Jump to content


how come you don't take skill/win rate into account in matchmaking


  • Please log in to reply
1958 replies to this topic

Staz211 #1901 Posted Mar 16 2018 - 17:25

    Major

  • Players
  • 27045 battles
  • 3,906
  • Member since:
    06-11-2012

View PostNixeldon, on Mar 16 2018 - 11:13, said:

I've posted dozens of underdog and balanced steamrolls in these threads.

 

1) Those are extremely rare.

2) It will be ignored because it doesn't fit the narrative.

3) Because the colors matched, at least the other team "had a chance".

4) This example is anecdotal, but the others aren't.

5) But what if MM had swapped X player for Y player.

6) etcetera 

 

And there it is. 

hatman2 #1902 Posted Mar 16 2018 - 20:56

    Sergeant

  • Players
  • 18416 battles
  • 182
  • [1MC] 1MC
  • Member since:
    12-23-2012
Skill based MM seems to work in Star-craft, league of legends, DOTA....... and all the other games that are e-sports. I'm just saying. 

Nixeldon #1903 Posted Mar 16 2018 - 21:02

    Captain

  • Players
  • 59702 battles
  • 1,957
  • [PRTSN] PRTSN
  • Member since:
    10-30-2011

View Posthatman2, on Mar 16 2018 - 14:56, said:

Skill based MM seems to work in Star-craft, league of legends, DOTA....... and all the other games that are e-sports. I'm just saying. 

Your point is what?

 



hatman2 #1904 Posted Mar 16 2018 - 21:04

    Sergeant

  • Players
  • 18416 battles
  • 182
  • [1MC] 1MC
  • Member since:
    12-23-2012

View PostNixeldon, on Mar 16 2018 - 14:02, said:

Your point is what?

 

 

That it could work in this game. 

 

They were/are desperately trying to make this game an e-sport, it's something they should do.


Edited by hatman2, Mar 16 2018 - 21:10.


Nixeldon #1905 Posted Mar 16 2018 - 21:14

    Captain

  • Players
  • 59702 battles
  • 1,957
  • [PRTSN] PRTSN
  • Member since:
    10-30-2011

View Posthatman2, on Mar 16 2018 - 15:04, said:

That it could work in this game. 

It isn't a question of could it work. Anything could work, but how well. It is a question of what it should accomplish followed by how to do it if it is even practical. You ignored some significant differences between WOT and the games you mentioned.

 

 



hatman2 #1906 Posted Mar 16 2018 - 21:41

    Sergeant

  • Players
  • 18416 battles
  • 182
  • [1MC] 1MC
  • Member since:
    12-23-2012

View PostNixeldon, on Mar 16 2018 - 14:14, said:

It isn't a question of could it work. Anything could work, but how well. It is a question of what it should accomplish followed by how to do it if it is even practical. You ignored some significant differences between WOT and the games you mentioned.

 

 

 

I don't disagree with you. But I believe that the games that are successful e-sports, the games that are selling out arenas and drawing huge audiences on twitch all have some kind of skill based MM. It seems like that what WOT is trying to be. 

StiffWind #1907 Posted Mar 16 2018 - 21:50

    Major

  • -Players-
  • 17338 battles
  • 2,074
  • [MOV] MOV
  • Member since:
    03-15-2017

View Post_Tsavo, on Mar 16 2018 - 11:09, said:

 

Because every game that G platoon is on my team, there's a game where they're on the enemy team.

 

Yeah, a single match may feel unfair, but it was assembled without malice or ill intent.

 

A compressed win rate is no bueno, and not fair to the playerbase.

 

Malice or ill intent?  Neither of these have anything to do with anything being discussed.  We are discussing the fact of the matter....not anyone's intentions.  We all know it's about making money.

...and what we have know is not a normal win rate...it's an expanded win rate.  Made so by the advantages and disadvantages WoT's MM hands out.  Changing things would not compress it...it would normalize the win rate into something more closely aligned with a player's skill.


 



Kliphie #1908 Posted Mar 16 2018 - 21:53

    Major

  • Players
  • 32045 battles
  • 4,313
  • [IOC] IOC
  • Member since:
    07-20-2012

View PostStiffWind, on Mar 16 2018 - 15:50, said:

 

Malice or ill intent?  Neither of these have anything to do with anything being discussed.  We are discussing the fact of the matter....not anyone's intentions.  We all know it's about making money.

...and what we have know is not a normal win rate...it's an expanded win rate.  Made so by the advantages and disadvantages WoT's MM hands out.  Changing things would not compress it...it would normalize the win rate into something more closely aligned with a player's skill.


 

 

There's a cogent thought in there somewhere.  

Nixeldon #1909 Posted Mar 16 2018 - 21:57

    Captain

  • Players
  • 59702 battles
  • 1,957
  • [PRTSN] PRTSN
  • Member since:
    10-30-2011

View Posthatman2, on Mar 16 2018 - 15:41, said:

I don't disagree with you. But I believe that the games that are successful e-sports, the games that are selling out arenas and drawing huge audiences on twitch all have some kind of skill based MM. It seems like that what WOT is trying to be. 

There should be a logical progression to e-sports instead of a differing formats. I doubt it matters on NA. Tournaments and Esports are canceled on the NA server.

 

View PostStiffWind, on Mar 16 2018 - 15:50, said:

Malice or ill intent?  Neither of these have anything to do with anything being discussed.  We are discussing the fact of the matter....not anyone's intentions.  We all know it's about making money.

...and what we have know is not a normal win rate...it's an expanded win rate.  Made so by the advantages and disadvantages WoT's MM hands out.  Changing things would not compress it...it would normalize the win rate into something more closely aligned with a player's skill.

WOT has a normal distribution.

 

What should the win rate distribution be, Stiffy?

 

View PostKliphie, on Mar 16 2018 - 15:53, said:

There's a cogent thought in there somewhere.  

Nope. There never has been and probably never will be.



da_Rock002 #1910 Posted Mar 16 2018 - 22:04

    Major

  • -Players-
  • 8805 battles
  • 3,394
  • Member since:
    11-24-2016

 kk


 

 

 

the forum works about like the game
 


Edited by da_Rock002, Mar 16 2018 - 22:17.


da_Rock002 #1911 Posted Mar 16 2018 - 22:07

    Major

  • -Players-
  • 8805 battles
  • 3,394
  • Member since:
    11-24-2016

rem

 

 removed

 


Edited by da_Rock002, Mar 16 2018 - 22:13.


da_Rock002 #1912 Posted Mar 16 2018 - 22:10

    Major

  • -Players-
  • 8805 battles
  • 3,394
  • Member since:
    11-24-2016

View PostNixeldon, on Mar 16 2018 - 15:14, said:

It isn't a question of could it work. Anything could work, but how well. It is a question of what it should accomplish followed by how to do it if it is even practical. You ignored some significant differences between WOT and the games you mentioned.

 

 

 

 

It does work in this game.    60% of the battles are NOT skill skewed.  They are not because MM produces a bell curve with 100% of the matches.   One end/side of the curve, about 20% of the battles, has 'team 1' ridiculously more skillful than 'team 2'.    On the other end/side of the curve, also 20%, the MM has 'team 2' ridiculously more skillful than 'team 1'.   


 

That produces two groups.   The group with 40% of the ridiculously skilled teams is one group.   Because of the overpowering effect the skill skewing produces, both sides of the bell curve produce the same result,  the OP side stomps the crap out of the weak side.   When a extremely strong team meets a weak one, it's natural for them to rolfstomp. 

   The other group of battles, the remaining 60%, is a group of filtered matchups.   The screwed skill matchups are not a part of the 60%.    What you have is a roughly created SBMM result.


 

We play SBMM matches all day long, about 60% of the time.    They prove to be much better battles.   They also include some blowouts, but these truly are 'naturally occurring'.     SBMM works pretty damn good.   And it's already in place.    Get rid of the 40% and their crummy impact on the quality of the WoT product, and 100% of the games would be as good as those 60% WG gives us already. 


 

We don't need perfect MM.   60% of the MM is pretty damn good.   It has excellent equipment matching, and battle tier balancing that works.   To turn the 40% into as good matchups as the present 60% is actually simple.    Simply swap a few pairs of players to move skilled players to the other team, and a few of the newbies to the power side.


 

SBMM already is working.



Kliphie #1913 Posted Mar 16 2018 - 22:11

    Major

  • Players
  • 32045 battles
  • 4,313
  • [IOC] IOC
  • Member since:
    07-20-2012
nvm

Edited by Kliphie, Mar 16 2018 - 22:11.


Hurk #1914 Posted Mar 16 2018 - 22:26

    Major

  • Players
  • 55234 battles
  • 17,374
  • [KGR] KGR
  • Member since:
    09-30-2012

View Posthatman2, on Mar 16 2018 - 12:56, said:

Skill based MM seems to work in Star-craft, league of legends, DOTA....... and all the other games that are e-sports. I'm just saying. 

View Posthatman2, on Mar 16 2018 - 13:04, said:

That it could work in this game. 

They were/are desperately trying to make this game an e-sport, it's something they should do.

and that is where a lot of people disagree. 

I never played starcraft or LOL because i did not like MOBAs for the most part, i did play DOTA with friends. that game was massively punishing with many real time skill based aspects that forced you to think on your feet or fail. I played with a few people much better than i was and focused on learning a couple heroes so that i always had a pick available, then learned how to play those heroes properly in team fights. the result was i was an above average "noob". IE i did not play ranked, but did better than average in randoms. 

I actually enjoy HoTS because its DOTA light... many less economies to juggle... oh, and i also play PvE only these days, but before i did, i was again, able to sustain more than a 50% win rate in "random" matching. 

 

when i started playing WoT, one of the appeals is that it wasnt yet another MOBA. the random matching was harsh, but at the same time fair over time. 

I recently played in ranked.... ranked worked by the way, by the time i got to tier 9, the teams were mostly green + players as the bad players were failing often enough to lose rank faster. but even in ranked, playing mostly arty that has limited impact to the battles, i was again able to rank up to 9 playing only a few battles a night. 

 

my point is, im a skilled enough player to do well regardless of the system you put me in. changing the rules so to speak, doesnt suddenly make me a bad player. but the amount of work those games take removes a lot of the fun factor. in short, with SBMM full time, i would simply play less. because, and since, im a good player, i would be forced to carry more. thats not fun. 

 

frankly, the whole e-sports thing is exactly what the game DOES NOT need. what it really needs is a clear separation between professional players and normal players like other games have with ranked play. your average pubbie should not be in a battle with a 3 man G platoon. they should be in a ranked ladder, working on things that matter to them and their clam instead. I would even go so far as to prevent tounament/professional players from participating in unranked play. 

But... i dont think the NA market has a large enough playerbase to support the separation. so i would rather we keep what we have now. 

 

or they could always add a PvE mode... and i will happily never PvP tanks again. 



spud_tuber #1915 Posted Mar 16 2018 - 22:55

    Major

  • Players
  • 56788 battles
  • 7,707
  • Member since:
    08-26-2013

View Posthatman2, on Mar 16 2018 - 13:56, said:

Skill based MM seems to work in Star-craft, league of legends, DOTA....... and all the other games that are e-sports. I'm just saying. 

I thought they all used a bracket/ladder system, not a balanced method?



NeatoMan #1916 Posted Mar 16 2018 - 23:15

    Major

  • Players
  • 27740 battles
  • 19,668
  • Member since:
    06-28-2011

View Postda_Rock002, on Mar 16 2018 - 16:10, said:

One end/side of the curve, about 20% of the battles, has 'team 1' ridiculously more skillful than 'team 2'.    On the other end/side of the curve, also 20%, the MM has 'team 2' ridiculously more skillful than 'team 1'.  

Your use of "ridiculously more skill" is debatable.   If the matches are indistinguishable by their game play then they aren't ridiculously more skilled.   That 40% number you all keep bantering about comprises many games that are very competitive.  You just don't like who is winning.



Hurk #1917 Posted Mar 16 2018 - 23:52

    Major

  • Players
  • 55234 battles
  • 17,374
  • [KGR] KGR
  • Member since:
    09-30-2012

View PostSoTrue, on Mar 16 2018 - 01:16, said:

The funny/sad part is YOU ARE PLAYING SBMM 60% of the time RIGHT NOW.  Shocking you don't seem to mind.

erm... no. im not. there is nothing skills based about it. 

 

this is the part where you are completely disconnected from the reality of the game.

the game is NOT 1v1. no one, barring some serious cheese situation, is ever going to have a 100% win or loss rate over any significant number of battles. 

this isnt due to "rigging" or "MM deciding". its due to the fact that its FIFTEEN PLAYERS VS FIFTEEN PLAYERS. 

 

the HUMAN RANDOM factor in EVERY MATCH prevents you from EVER achieving more wins than happens now. sure, maps matter a little. sure tank matters some as well, but ultimately, since the enemy team is all HUMAN, there is nothing to predict and learn that will allow anyone to ever say "i am in complete control of the outcome of this match". 

 

that isnt rigging. its random. you keep think that SBMM will somehow make that better... it wont. it will make it far worse using the metrics you are proposing here... instead of having zero input in 60% of all matches, you will have zero input in ~80% of all matches. 

thats what i keep trying to tell you, you arent listening. ive played SBMM games... ive watched enemy teams' handicap carry me into territory with lopsided game play before. i dont want that in tanks too. its a ton of work and not any more fun. 



_Tsavo_ #1918 Posted Mar 17 2018 - 04:15

    Major

  • Beta Testers
  • 44885 battles
  • 19,090
  • [TSAVO] TSAVO
  • Member since:
    02-16-2011

View Posthatman2, on Mar 16 2018 - 14:56, said:

Skill based MM seems to work in Star-craft, league of legends, DOTA....... and all the other games that are e-sports. I'm just saying. 

 

Those aren't 15 v 15.

RenamedUser_1028600066 #1919 Posted Mar 17 2018 - 04:17

    Corporal

  • -Players-
  • 6291 battles
  • 67
  • Member since:
    11-26-2017
Because it would hurt too many feelings. 

SoTrue #1920 Posted Mar 17 2018 - 21:07

    Major

  • Players
  • 31802 battles
  • 3,302
  • Member since:
    04-01-2011

View PostHurk, on Mar 16 2018 - 14:52, said:

erm... no. im not. there is nothing skills based about it.

 

this is the part where you are completely disconnected from the reality of the game.

the game is NOT 1v1. no one, barring some serious cheese situation, is ever going to have a 100% win or loss rate over any significant number of battles.

this isnt due to "rigging" or "MM deciding". its due to the fact that its FIFTEEN PLAYERS VS FIFTEEN PLAYERS.

 

the HUMAN RANDOM factor in EVERY MATCH prevents you from EVER achieving more wins than happens now. sure, maps matter a little. sure tank matters some as well, but ultimately, since the enemy team is all HUMAN, there is nothing to predict and learn that will allow anyone to ever say "i am in complete control of the outcome of this match".

 

that isnt rigging. its random. you keep think that SBMM will somehow make that better... it wont. it will make it far worse using the metrics you are proposing here... instead of having zero input in 60% of all matches, you will have zero input in ~80% of all matches.

thats what i keep trying to tell you, you arent listening. ive played SBMM games... ive watched enemy teams' handicap carry me into territory with lopsided game play before. i dont want that in tanks too. its a ton of work and not any more fun.

 

really?  how daft are you.  It's been proven over and over by the data everyone has collected that about 60% of the existing games you player occur between roughly equally skilled teams - I.E. SKILL BASED MM.   How can you not know this?  It's the bottom 20% and top 20% crap battles we are trying to get rid of.




1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users