202 replies to this topic

[Deathlike]

aniviron, on 06 April 2014 - 09:00 PM, said:

Tangentially related to the matchmaker; I too feel like when it rains it pours in regards to maps. Most nights I play, I get about 6-7 games, and at least four or five will be on the same map, almost without fail. What map that is changes randomly, but if I start the night off with one map, I'm going to keep dropping on that map over and over, it seems. Does anyone else get that?

It could be that PGI forgot to initialize the seeding for randomize().

But, I personally cannot get that out of my system.

The game to me at time runs like this:
1) Play map (say Mordor)
2A) Drop into same map, particularly after a loss
2B ) Drop into same map, after a win (especially if it's Mordor)
3) Massively gripe as I'm probably on the same map again
4) Curse the RNG after dropping into same map after dropping into a different map later

It's a theme.

Joseph Mallan, on 07 April 2014 - 11:00 AM, said:

lol pic

Where did you find this?

[Charons Little Helper]

Abivard, on 07 April 2014 - 12:58 PM, said:

It does not work in randomly assigned team multiplayer games. This is fact, not opinion, everyone is entitled to their own opinions, they are not entitled to their own facts!

Yes it does.

That is a fact.

Look it up.

Learn math.

[Abivard]

[

Charons Little Helper, on 07 April 2014 - 01:56 PM, said:

Yes it does.

That is a fact.

Look it up.

Learn math.

Show me the link, I can not seem to find anything that says you are correct that is from any sort of reliable source. Could it be because it simply does not exist? that would explain why no one can seem to find these facts you claim are proven.

[Abivard]

Well here is a link for my case:
http://research.micr...ll/details.aspx

I am sure you will be able to follow the math, it isn't very hard at all.

[Targetloc]

Abivard, on 07 April 2014 - 02:18 PM, said:

Well here is a link for my case:
http://research.micr...ll/details.aspx

I am sure you will be able to follow the math, it isn't very hard at all.

I'm not sure how that helps your case.

It states Elo was not designed for team games. Okay.
It claims it can find the proper ranking for players faster. Okay.

It doesn't claim that Elo doesn't work.
It uses an extremely similar system to what we have now. It doesn't take weapons or tonnage or hard-counters into account. All it looks at is win/loss.

Quote

But how does the TrueSkill ranking system incorporate the game outcome of a team match? In this case, the team's skill is assumed to be the sum of the skills of the players. The algorithm determines the sum of the skills of the two teams and uses the above two equations where (μwinner,σ2winner) and (μwinner,σ2loser) are the mean skills and skill variances of the winning and losing team, respectively.

[Roadkill]

Abivard, on 07 April 2014 - 02:18 PM, said:

Well here is a link for my case:
http://research.micr...ll/details.aspx

I am sure you will be able to follow the math, it isn't very hard at all.

You're right, it isn't very hard at all.

What you're wrong about is your conclusion, which has been explained to you in great detail before. Yet you continue to spout this nonsense about Elo not working in multiplayer games.

Trueskill is derived from Elo. It is, at its core, an Elo system. You can't say on one hand that Elo doesn't work for multiplayer games, and then on the other hand that Trueskill is proof of that because Trueskill is Elo.

Pretty much all Trueskill does is help speed up how quickly your rating stabilizes. That's it.

[Abivard]

Targetloc, on 07 April 2014 - 03:04 PM, said:

I'm not sure how that helps your case.

It states Elo was not designed for team games. Okay.
It claims it can find the proper ranking for players faster. Okay.

It doesn't claim that Elo doesn't work.
It uses an extremely similar system to what we have now. It doesn't take weapons or tonnage or hard-counters into account. All it looks at is win/loss.

You are being mystified. Not a single person has ever claimed Elo doesn't work for what it was designed for.

The point is, it can not work for MWO or any other random team multi-player game.

If it could, NO ONE would be working on a system that CAN work!

[Roadkill]

Abivard, on 07 April 2014 - 03:19 PM, said:

The point is, it can not work for MWO or any other random team multi-player game.

If it could, NO ONE would be working on a system that CAN work!

WRONG.

My Ferrari is faster than your Toyota. Does that mean your Toyota doesn't work?

The work on systems like Trueskill is to speed up how quickly your rating stabilizes. Trueskill is faster than Elo in a multiplayer environment.

That does not mean that Elo doesn't work. In fact, it proves that Elo does work. Trueskill (and others like it) just work faster.

[Roland]

While it's true that elo rating should theoretically work, it is possible that there is so much obfuscation in the signal due to the impact of the other 23 players in the game, that it may take an impossibly large number of games in order for the system to arrive at a correct ranking.

So, for instance, it may take the elo system a million games to correctly arrive at a meaningful rating for a player in this game. In such a case, while it works in the abstract theoretical sense, it doesn't work in any practical sense because it takes to many games to arrive at a correct rating.

[Abivard]

Roadkill, on 07 April 2014 - 03:18 PM, said:

You're right, it isn't very hard at all.

What you're wrong about is your conclusion, which has been explained to you in great detail before. Yet you continue to spout this nonsense about Elo not working in multiplayer games.

Trueskill is derived from Elo. It is, at its core, an Elo system. You can't say on one hand that Elo doesn't work for multiplayer games, and then on the other hand that Trueskill is proof of that because Trueskill is Elo.

Pretty much all Trueskill does is help speed up how quickly your rating stabilizes. That's it.

Do you read what you write? really, do you?
Trueskill is Elo? If you can not comprehend how stupid and illogical that statement is, you have no place in any kind of discussion such as this.
Everything under the sun DERIVES from something else. That does not mean everything is the same nor does it mean that a saw will work as a hammer, they are both tools but perform different functions.

IF Elo as it was first presented worked across everything in the entire universe without modification as you seem to assert.

Why are so many DIFFERENT systems being used or researched because basic Elo is giving false results?

Why do the chess federations no longer use the original Elo?

Why can you not provide a single scholarly reference to basic Elo working in this environment?

[Roadkill]

Roland, on 07 April 2014 - 03:30 PM, said:

While it's true that elo rating should theoretically work, it is possible that there is so much obfuscation in the signal due to the impact of the other 23 players in the game, that it may take an impossibly large number of games in order for the system to arrive at a correct ranking.

So, for instance, it may take the elo system a million games to correctly arrive at a meaningful rating for a player in this game. In such a case, while it works in the abstract theoretical sense, it doesn't work in any practical sense because it takes to many games to arrive at a correct rating.

I think you participated in that "other" thread as well.

It's not an impossibly large number of games. It's no more than 200-300, and may be as few as 50 depending on how confident you want your ratings to be.

But perhaps more to the point, the rating variability introduced by multiplayer Elo is entirely washed out in the noise caused by PGI's matchmaker, which itself is derived from the small player base. When your matchmaker is starting with a window of +/- 175 points and rapidly expands its search to +/- 1000 points, that tells you everything you need to know.

It's not Elo. It's the matchmaker.

[Khobai]

Quote

Elo by itself in a one on one works, Elo in a random multiplayer match like MWO does not work.

Actually ELO does work in multiplayer games provided one person can significantly alter the outcome of the game. That was very much the case in 8v8 and I would argue that ELO could indeed work in 8v8.

In 12v12 however its very difficult for one person to alter the outcome of the game no matter how good they play. So in 12v12 Elo doesnt really work.

[Roadkill]

Abivard, on 07 April 2014 - 03:32 PM, said:

Why can you not provide a single scholarly reference to basic Elo working in this environment?

You've never read them when they were provided in the past, so why should I link them again?

[Roadkill]

Abivard, on 07 April 2014 - 03:32 PM, said:

Why are so many DIFFERENT systems being used or researched because basic Elo is giving false results?

It isn't giving false results. The new systems arrive at the same valid ratings more quickly than Elo. And that is why they're being researched. Not because Elo doesn't work, but because the new systems work faster.

I've never claimed that Elo is perfect. It is possible to improve upon Elo.

You, however, have claimed that Elo doesn't work. You are wrong. Elo does work in this environment. It's mathematically proven to work.

[Abivard]

Roland, on 07 April 2014 - 03:30 PM, said:

While it's true that elo rating should theoretically work, it is possible that there is so much obfuscation in the signal due to the impact of the other 23 players in the game, that it may take an impossibly large number of games in order for the system to arrive at a correct ranking.

So, for instance, it may take the elo system a million games to correctly arrive at a meaningful rating for a player in this game. In such a case, while it works in the abstract theoretical sense, it doesn't work in any practical sense because it takes to many games to arrive at a correct rating.

Which I could agree to in Theory, but!

Also in theory, because of all the obfuscation or static of the other 23 players your true rating will never be able to be discerned through all of the clutter. As the sample size increases so do ALL of the results of ALL the players.

The law of large numbers as so often quoted and misapplied must be taken in the following context:

[color=#000000]With a large enough sample, any outrageous thing is likely to happen (Diaconis and Mosteller 1989). [/color]

[Macbrea]

You are as bad as the other 23 players in your match. Their teams average elo is close to your teams average elo. And almost always, your elo is close to your teams. I am not sure why people feel they are super special and should get their own bracket.

[Roadkill]

Abivard, on 07 April 2014 - 03:43 PM, said:

The law of large numbers as so often quoted and misapplied

ROFLMAO.

Actually, this is pretty much exactly what the law of large numbers is all about.

[Roland]

Roadkill, on 07 April 2014 - 03:34 PM, said:

I think you participated in that "other" thread as well.

It's not an impossibly large number of games. It's no more than 200-300, and may be as few as 50 depending on how confident you want your ratings to be..

I don't believe you have any mathematical basis for this statement, at all.

If you believe otherwise, please show your math, but I seriously doubt that you are considering the obfuscating factors at hand.

Edited by Roland, 07 April 2014 - 03:48 PM.

[Roadkill]

Not getting into another one of these threads.

Abivard has his point, and it's wrong, but he can't see that, and it's his point dammnit so he's going to continue stating it over and over and over and over and over and over and over and over and over and over and over and over and over and over and over...

until everyone else gets tired of explaining it to him and gives up. So I'll just pre-empt the shenanigans and leave early this time.

TL;DR Elo works fine in multiplayer environments. Abivard is convinced otherwise despite multiple research papers that have been presented to him in the past, so there's no point in arguing with him.

[BLOOD WOLF]

Roadkill, on 07 April 2014 - 03:51 PM, said:

Not getting into another one of these threads.

Abivard has his point, and it's wrong, but he can't see that, and it's his point dammnit so he's going to continue stating it over and over and over and over and over and over and over and over and over and over and over and over and over and over and over...

until everyone else gets tired of explaining it to him and gives up. So I'll just pre-empt the shenanigans and leave early this time.

TL;DR Elo works fine in multiplayer environments. Abivard is convinced otherwise despite multiple research papers that have been presented to him in the past, so there's no point in arguing with him.

well stated.