlining my point?

BTW, You consistently completely disregard the fact that all but the most FUBARed Team-Based elo rankings are not using the same algorithms as 1:1 pairings, but a highly modified system. Again, like the telephone, Elo has evolved.

Ironically, irrelevant to your unfounded personal attack, you agree with me that 95% don`t understand how the system works, and therefore by logical correlation should not be acting like they do in their complaints.

Any more points of mine you would like to inadvertently back up?

Edited by Zerberus, 17 June 2013 - 01:48 PM.

[Fyrerock]

I work 12 hour shift work, so my play time is all over the place and I have noticed how aggressive I am able to play depends a lot on what time I log into the game. Early saturday mornings I can do almost anything I want and get very little damage. Prime time if I stick my head above a hill for more then 5 seconds I will take a lot of damage. I have a feeling that how well they can match up players depends a lot on the number of players playing at one time, and off prime time they will pick almost any skill level to play with each other.

[Ningyo]

Much of the problem with the ELO system they have stems from it rewarding everyone on the team equally for a win or a loss.

For example if My team average ELO 1600, goes against enemy team average ELO 1700 and:

We lose 0:8, with me doing 50 damage, my ELO drops the same as if
We lose 7:8 with me getting 7 kills and 1500 damage, and the rest of my team getting under 100 dmg
(no I have never done this well)

This would be fine IF the teams were constant, because then its a TEAMs ELO. For individual ELO though basing it all on a random teams performance that changes constantly is really going to fail.


What they need to do is have it be based on a formula such as: that includes how you do in your match (match score is probably a decent thing to use, though really would be good to include things like capturing points and such. IN addition to who wins or loses)

Admittedly this would add some complexity to properly balancing the ELO gain and loss formulas. But it also might actually work.

[JokerVictor]

Zerberus, on 17 June 2013 - 01:42 PM, said:

Don`t you have a list to work on? Or was one counter example all you could dig up?

Alternatively: where is your improved solution? Or do you prefer to just take the easy way out and complain despite your higher education being beneficial to your cause, further underlining my point?

Ironically, irrelevant to your unfounded personal attack, you agree with me that 95% don`t understand how the system works, and therefore by logical correlation should not be acting like they do in their complaints.

Any more points of mine you would like to underline?

I do have to provide a better alternative to be right? No, I don't. It's not my job to design these systems, but I'm perfectly qualified to point out their flaws as an end user.

If you don't want unfounded personal attacks thrown your way, then leave them out of your counter argument. Saying 95% of people 'don't understand this' as an aside to your crap examples is called a logical fallacy. L2Argue.

Quote

Irrelevant Appeals

Explanation

Irrelevant appeals attempt to sway the listener with information that, though persuasive, is irrelevant to the matter at hand. There are many different types of irrelevant appeal, many different ways of influencing what people think without using evidence. Each is a different type of fallacy of relevance.

Currently the only game besides this one that uses Elo as a matching system in random matches is League of Legends. Out of all of your examples, this is the only valid thing to compare it to, because in any ORGANIZED TEAM SPORT (as I pointed out earlier) using an Elo ratings system evaluates the entire team. A team which doesn't change from game to game.

And even taking all that into account, League of Legends isn't directly comparable. There are entirely too many variables at play in team composition in MWO to get a good read of performance and skill based solely on win/loss. Then add in weight class matching and the wildly different capabilities of the different mech classes on top of that and you've got a mess.

Going back to LoL... LoL also has pre-game lobbies, so the newly formed team can quickly set themselves up before being thrown to the wolves. Currently, we get a gaggle of idiots with completely random everything except for a roughly 'equal' average elo score between the teams. And you wonder why there are all these threads popping up asking why it's completely terrible 80-90% of the time?

Oh right, it's because 95% of people just 'don't understand' a terrible system when they see one.

[Zerberus]

JokerVictor, on 17 June 2013 - 02:00 PM, said:

Currently the only game besides this one that uses Elo as a matching system in random matches is League of Legends. Out of all of your examples, this is the only valid thing to compare it to, because in any ORGANIZED TEAM SPORT (as I pointed out earlier) using an Elo ratings system evaluates the entire team. A team which doesn't change from game to game.

No, it isn`t.

CS:Go uses Elo, as does Guild Wars, WoW used it for the Arena until 2010, and numerous other online games which would require more than precursorily reading the sources I freely provided in a previous post. If you don`t read documents handed to you, you are handicapping yourself. L2Research in a manner that someone with a "science background" should have been taught. This is stated in this form solely in counterpoint to your equally condescending L2Argue.

In light of your proven capacity for in depth research as well as your obviously holier than thou self-perception, I don`t think I need to concern myself with the rest of your statement or with any future ones.

Edited by Zerberus, 17 June 2013 - 02:17 PM.