Jump to content


Finally a proper MM...enjoy..:)


  • Please log in to reply
206 replies to this topic

Alabamatick #61 Posted 21 March 2015 - 04:49 PM

    Lieutenant Сolonel

  • Player
  • 31849 battles
  • 3,336
  • Member since:
    04-11-2011

View Postthoso1973, on 21 March 2015 - 09:34 AM, said:

It's not just that AW's matchmaker algorithm is being designed to account for tank modules and player stats. It will also prevent platoons driving tanks several tiers apart from joining a game, hard cap max artillery at 3 per team and reduce the tier spread in any game to +/-1 if enough players are online.

 

^ MM mechanisms that The potatoes have wished to see implemented in WoT since they started playing the game.

Fixed that for you:harp:

View PostYuri_Yslin, on 21 March 2015 - 11:12 AM, said:

if you distrubte the players evenly (the "3-3" example), you end up with forced 50% WR over a large number of battles. There is absolutely no way around this, so if they go for it, they should also remove WR from display, because it's meaningless: the game enforces 50% WR, so what's the point of even displaying it?

 

Bot is going to have 50% WR, and a super unicum would have 50% WR also. What a great system :P

This ^, although i have to disagree slightly as super unicums could probably manage upto 51%, average players would have 50% and even the meat and potatoes would manage 49% :trollface:

 

It would also lead to more draws, how interesting, yawn :sceptic:

 

Edit: added troll face


Edited by Alabamatick, 21 March 2015 - 05:08 PM.


Gl0cK_17 #62 Posted 21 March 2015 - 04:52 PM

    Sergeant

  • Player
  • 34794 battles
  • 204
  • Member since:
    02-02-2011

View Postthoso1973, on 21 March 2015 - 03:33 PM, said:

It's kind of funny how some players arguing against skillbased mm, have become so institutionalized by playing under the terms of the current broken mm, that they perceive it as a 'punishment' if they were ever matched against equally skilled players at a consistent rate. :) 

 

You know whats really funny, that people bring skill into the argument as beeing one of the MMs parameters.

While AWs MM will  only look at vehicle, vehicle upgrades and ammount of battles played in a vehicle. Theres is nothing skill based in there, only the assumption that someone with more games played should be better skilled then someelse. And we all know how true that is right? ;)



thoso1973 #63 Posted 21 March 2015 - 05:00 PM

    Staff Sergeant

  • Player
  • 46187 battles
  • 440
  • [TEC] TEC
  • Member since:
    03-07-2013

View Postlord_chipmonk, on 21 March 2015 - 03:48 PM, said:

 

You really are, or at least that is how you come across. Your point doesn't change the thrust of Yuri's argument, which remains. Your comment, while correct, was pedantic and facile. Yuri remains fundamentally correct. 

 

Yuri corrected LegioCenturion's statement in his post, arguing that 'math' backs up his point.

 

Never the less, what Legio said and is saying is entirely accurate. A battle between two equally skilled players/teams have a higher probability to lead to an unpredictable result than a battle between players/teams on different skill levels on a per-game basis. That's the whole point or at least half of it. Better and more enjoyable gameplay is the other half among most proponents of skillbased matchmaking.



Homer_J #64 Posted 21 March 2015 - 05:06 PM

    Field Marshal

  • Beta Tester
  • 31518 battles
  • 34,514
  • [WJDE] WJDE
  • Member since:
    09-03-2010

View PostAlabamatick, on 21 March 2015 - 03:49 PM, said:

 

This ^, although i have to disagree slightly as super unicums could probably manage upto 51%,

 

Surely super unicums will suffer worst because by definition there will only very rarely be two super unicums in the queue together so the only way the system can balance them is put several very good players on the opposite team.

Alabamatick #65 Posted 21 March 2015 - 05:07 PM

    Lieutenant Сolonel

  • Player
  • 31849 battles
  • 3,336
  • Member since:
    04-11-2011

View PostHomer_J, on 21 March 2015 - 04:06 PM, said:

 

Surely super unicums will suffer worst because by definition there will only very rarely be two super unicums in the queue together so the only way the system can balance them is put several very good players on the opposite team.

Actually i forgot to add troll face



lord_chipmonk #66 Posted 21 March 2015 - 05:12 PM

    Field Marshal

  • Player
  • 35729 battles
  • 10,279
  • [-HOW-] -HOW-
  • Member since:
    12-23-2012

View Postthoso1973, on 21 March 2015 - 05:00 PM, said:

 

Yuri corrected LegioCenturion's statement in his post, arguing that 'math' backs up his point.

 

Never the less, what Legio said and is saying is entirely accurate. A battle between two equally skilled players/teams have a higher probability to lead to an unpredictable result than a battle between players/teams on different skill levels on a per-game basis. That's the whole point or at least half of it. Better and more enjoyable gameplay is the other half among most proponents of skillbased matchmaking.

 

When you say  "higher probability to lead to an unpredictable result" I assume you mean that you (the observer) have the lowest probability (50%, ignoring the probability of a draw for the sake of ease of argument) of predicting the result. This is true, and something Yuri acknowledges when he discusses the difference between win change and win rate. As I say, this might be true, but does not undermine of disprove the point that Yuri is making. There is a difference between predicting the result on a game by game basis (where there are many variables capable of changing the result) and on a large statistical sample (where most of the variables average out). 



PanzerKFeldherrnhalle #67 Posted 21 March 2015 - 05:34 PM

    Warrant Officer

  • Player
  • 2873 battles
  • 816
  • Member since:
    09-14-2013

View PostYuri_Yslin, on 21 March 2015 - 03:42 PM, said:

 

You don't seem to understand that over big batches (10000+ or so attempts) there is little to none "unpredictability".

 

Yes, noobs may win vs. pros in a single match.

But let those teams play 1000 times against each other and check results.

 

Winchance determines 1 game. It can be pretty invalid.

 

Your answers are way off track suggesting lack of comprehension. Simply put and I cannot make it any simpler so please do make an effort:

50% win-chance isn't determinant of anything. It is in fact the only way you cannot predict the outcome of a game. Likewise, the outcome of 10000 games.

 

Your statement that balanced match-ups are the equivalent of 50% rates is unintelligent.

 

The reason you make it is that you've gotten used to the World of Tanks' manipulative matchmaking, system that consistently generates uneven match-ups (ironically for you) with the purpose of balancing win-rates.

 



_Grim_ #68 Posted 21 March 2015 - 05:39 PM

    Second Lieutenant

  • Player
  • 22990 battles
  • 1,431
  • [-DFA-] -DFA-
  • Member since:
    04-13-2011

View PostLegioCenturion, on 21 March 2015 - 06:34 PM, said:

Your statement that balanced match-ups are the equivalent of 50% rates is unintelligent.

 

So I guess 3000 years of maths are also "unintelligent" :facepalm:....

pallie_the_artillerist #69 Posted 21 March 2015 - 05:49 PM

    Second Lieutenant

  • Player
  • 19949 battles
  • 1,364
  • [TRTD] TRTD
  • Member since:
    01-12-2013

View PostLegioCenturion, on 21 March 2015 - 05:34 PM, said:

he World of Tanks' manipulative matchmaking, system that consistently generates uneven match-ups (ironically for you) with the purpose of balancing win-rates.

 

 

You can't say Yuri is stupid and wrong and t hen post this senseless drivel without making yourself look stupid :/

thoso1973 #70 Posted 21 March 2015 - 05:53 PM

    Staff Sergeant

  • Player
  • 46187 battles
  • 440
  • [TEC] TEC
  • Member since:
    03-07-2013

View Postlord_chipmonk, on 21 March 2015 - 04:12 PM, said:

 

When you say  "higher probability to lead to an unpredictable result" I assume you mean that you (the observer) have the lowest probability (50%, ignoring the probability of a draw for the sake of ease of argument) of predicting the result. This is true, and something Yuri acknowledges when he discusses the difference between win change and win rate. As I say, this might be true, but does not undermine of disprove the point that Yuri is making. There is a difference between predicting the result on a game by game basis (where there are many variables capable of changing the result) and on a large statistical sample (where most of the variables average out). 

 

"Higher probability to lead to an unpredictable result" simply means that (if we ignore tied results as you suggest) the two possible outcomes of a battle have an equal chance of materializing for your team. It's basically 50/50 - win/loss. That is the highest possible degree of an unpredictable result. It is more likely to see yourself in a 50/50 battle if the teams are evenly skilled than the opposite. Hence, skilled matchmaking leads to a higher probability of a 50/50 game aka an unpredictable result.

 

A battle between two teams far apart judging by player skills/stats (and tank tiers being comparable between the teams), will be closer to, say as an example, a 70/30 game. Hence the probable outcome is less unpredictable. If you guess that the better team will win, you have a 70% chance of being right, as opposed to 50% on either team in the example above.

 

You are welcome to browse through the posts I've made on this issue at the forum. You won't find a single one where I argue that skillbased mm will somehow increase my WR or WN8 stats or whatever. I don't really care about that. What I care about is the quality of the gameplay. As I see it, challenging gameplay is more desirable than the opposite.

 

 

@Alabamatick

I don't really know why you suppose that only tomatoes wish for fewer arty per game/team, no more tier X players dragging a couple of tier II players into a battle via platooning just to troll or a lesser tier spread when the teams are composed by the mm. As far as I know, those things are very high on most WoT players wishlist, many of them blue players.

Link to post #61">#61

 


Alabamatick #71 Posted 21 March 2015 - 06:05 PM

    Lieutenant Сolonel

  • Player
  • 31849 battles
  • 3,336
  • Member since:
    04-11-2011

View Postthoso1973, on 21 March 2015 - 04:53 PM, said:

 

@Alabamatick

I don't really know why you suppose that only tomatoes wish for fewer arty per game/team, no more tier X players dragging a couple of tier II players into a battle via platooning just to troll or a lesser tier spread when the teams are composed by the mm. As far as I know, those things are very high on most WoT players wishlist, many of them blue players.

Link to post #61">#61

 

I was actually just generalising on the wishes of a fairer :sceptic: mm, not the specifics of the whole proposition

 

the only thing i agree with is, no troll platoons and fewer spg's, (however fewer spg's per game will probably lead to  fewer games with no spg's)

 

and as for many players wishlist, how many is many


Edited by Alabamatick, 21 March 2015 - 06:10 PM.


lord_chipmonk #72 Posted 21 March 2015 - 06:10 PM

    Field Marshal

  • Player
  • 35729 battles
  • 10,279
  • [-HOW-] -HOW-
  • Member since:
    12-23-2012

View Postthoso1973, on 21 March 2015 - 05:53 PM, said:

 

"Higher probability to lead to an unpredictable result" simply means that (if we ignore tied results as you suggest) the two possible outcomes of a battle have an equal chance of materializing for your team. It's basically 50/50 - win/loss. That is the highest possible degree of an unpredictable result. It is more likely to see yourself in a 50/50 battle if the teams are evenly skilled than the opposite. Hence, skilled matchmaking leads to a higher probability of a 50/50 game aka an unpredictable result.

 

A battle between two teams far apart judging by player skills/stats (and tank tiers being comparable between the teams), will be closer to, say as an example, a 70/30 game. Hence the probable outcome is less unpredictable. If you guess that the better team will win, you have a 70% chance of being right, as opposed to 50% on either team in the example above.

 

You are welcome to browse through the posts I've made on this issue at the forum. You won't find a single one where I argue that skillbased mm will somehow increase my WR or WN8 stats or whatever. I don't really care about that. What I care about is the quality of the gameplay. As I see it, challenging gameplay is more desirable than the opposite.

Yes, that's what I thought you meant. I understand what you're saying. 



thoso1973 #73 Posted 21 March 2015 - 06:22 PM

    Staff Sergeant

  • Player
  • 46187 battles
  • 440
  • [TEC] TEC
  • Member since:
    03-07-2013

View PostAlabamatick, on 21 March 2015 - 05:05 PM, said:

I was actually just generalising on the wishes of a fairer :sceptic: mm, not the specifics of the whole proposition

 

the only thing i agree with is, no troll platoons and fewer spg's, (however fewer spg's per game will probably lead to  fewer games with no spg's)

 

and as for many players wishlist, how many is many

 

Fair enough.

 

As for 'how many is many' - on the first page only of this very General Discussion sub-forum, there are currently 3 arty specific threads being kept alive by a wide variety of forum members, 2 of those specifically addressing the number of artys in each game and why that number should be reduced.

 

If WG really wanted to implement all the three mm criteria I mentioned, they could. With relative ease, I might add. They have all the parameters necessary to make it work available to them now.



Vestrick64 #74 Posted 21 March 2015 - 06:27 PM

    Lieutenant General

  • Player
  • 11492 battles
  • 6,210
  • Member since:
    03-02-2013

View Postthoso1973, on 21 March 2015 - 05:00 PM, said:

 

Yuri corrected LegioCenturion's statement in his post, arguing that 'math' backs up his point.

 

Never the less, what Legio said and is saying is entirely accurate. A battle between two equally skilled players/teams have a higher probability to lead to an unpredictable result than a battle between players/teams on different skill levels on a per-game basis. That's the whole point or at least half of it. Better and more enjoyable gameplay is the other half among most proponents of skillbased matchmaking.

 

Unpredictable result - yes.

But this unpredictable result will net you 50% (or 49% if you want to be accurate) WR nevertheless over a large batch of games.

 

Flipping a coin is extremely unpredictable. But do it 10000 times and you'll see that you're heading towards 50% distrubution of both sides.


Same with WR.

 

I'm not sure if this won't actually net an even short


Edited by Yuri_Yslin, 21 March 2015 - 06:31 PM.


iBeNoob #75 Posted 21 March 2015 - 06:29 PM

    Staff Sergeant

  • Player
  • 2517 battles
  • 377
  • Member since:
    03-30-2014

I'm still trying to find the skill in that MM formula.

Anyone know where it is?


Edited by iBeNoob, 21 March 2015 - 06:30 PM.


SovietBias #76 Posted 21 March 2015 - 06:31 PM

    Second Lieutenant

  • Player
  • 37407 battles
  • 1,347
  • Member since:
    06-10-2013

View PostiBeNoob, on 21 March 2015 - 06:29 PM, said:

I'm still trying to find the skill in that MM formula.

Anyone knows where it is?

 

Obviously,  it's the 3 spg hard cap :trollface:

PanzerKFeldherrnhalle #77 Posted 21 March 2015 - 06:31 PM

    Warrant Officer

  • Player
  • 2873 battles
  • 816
  • Member since:
    09-14-2013

View Postthoso1973, on 21 March 2015 - 04:53 PM, said:

 

"Higher probability to lead to an unpredictable result" simply means that (if we ignore tied results as you suggest) the two possible outcomes of a battle have an equal chance of materializing for your team. It's basically 50/50 - win/loss. That is the highest possible degree of an unpredictable result. It is more likely to see yourself in a 50/50 battle if the teams are evenly skilled than the opposite. Hence, skilled matchmaking leads to a higher probability of a 50/50 game aka an unpredictable result.

 

A battle between two teams far apart judging by player skills/stats (and tank tiers being comparable between the teams), will be closer to, say as an example, a 70/30 game. Hence the probable outcome is less unpredictable. If you guess that the better team will win, you have a 70% chance of being right, as opposed to 50% on either team in the example above.

 

You are welcome to browse through the posts I've made on this issue at the forum. You won't find a single one where I argue that skillbased mm will somehow increase my WR or WN8 stats or whatever. I don't really care about that. What I care about is the quality of the gameplay. As I see it, challenging gameplay is more desirable than the opposite.

 

 

@Alabamatick

I don't really know why you suppose that only tomatoes wish for fewer arty per game/team, no more tier X players dragging a couple of tier II players into a battle via platooning just to troll or a lesser tier spread when the teams are composed by the mm. As far as I know, those things are very high on most WoT players wishlist, many of them blue players.

Link to post #61">#61

 

 

Thoso1973,

 

There's at least 3 or 4 very well known streamers to whom the game has no secrets and their opinion on topics such as artillery are unanimous. Or even players like Garbad that sometimes make appearance on the forum. Of course it's not only their opinion that matters. Nor is only bad players opinions that matter. However, the notion that only bad players wish for X, Y or Z is silly and only finds echo in forums such as this.

 

There are some major issues with the game (an excellent game no doubt about it) and the matchmaking or artillery are undoubtedly part of them.



Vestrick64 #78 Posted 21 March 2015 - 06:39 PM

    Lieutenant General

  • Player
  • 11492 battles
  • 6,210
  • Member since:
    03-02-2013

I'm afraid your "skill MM" will only lead to a greater amount of 3min games.

 

When good players clash and someone screws up, the rest of the team takes advantage of that quickly and snowballs the game fast.


It's not uncommon to have games in WoT where worst players are 1200+ WN8 players, esp. on tier X in WoT and they are nowhere near the mythical gaming valhalla people dream about.



Alabamatick #79 Posted 21 March 2015 - 06:41 PM

    Lieutenant Сolonel

  • Player
  • 31849 battles
  • 3,336
  • Member since:
    04-11-2011

View Postthoso1973, on 21 March 2015 - 05:22 PM, said:

 

Fair enough.

 

As for 'how many is many' - on the first page only of this very General Discussion sub-forum, there are currently 3 arty specific threads being kept alive by a wide variety of forum members, 2 of those specifically addressing the number of artys in each game and why that number should be reduced.

 

If WG really wanted to implement all the three mm criteria I mentioned, they could. With relative ease, I might add. They have all the parameters necessary to make it work available to them now.

Did you read all my post, i did say i didn't disagree with arty reduction, i just said that it would lead to fewer games with 0 spg's

 

atm we do get a few with 0, which is good, but the how many is may comment was about a fairer mm

 

(bearing in mind that most players don't come to the forums and the many players who have no or very little complant with the mm, therefore don't come here to complain)



lord_chipmonk #80 Posted 21 March 2015 - 06:43 PM

    Field Marshal

  • Player
  • 35729 battles
  • 10,279
  • [-HOW-] -HOW-
  • Member since:
    12-23-2012
Ultimately, AW have their mm system, I shall try it when able, WoT has its mm system (which is I believe being reviewed by WG anyway), which I personally play and enjoy. Many of us will not agree on a `perfect' mm. If both WoT and AW go in slightly different directions, we can have both. Sounds good to me...




1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users