Stat Based MM

[Dragnorak]

There have been a lot of discussions over the years re the quality of ones team and not forgetting of course the dirge of 40% Wr players.

 

So with a bit of thought would a stats based MM actually work do you think so that you are matched up more often with players of your skill level where available?

[Egoleter]

Wargaming will not implement any form of skill based matchmaking. It will also not solve the percieved problem in any way.

[Dragnorak]

7 minutes ago, Egoleter said:

Wargaming will not implement any form of skill based matchmaking. It will also not solve the percieved problem in any way.

 

Well I think they are missing a trick. You could have 3 levels of server for example and you choose where you want to play. It could be implemented in many ways that are dooable.

 

Up to 48% WR beginners server

49% to 56 average to good

57 and above Unicum tuff

 

I dunno but there could be many ideas around this if it was thought through. Surely they would retain more customers as the potatoes would be happy being potatoes and the good players would not keep getting frustrated with said taters as they would never come across them!

 

 

[ColonelPete]

11 minutes ago, Dragnorak said:

There have been a lot of discussions over the years re the quality of ones team and not forgetting of course the dirge of 40% Wr players.

 

So with a bit of thought would a stats based MM actually work do you think so that you are matched up more often with players of your skill level where available?

No, as discussed often enough.

 

According to maplesysrup, players with 58%+ winrating make up around 2.3% of the playerbase on the asia server.

Good luck finding teams for 62% WR players

http://maplesyrup.sweet.coocan.jp/wows/ranking/tableofaccounts.html

 

 

[Shaka_D]

The simply truth is its been discussed on numerous forums and will likely never happen, but to be honest, I don't mind having people off all abilities on my team - just as long as the opposing team has a similar number / spread of mixed ability so we all have a reasonably similar number of noobs, newbies, average Joe's, semi pro's and pro's. It's when the mm by its random nature pits a majority group of good players versus a majority group of baddies against each other that causes problems

[Pikkozoikum]

1 minute ago, Dragnorak said:

 

Well I think they are missing a trick. You could have 3 levels of server for example and you choose where you want to play. It could be implemented in many ways that are dooable.

 

Up to 48% WR beginners server

49% to 56 average to good

57 and above Unicum tuff

 

I dunno but there could be many ideas around this if it was thought through. Surely they would retain more customers as the potatoes would be happy being potatoes and the good players would not keep getting frustrated with said taters as they would never come across them!

 

 

That would devide the playerbase pretty much. A server with only 57% and higher would end to a low playerbase-server

 

What I just would wish, if the would a matchmaking with personal-team-rating

 

The the different PRs of different players sums to a total number, which has to fit with the total number of the enemy team. In a case of a 3v3 to make it easier

 

A unicorm has 1800 pr

average as 1500

and bad player with 1200

=  4500 PR total

 

the other team has only average players but no unicums

1550

1550

1400

=4500 PR total

 

The match would be some how balanced without having only unicums in on team and only bad players in the other team

 

[Yosha_AtaIante]

Vor 21 Minuten, Dragnorak sagte:

There have been a lot of discussions over the years re the quality of ones team and not forgetting of course the dirge of 40% Wr players.

 

So with a bit of thought would a stats based MM actually work do you think so that you are matched up more often with players of your skill level where available?

It will be the same as always. The 60% players will drop to 50% and the 70% will be the new 60%.

 

[Dragnorak]

4 minutes ago, ColonelPete said:

No, as discussed often enough.

 

According to maplesysrup, players with 58%+ winrating make up around 2.3% of the playerbase on the asia server.

Good luck finding teams for 62% WR players

http://maplesyrup.sweet.coocan.jp/wows/ranking/tableofaccounts.html

 

 

 

Strewth I didnt know it was that low. Where can you see that? 

 

So most players then fall below 58?

[Dragnorak]

2 minutes ago, Pikkozoikum said:

That would devide the playerbase pretty much. A server with only 57% and higher would end to a low playerbase-server

 

What I just would wish, if the would a matchmaking with personal-team-rating

 

The the different PRs of different players sums to a total number, which has to fit with the total number of the enemy team. In a case of a 3v3 to make it easier

 

A unicorm has 1800 pr

average as 1500

and bad player with 1200

=  4500 PR total

 

the other team has only average players but no unicums

1550

1550

1400

=4500 PR total

 

The match would be some how balanced without having only unicums in on team and only bad players in the other team

 

 

Actually yes that's not a bad idea to be honest.

[DFens_666]

6 minutes ago, Pikkozoikum said:

The the different PRs of different players sums to a total number, which has to fit with the total number of the enemy team. In a case of a 3v3 to make it easier

 

A unicorm has 1800 pr

average as 1500

and bad player with 1200

=  4500 PR total

 

the other team has only average players but no unicums

1550

1550

1400

=4500 PR total

 

The match would be some how balanced without having only unicums in on team and only bad players in the other team

3 minutes ago, Dragnorak said:

Actually yes that's not a bad idea to be honest.

 

Thats just horrible. Thats what i exprienced during the first 2 ranked sprint seasons. The games just felt so iffy. And i started to check the players afterwards, and almost every match it was like this:

Division mate and me both with >62% WR. Our other 4 Guys had the lowest WR from all the 12 players in the game, usually like 40-48%.

On the enemy Team, we had 48-55%~ WR players. Basicly that made both teams have an equal WR. Im not sure if that was a coincedence, but maybe they tested something? Im pretty sure that it was rigged

 

Anyway, those games basicly ended in a loss always. 4 extremely bad players just died because they cant angle and cant shot and dont know what objectives are. You cant kill above average players that fast, as they have atleast some brains to migitate damage. And playing 2vs6... almost impossible, especially on T5 (what it was back then).

Those were one of the most horrible games ive played, also a reason why i dont play ranked sprint anymore after that, only a couple.

[Pikkozoikum]

8 minutes ago, Dragnorak said:

 

Actually yes that's not a bad idea to be honest.

Yes, even personal ratings are already caluclated by sides like wows-numbers.com/

 

[Jethro_Grey]

5 minutes ago, Pikkozoikum said:

That would devide the playerbase pretty much. A server with only 57% and higher would end to a low playerbase-server

 

What I just would wish, if the would a matchmaking with personal-team-rating

 

The the different PRs of different players sums to a total number, which has to fit with the total number of the enemy team. In a case of a 3v3 to make it easier

 

A unicorm has 1800 pr

average as 1500

and bad player with 1200

=  4500 PR total

 

the other team has only average players but no unicums

1550

1550

1400

=4500 PR total

 

The match would be some how balanced without having only unicums in on team and only bad players in the other team

 

Somehow? One does not magically balance Things. 

 

Besides, you can't only take PR as criteria to balance teh MM, that wouldn't work. 

You have to take ships and their weighting into account too. 

 

Imagine a division of super unicums with a few average players mixed in. 

In order to balance that you either have to find equally strong players, or if the MM can't find any in their range, give the other team way more players. 

And with the amount of bad and below average players in the playerbase, you'd have 6players vs 20 players games. (exagerated)

 

This would create an Imbalance, so no. PR based, or even skill based MM in WoWS is a phantasy that'll never come true. 

And if it ever would, it'd turn into a nightmare pretty damn fast. 

[__Helmut_Kohl__]

Your dreams of "perfect MM balance" would mean more or less 50% WR for everyone.

If you are playing well, you just get punished by MM.

 

Why even care to get better then?

I suggest instead that you do just that and stop blaming your high tier losses on MM.

 

[Pikkozoikum]

19 minutes ago, DFens_666 said:

 

Thats just horrible. Thats what i exprienced during the first 2 ranked sprint seasons. The games just felt so iffy. And i started to check the players afterwards, and almost every match it was like this:

Division mate and me both with >62% WR. Our other 4 Guys had the lowest WR from all the 12 players in the game, usually like 40-48%.

On the enemy Team, we had 48-55%~ WR players. Basicly that made both teams have an equal WR. Im not sure if that was a coincedence, but maybe they tested something? Im pretty sure that it was rigged

 

Anyway, those games basicly ended in a loss always. 4 extremely bad players just died because they cant angle and cant shot and dont know what objectives are. You cant kill above average players that fast, as they have atleast some brains to migitate damage. And playing 2vs6... almost impossible, especially on T5 (what it was back then).

Those were one of the most horrible games ive played, also a reason why i dont play ranked sprint anymore after that, only a couple.

4 extrem bad players and 2 extrem good players vs 6 average is not a balance. 3 Bad and 3 good is a balance against and average team. Thus the PR would likely to swap one extrem bad with an average, and you would have in your case a more fair team.

 

Beside that. You got this result without any matchmaking, that's the evidence, that it is purely random and not even close to fairness. You can end up with a team full of bad players, while the team ends up full with good players. No balance. Balancing with PR would balance it.

 

 

You have also to consider, that in your case it's impossbile to generate a real fair team, since all unicums are on one side. No fair team balance possible. So i can only balanced for a teams or generate unfair teams with all the unicums on one side.

 

To prevent that, it could be also possible to build a filter for ranked games. That players with rating from mimium rating up to 1600 get matched and in another filter only players with 1500 rating and maximum rating get matched

Thus there are matches where the highest rating is 1600 from bad players up to average players and then there are matches with only average players up to unicums.

[Dragnorak]

4 minutes ago, __Helmut_Kohl__ said:

Your dreams of "perfect MM balance" would mean more or less 50% WR for everyone.

If you are playing very well, you just get punished by MM.

 

Why even care to get better then?

I suggest instead that you do just that and stop blaming your losses on MM.

 

 

Here we go the first post to diverge from my original question to a veiled personal attack on me.

 

So just to SHUT you up before you get on your high horse:

 

I was NOT blaming any of my losses on MM

 

Got IT?>???

 

I was attempting to start a conversation on the merits of a better MM system than we currently have. Maybe you struggle to read and understand English in which case I forgive you.

[Dragnorak]

8 minutes ago, DFens_666 said:

 

Thats just horrible. Thats what i exprienced during the first 2 ranked sprint seasons. The games just felt so iffy. And i started to check the players afterwards, and almost every match it was like this:

Division mate and me both with >62% WR. Our other 4 Guys had the lowest WR from all the 12 players in the game, usually like 40-48%.

On the enemy Team, we had 48-55%~ WR players. Basicly that made both teams have an equal WR. Im not sure if that was a coincedence, but maybe they tested something? Im pretty sure that it was rigged

 

Anyway, those games basicly ended in a loss always. 4 extremely bad players just died because they cant angle and cant shot and dont know what objectives are. You cant kill above average players that fast, as they have atleast some brains to migitate damage. And playing 2vs6... almost impossible, especially on T5 (what it was back then).

Those were one of the most horrible games ive played, also a reason why i dont play ranked sprint anymore after that, only a couple.

 

Yes I can see your point. Do you have any ideas for a better MM system than the current one we have? Or is it the best of a bad lot of ideas and that's it do you think.

[HMS_Kilinowski]

35 minutes ago, Dragnorak said:

There have been a lot of discussions over the years re the quality of ones team and not forgetting of course the dirge of 40% Wr players.

 

So with a bit of thought would a stats based MM actually work do you think so that you are matched up more often with players of your skill level where available?

Such a MM would even out Winrates, so they converge to 50%. The MM would the be very dynamic as it continously tries to balance minor differences in WR.

We would get more exciting interesting battles, I am sure, but there would be no incentive to play good, as you only get more of a handicap, the better you are. The only way to make this work would be a system, where more handicap also means higher rewards and some sort of bragging status.

[Pikkozoikum]

5 minutes ago, Jethro_Grey said:

Somehow? One does not magically balance Things. 

 

Besides, you can't only take PR as criteria to balance teh MM, that wouldn't work. 

You have to take ships and their weighting into account too. 

 

Imagine a division of super unicums with a few average players mixed in. 

In order to balance that you either have to find equally strong players, or if the MM can't find any in their range, give the other team way more players. 

And with the amount of bad and below average players in the playerbase, you'd have 6players vs 20 players games. (exagerated)

 

This would create an Imbalance, so no. PR based, or even skill based MM in WoWS is a phantasy that'll never come true. 

And if it ever would, it'd turn into a nightmare pretty damn fast. 

It gets closer to balance and it will be 100% balanced in case of PRs. PRs are players overall performance.

 

The ship is somthing, tha I wouldn't take to account, but more like the the Tier, if top tier or bottom tier. A unicum player in bottom tier should be less valued.

 

It's not about to get a 100% perfect balancing, just to get at least any kind of balancing and summing up the team-rating would work to create a balancing of teams

[__Helmut_Kohl__]

5 minutes ago, Dragnorak said:

Here we go the first post to diverge from my original question to a veiled personal attack on me.

 

So just to SHUT you up before you get on your high horse:

 

I was NOT blaming any of my losses on MM

 

Got IT?>???

 

I was attempting to start a conversation on the merits of a better MM system than we currently have. Maybe you struggle to read and understand English in which case I forgive you.

 

You still didn't address my point, that in your system everyone will get back to 50% WR, preventing the system from working correctly, and my point that it punishes good play.

 

And your sad rage post just confirms my assumption.

[mtm78]

38 minutes ago, Pikkozoikum said:

That would devide the playerbase pretty much. A server with only 57% and higher would end to a low playerbase-server

 

What I just would wish, if the would a matchmaking with personal-team-rating

 

The the different PRs of different players sums to a total number, which has to fit with the total number of the enemy team. In a case of a 3v3 to make it easier

 

A unicorm has 1800 pr

average as 1500

and bad player with 1200

=  4500 PR total

 

the other team has only average players but no unicums

1550

1550

1400

=4500 PR total

 

The match would be some how balanced without having only unicums in on team and only bad players in the other team

 

 

PR puts to much weight in damage. Yes you can disagree, that's ok. 

 

And no this idea been dismissed before, it would mean good people always have to carry bad people, indirectly bad people still getting carried by good people, so nothing is solved. No more fully good team vs fully good team out of random luck, no because every good players is being pulled in matches where they have to compensate for the rest of their team. Doesn't sound that much fun to me.

 

 

[Dragnorak]

21 minutes ago, __Helmut_Kohl__ said:

 

You still didn't address my point, that in your system everyone will get back to 50% WR, preventing the system from working correctly, and my point that it punishes good play.

 

And your sad rage post just confirms my assumption.

 

It's not sad and I wasn't raging.

 

Anyway to your points. 

 

It's not my system. I wanted to open dialogue. It's what forums are for don't you know. To share ideas and talk about stuff and see if you can find new ideas or consensus. So not my system just an opening idea which has already morphed into discussion and new ideas. 

 

So yes maybe WR on its own as you rightly say is not the best idea but maybe a mix of WR and PR and Tier would work better than what we have at the moment. 

[__Helmut_Kohl__]

2 minutes ago, mtm78 said:

every good players is being pulled in matches where they have to compensate for the rest of their team. Doesn't sound that much fun to me.

 

Yep, that's what I mean.

If you play well, you are being punished by having to play in even worse teams.

[Dragnorak]

1 minute ago, __Helmut_Kohl__ said:

 

Yep, that's what I mean.

If you play well, you are being punished by having to play in even worse teams.

 

But what we have at the moment is totally random and as more and more potatoe players reach higher ranks your chances of getting a potatoe team increase exponentially. 

[mtm78]

8 minutes ago, Dragnorak said:

So yes maybe WR on its own as you rightly say is not the best idea but maybe a mix of WR and PR and Tier would work better than what we have at the moment. 

 

So... you want to reward farming dmg? That's how you get your purple PR, do moar dmg in the ship as X% of other players. Sad to, because it creates this ''we're unicum, we need to do our dmg to keep our unicum status so if we can do A and win the game, or B and farm lots more dmg but increase chances of loosing', still choosing B. Effective dmg matters a lot more, and guess what this effectiveness is already shown in the winrate itself. Therefore PR weighting of dmg vs wr is imo off the mark. Then again, using WR is only applicable over large enough datasets. 

 

4 minutes ago, Dragnorak said:

But what we have at the moment is totally random and as more and more potatoe players reach higher ranks your chances of getting a potatoe team increase exponentially. 

 

Wait, since when is this about ranked?

[__Helmut_Kohl__]

5 minutes ago, Dragnorak said:

But what we have at the moment is totally random and as more and more potatoe players reach higher ranks your chances of getting a potatoe team increase exponentially. 

 

And the solution is to not make it random and instead forcing good players into bad teams?