Jump to content


Math and Matchmaking

Match-Making Matchmaking MM Balance rework Tier Random-Battles Random Random battles Un-even

  • Please log in to reply
103 replies to this topic

CardinalMite #41 Posted 12 September 2020 - 06:09 AM

    Captain

  • Member
  • 32109 battles
  • 1,389
  • [PBKAC] PBKAC
  • Member since:
    04-18-2014

View PostAlexTheKid72, on 11 September 2020 - 09:46 PM, said:

Why would queues be longer if the matchmaker first pick 30 tanks that are class-equal, then apply a bit of player stats to decide which heavies go on which team et al?
It's not going to be 45%ers vs 45%ers, but the skill levels would resemble TEAM balance.
Unicums and potatoes are still on the same team, but the mm does team pick-for-pick from best to worst stats. AFTER mm has put 30 people into a battle. Not choosing the 30 based on skill, just balancing total team skill.
If theres 4x60% high stat players, they don't end up on the same team because tank type is the only parameter. Each team gets 2. No extra queue time, just an extra quick math calculation before battle starts.

Exactly which stat do you use for this?

 

Overall win rate? Overall Win rate in the class of tank they are playing? Overall win rate in the tier they are queuing in? Overall win rate in the specific tank they are playing?

 

But you have 2 52% overall players in the match. One has 30k games, plays mainly tiers 7,8,9,10 is playing a tier 9 LT that they have a 60% win rate in after 200 games in it. The other has 30k games, has a 48% highest win rate in LTs at tier 9 playing a tier 9 LT, 5k of the games they have are in a tier 5 T67 with a 54% win rate in it. Most of the rest of their tanks played are < tier 7.

 

People put out whines that if only MM had SBMM everything would be more balanced, more fair that it is an easy fix. I don't think that is true at all. I don't think people really think through what it looks like in practise.


“Holiday ops is balanced for Asia because a good player opening 11 boxes will get just as many decorations as a noob opening 75..”—.Murazor new head of global festive events.

 

"It's like you literally dont read your own posts, let alone others, make up instead what you wished others said, then put a non-sensical counter argument to that imagined post."

 

"Never argue with an idiot, they drag you down to their level then beat you with experience. But then if I limited myself to only arguing with non-idiots on these forums I'd literally never post"


AlexTheKid72 #42 Posted 12 September 2020 - 06:45 AM

    Sergeant

  • Member
  • 28850 battles
  • 199
  • [SABRE] SABRE
  • Member since:
    05-28-2017

View PostCardinalMite, on 12 September 2020 - 06:09 AM, said:

Exactly which stat do you use for this?

 

Overall win rate? Overall Win rate in the class of tank they are playing? Overall win rate in the tier they are queuing in? Overall win rate in the specific tank they are playing?

 

But you have 2 52% overall players in the match. One has 30k games, plays mainly tiers 7,8,9,10 is playing a tier 9 LT that they have a 60% win rate in after 200 games in it. The other has 30k games, has a 48% highest win rate in LTs at tier 9 playing a tier 9 LT, 5k of the games they have are in a tier 5 T67 with a 54% win rate in it. Most of the rest of their tanks played are < tier 7.

 

People put out whines that if only MM had SBMM everything would be more balanced, more fair that it is an easy fix. I don't think that is true at all. I don't think people really think through what it looks like in practise.

I just remember picking teams as a kid. Seemed to work well.

Not sure what stat you should use, but recents in the tank you're playing would probably work? WarGaming have all the metrics, I'm sure there's one that only half of us would hate.
I wouldn't call it sbmm .... more skill based team distribution. 

edit : I don't actually mind current mm that much, but there's always room for improvement


Edited by AlexTheKid72, 12 September 2020 - 06:52 AM.

signature.png

Ezz #43 Posted 12 September 2020 - 07:27 AM

    How many flipping posts do I need to get past Major?

  • Beta-Tester
  • 74444 battles
  • 36,903
  • [PBKAC] PBKAC
  • Member since:
    07-17-2012

View PostThe_Big_M, on 12 September 2020 - 01:43 AM, said:

 

The reference to 'who cares' was about the stats, not about the game. Not sure how you took it another way.

 

As for mm, it's a fundamental part of the game, so obviously intrinsic to the enjoyment.

How much you win, ie what your stats are, is a fundamental part of the game, so obviously intrinsic to the enjoyment.


Edited by Ezz, 12 September 2020 - 07:31 AM.

Who the [edited] are you? Get Spoofed! "wouldn't be a proper WG balance change if they didn't [edited] something up after all "

>9000 cynicism brought to you by P2W, Balance TM and the Cartoon Connection

R. Pubbie: "why are all PBKAC players so rude, arrogant and nasty? and why do Mods favor them?"


CardinalMite #44 Posted 12 September 2020 - 07:43 AM

    Captain

  • Member
  • 32109 battles
  • 1,389
  • [PBKAC] PBKAC
  • Member since:
    04-18-2014

View PostAlexTheKid72, on 11 September 2020 - 10:45 PM, said:

I just remember picking teams as a kid. Seemed to work well.

Not sure what stat you should use, but recents in the tank you're playing would probably work? WarGaming have all the metrics, I'm sure there's one that only half of us would hate.
I wouldn't call it sbmm .... more skill based team distribution. 

edit : I don't actually mind current mm that much, but there's always room for improvement

So you played the tank 1 game, it was a landslide win. Your recents in that tank are 100% win rate. You play the tank 9 more times, you are blessed by winning 6 of those, you didn't do all that much or contribute all that much to the wins, but hey everyone gets win and loss streaks right? You are a 70% win rate unicum in the tank after 10 battles. Say your overall account win rate is 48%, you play another 90 battles in the tank, only because you are a 48% player on average, and the tank is higher tiers where you average more like 44%, as the number of games increases towards that hundredth game your win rate for the tank drops to 45%. MM would have matched you completely wrongly for most of those matches. As it would have done for most of the other players on the teams, most of the time. 

 

Picking teams as a kid worked well because human beings don't measure the capability of their fellows by statistics, they don't need to because they generally know all the other kids being picked and how good a player they are without particular measures. MM doesn't know one player ID from another by any means other than statistics.

 

Point is the upshot is that your skill matching per match is meaningless, the stats only work over statistical samples, even there you get anomalies. How do you compare someone who played a hundred battles in a tank that they ground from completely stock, versus someone that played 100 battles but free XP'd all the modules? The kind of matching you talk about is only worth doing if it is accurate, if it isn't then it literally would be flat out worse even by the goals you set for MM in introducing that type of matching.


“Holiday ops is balanced for Asia because a good player opening 11 boxes will get just as many decorations as a noob opening 75..”—.Murazor new head of global festive events.

 

"It's like you literally dont read your own posts, let alone others, make up instead what you wished others said, then put a non-sensical counter argument to that imagined post."

 

"Never argue with an idiot, they drag you down to their level then beat you with experience. But then if I limited myself to only arguing with non-idiots on these forums I'd literally never post"


Ezz #45 Posted 12 September 2020 - 07:46 AM

    How many flipping posts do I need to get past Major?

  • Beta-Tester
  • 74444 battles
  • 36,903
  • [PBKAC] PBKAC
  • Member since:
    07-17-2012
If everyone was in the same tanks and always had been, then sbmm might be able to make teams relatively even.

Who the [edited] are you? Get Spoofed! "wouldn't be a proper WG balance change if they didn't [edited] something up after all "

>9000 cynicism brought to you by P2W, Balance TM and the Cartoon Connection

R. Pubbie: "why are all PBKAC players so rude, arrogant and nasty? and why do Mods favor them?"


Ezz #46 Posted 12 September 2020 - 08:45 AM

    How many flipping posts do I need to get past Major?

  • Beta-Tester
  • 74444 battles
  • 36,903
  • [PBKAC] PBKAC
  • Member since:
    07-17-2012

View PostThe_Big_M, on 12 September 2020 - 02:04 AM, said:

 

WG's 3:5:7 is still in place. So you see this SBMM as equivalent, which means there's nothing fundamentally wrong with it, just that you don't like it. Which is fair enough, but not an actual justification that it doesn't work.

 

357 is still in place, but got significantly overhauled. It wasn't just me who disliked giving more influence to fewer players, there was enough backlash for wg to undertake the overhaul. It was fundamentally bad from the outset which is why wg need to make such dramatic changes. They now use a range of templates and altered the frequencies such that the eponymous 357 is very rare nowadays. As to its relation to sbmm, this was about the influence of a match being down to a smaller pool of players. Which, when other factors are taken into account such as tank and class, can lead to even more imbalances. Imagine your team's good are in a brit light or arty, while the other team good is in a chieftain or 279.

 

View PostThe_Big_M, on 12 September 2020 - 02:04 AM, said:

As for the ratios, as given above that range will be 100% of influence now as they're evenly matched. It's not player vs player but team vs team - both having an equal chance. The 20% has now become the 100%. The expected contribution on the better player will be higher as a proportion of the team, but that keeps the challenge up rather than just making it a cake walk.  So you need to bring your A game each time, whatever type of player you are (though there are still all sorts of variables in the mix as at present, so it's not as dramatic a change as made out). 

You aren't getting the point of the 40:40:20 illustration. It references the spread of win rates in the population, which for simplicity i used 20% (ie between 40%ers and 60%ers). Under an sbmm that spread will be lower, hence the model would likely reduce to something like 49:49:2 - where the vast majority of the player pool will be between 48% and 50% (assuming mean stays at 49% given draws). Remember, under a sbmm, a 40%er is going to get no longer going to be vs an average player, they'll be vs a shitter. Their win rate will be close to 50%. And a good will no longer be vs an average player, they will be vs a good player, so their win rate will also be close to 50%.

 

Or to use the terms used in the model but for a shitter, you will be able to be shit and you'll win 49% of the time. And you can be shit and lose 49% of the time. While for a good player, you can be good and win 49%, or you can be good and lose 49% of the time. If this doesn't send off alarm bells for any of the sbmm people, then i suspect their alarm is broken.

 

View PostThe_Big_M, on 12 September 2020 - 02:04 AM, said:

Yes, because there is better balance, that means the challenge will regularly be tougher, rather than having so many cakewalks. If there is an implication on rewards that could be rebalanced to include draws if they became oversized as an outcome.

Tougher challenge or less challenge. For a good player it will be almost impossible to achieve the performance they do currently. 60%ers won't exist. While for a shitter, it will be near impossible to do as badly as they do now. They could play even worse than they do now and still win more than present. And yes, this is assuming rewards aren't rebalanced.

 

As i've suggested multiple times in the the previous sbmm discussions, the concept could work if wg completely overhaul rewards. They'd first need a skill metric other than win rate. Doesn't matter what it is but as everyone will be winning about the same amount, it clearly can't be win rate. Then for those that are considered good under that metric, they would need considerably more rewards every battle - eg a 3x multiplier. Average players don't need a tweak, while bads would get a third as much as they do now (assumes a 3 bucket skill system - naturally it would need to be more granular). Hence the incentive to actually play well and improve would be retained. But let's be honest, while it would be hilarious the sorts of threads that would come out due to such a system, it would not be a step forward for the game.

 

Finally, there is this underlying assertion that sbmm will make battles closer and stop steam rolls. The closest things to sbmm we have currently show no evidence that this will be the case. If anything, they show the opposite. Maps, teams, tanks just simply don't interact for there ever to be a balance like that. Consider an sbmm where same tanks are driven by same skill players (which as noted would be a shit fight for the mm trying to find). Next those players would have to go and fight over the same part of the map in a 1v1. And then, once they finished fighting, they must not then contribute to any of the other 1v1s. Then, and only under those farcical limitations, would we see battles that would end up fairly equal all the time.

 

Or to put it another way, steamrolls are rarely due to full team skill imbalances. They are mostly due to one team playing one area, starting the snow ball and from that point it's gg. 3 tanks early turns into 5 tanks when that mob show up somewhere else. Finish off the battle and it's a 7 tank or more lead. It's just how battle momentum works.


Edited by Ezz, 12 September 2020 - 08:46 AM.

Who the [edited] are you? Get Spoofed! "wouldn't be a proper WG balance change if they didn't [edited] something up after all "

>9000 cynicism brought to you by P2W, Balance TM and the Cartoon Connection

R. Pubbie: "why are all PBKAC players so rude, arrogant and nasty? and why do Mods favor them?"


NameWasStolenStresslevel #47 Posted 12 September 2020 - 08:55 AM

    Captain

  • Member
  • 50129 battles
  • 1,429
  • Member since:
    01-19-2015

While playing malinovka some time ago in my su130pm, I came across with an EBR105 on my team (who died obviously because he was in my team, and because we were playing malinovka). 

 

He called me "noob" because I didnt sit on cap point with out full HP autoloader TVP tier 10. While a patton tier X and another guy coming down the hill to decap base. The almighty EBR called me "noob" while on death camera - probably jealous of my 3 gun marks - just because instead of going to the uncappable cap, I went to seek the tier 8 premium french TD siting in that infamous bush near that infamous lake that SPG always sit on before drowing end game.

 

Do anyone need to bring stats here? Just by what I typed here, top tier EBR dead in malinovka ignoring a TD just waiting for the right moment to decap the TVP, while me "the damage red liner noob stat padder 3 gun marks" went to try to Spot the sneak TD before shot us to death as soon as Patton and the other guy come - and even if they dont come. ( we won that game because I not only spotted the french TD and got him killed, also killed the patton myself who obviously came shortly after I spotted the TD)

 

Check his stats, he was a 52% overall. While saw so many "group award" medals on his account that would explain why he can even be above 50%, but sucks whenever go solo.

 

Toxic shiter 52% player who SBMM would never help. But 52% isnt bad (its actually good, depending on what kind of player we talking about), it just that winrate can be a trick factor for the already congested MM (that deal with platoons, EBRs, OP tanks, and shit map rotation) to deal with.

 

Skill based MM in randons would kill the "random" of it. So keep it reserved to Ranked. Go play ranked. It happens once every 6months.

 

 


Edited by NameWasStolenStresslevel, 12 September 2020 - 08:58 AM.

Having META problems?

 #Hashtag break the META

     (HAHAGHa i must have drank somethin)

             https://worldoftanks.asia/en/hall-of-fame/vehicles/A100_T49/#wot&w_ts=overall&w_nb=500

         Currently Hall of Fame T49's number 1 in WTR!!! (parameters are *battles +500* and *since 2014's* ranking)


Ezz #48 Posted 12 September 2020 - 09:14 AM

    How many flipping posts do I need to get past Major?

  • Beta-Tester
  • 74444 battles
  • 36,903
  • [PBKAC] PBKAC
  • Member since:
    07-17-2012
The problem with ranked is bads don't get anything. That's why they like the idea of sbmm in randoms. They'd be getting more under the current system while goods would be getting less.

Who the [edited] are you? Get Spoofed! "wouldn't be a proper WG balance change if they didn't [edited] something up after all "

>9000 cynicism brought to you by P2W, Balance TM and the Cartoon Connection

R. Pubbie: "why are all PBKAC players so rude, arrogant and nasty? and why do Mods favor them?"


MagicalFlyingFox #49 Posted 12 September 2020 - 12:57 PM

    Destroyer of Tier 6 CW

  • Beta-Tester
  • 35498 battles
  • 13,849
  • [ATLUS] ATLUS
  • Member since:
    10-03-2012

The way they are talking about the state of random battles now is like the enemy team always has a unicum Chieftain/279e while their team has a red E5 as their heavy tank every single tier 8 game they play.

Always steamrolls because they have the better players in the better tanks.

 

Distributing good players on both sides will not make a difference in those situations, or it will even make it harder to switch the result from the almost predetermined one of the Chief doing 7k damage and winning.

 

 

This game has more issues with tank balance and map design than it does with the MM.

 

This entire argument was built on the premise that 0-15 wipeouts and steamrolls are increasingly commonplace. Statistically that is not true, otherwise we'd be having wipe-outs every 2nd game by now. We've been on these forums for years and have played this game for years. Every single patch, every single year there is a thread or comment saying how "There are more steamrolls now, this game is shit". Even back then, when we had better tools like vbaddict to analyse the results of games over large sample sizes the data showed no trend of increasing wipeouts.

 

Going back to the OP (original post), the real reason why the game is in such a shit state is as he says. OP (overpowered) premiums getting more and more stronger. Just look at the amount of premium tanks there are at tier 8 compared to tech tree tanks, and remember how shit the tech tree tanks are to their premium contemporaries. WG's game balancing has been a joke for years.


http://www.theuselessweb.com/

 A. Guy on 02 June 2018 - 12:40 AM, said:

Destroyer of Tier 6 CW... says it all about you.


The_Big_M #50 Posted 12 September 2020 - 02:03 PM

    Corporal

  • Beta-Tester
  • 69684 battles
  • 51
  • [TKSA] TKSA
  • Member since:
    09-28-2012

View PostCardinalMite, on 12 September 2020 - 08:09 AM, said:

Exactly which stat do you use for this?

 

Overall win rate? Overall Win rate in the class of tank they are playing? Overall win rate in the tier they are queuing in? Overall win rate in the specific tank they are playing?

 

But you have 2 52% overall players in the match. One has 30k games, plays mainly tiers 7,8,9,10 is playing a tier 9 LT that they have a 60% win rate in after 200 games in it. The other has 30k games, has a 48% highest win rate in LTs at tier 9 playing a tier 9 LT, 5k of the games they have are in a tier 5 T67 with a 54% win rate in it. Most of the rest of their tanks played are < tier 7.

 

People put out whines that if only MM had SBMM everything would be more balanced, more fair that it is an easy fix. I don't think that is true at all. I don't think people really think through what it looks like in practise.

 

Most likely PR. 

 

Remember, we're not trying to achieve perfection and don't expect it to be perfectly 50/50, just get rid of the rough edges which often exist- 40:40 as suggested by some.



The_Big_M #51 Posted 12 September 2020 - 02:05 PM

    Corporal

  • Beta-Tester
  • 69684 battles
  • 51
  • [TKSA] TKSA
  • Member since:
    09-28-2012

View PostEzz, on 12 September 2020 - 09:27 AM, said:

How much you win, ie what your stats are, is a fundamental part of the game, so obviously intrinsic to the enjoyment.

 

I dont believe that, given most don't use xvm which is about stats, most knock the win chance and other aspects, it's not a part of the promotional encouragement (which WG would use if it was actually attracting customers).

 

As said, of course some use it, but it's not the primary enjoyment for most. And I can't imagine anyone logging in if they didn't improve the game "just to improve their stats" which are pointless otherwise.

 

And it's still immaterial. As said, if stats need to change to better reflect new MM, they will just as they have in the past, and there's no perfect stat measurement now anyway.


Edited by The_Big_M, 12 September 2020 - 02:06 PM.


The_Big_M #52 Posted 12 September 2020 - 02:09 PM

    Corporal

  • Beta-Tester
  • 69684 battles
  • 51
  • [TKSA] TKSA
  • Member since:
    09-28-2012

View PostCardinalMite, on 12 September 2020 - 09:43 AM, said:

So you played the tank 1 game, it was a landslide win. Your recents in that tank are 100% win rate. You play the tank 9 more times, you are blessed by winning 6 of those, you didn't do all that much or contribute all that much to the wins, but hey everyone gets win and loss streaks right? You are a 70% win rate unicum in the tank after 10 battles. Say your overall account win rate is 48%, you play another 90 battles in the tank, only because you are a 48% player on average, and the tank is higher tiers where you average more like 44%, as the number of games increases towards that hundredth game your win rate for the tank drops to 45%. MM would have matched you completely wrongly for most of those matches. As it would have done for most of the other players on the teams, most of the time. 

 

Picking teams as a kid worked well because human beings don't measure the capability of their fellows by statistics, they don't need to because they generally know all the other kids being picked and how good a player they are without particular measures. MM doesn't know one player ID from another by any means other than statistics.

 

Point is the upshot is that your skill matching per match is meaningless, the stats only work over statistical samples, even there you get anomalies. How do you compare someone who played a hundred battles in a tank that they ground from completely stock, versus someone that played 100 battles but free XP'd all the modules? The kind of matching you talk about is only worth doing if it is accurate, if it isn't then it literally would be flat out worse even by the goals you set for MM in introducing that type of matching.

 

Yet it was stated previously that Ranked battles are SBMM... 

 

As said above, PR is the obvious metric that comes to mind. It's a generally good indication of general skill and that's what we're talking about. Not having teams of reds+ oranges with 2 yellows vs teams with 6 greens, 8 yellows and 1 blue.



Ezz #53 Posted 12 September 2020 - 02:09 PM

    How many flipping posts do I need to get past Major?

  • Beta-Tester
  • 74444 battles
  • 36,903
  • [PBKAC] PBKAC
  • Member since:
    07-17-2012

View PostThe_Big_M, on 12 September 2020 - 04:05 PM, said:

 

I dont believe that, given most don't use xvm which is about stats, most knock the win chance and other aspects, it's not a part of the promotional encouragement (which WG would use if it was actually attracting customers).

 

As said, of course some use it, but it's not the primary enjoyment for most. And I can't imagine anyone logging in if they didn't improve the game "just to improve their stats" which are pointless otherwise.

 

And it's still immaterial. As said, if stats need to change to better reflect new MM, they will just as they have in the past, and there's no perfect stat measurement now anyway.

The point was given the way the game is now, enjoyment for many people involves doing well. If i have a good day, that will be reflected in my stats. Similarly if i have a shit day, that too tends to be reflected. Hence the stats become a convenient way of measuring the levels of enjoyment i had in game.


Who the [edited] are you? Get Spoofed! "wouldn't be a proper WG balance change if they didn't [edited] something up after all "

>9000 cynicism brought to you by P2W, Balance TM and the Cartoon Connection

R. Pubbie: "why are all PBKAC players so rude, arrogant and nasty? and why do Mods favor them?"


Ezz #54 Posted 12 September 2020 - 02:12 PM

    How many flipping posts do I need to get past Major?

  • Beta-Tester
  • 74444 battles
  • 36,903
  • [PBKAC] PBKAC
  • Member since:
    07-17-2012

View PostThe_Big_M, on 12 September 2020 - 04:09 PM, said:

Yet it was stated previously that Ranked battles are SBMM... 

It's one way of achieving it. As you progress up the ranks you face more of the other players who have progressed and fewer of the qualification and div 3 randoms. By div 1 (especially early on) you hardly see any bads, and in fact most of the time you're facing only skilled players.

 

There are other methods too. One being pool based, where goods only play other goods, average players only vs avg, and bads only vs bads. While another method would be closer to a 'here's 30 players, team 1 gets the best player, team 2 gets the second best... etc..' Yet this final method has the glaring issue of tank tier class etc..


Edited by Ezz, 12 September 2020 - 02:14 PM.

Who the [edited] are you? Get Spoofed! "wouldn't be a proper WG balance change if they didn't [edited] something up after all "

>9000 cynicism brought to you by P2W, Balance TM and the Cartoon Connection

R. Pubbie: "why are all PBKAC players so rude, arrogant and nasty? and why do Mods favor them?"


The_Big_M #55 Posted 12 September 2020 - 02:25 PM

    Corporal

  • Beta-Tester
  • 69684 battles
  • 51
  • [TKSA] TKSA
  • Member since:
    09-28-2012

View PostEzz, on 12 September 2020 - 10:45 AM, said:

357 is still in place, but got significantly overhauled. It wasn't just me who disliked giving more influence to fewer players, there was enough backlash for wg to undertake the overhaul. It was fundamentally bad from the outset which is why wg need to make such dramatic changes. They now use a range of templates and altered the frequencies such that the eponymous 357 is very rare nowadays. As to its relation to sbmm, this was about the influence of a match being down to a smaller pool of players. Which, when other factors are taken into account such as tank and class, can lead to even more imbalances. Imagine your team's good are in a brit light or arty, while the other team good is in a chieftain or 279.

 

Quibbling at the edges like that doesn't negate the improvement. Not every match is going to have that, and not every match is going to rely on that. 

 

 

View PostEzz, on 12 September 2020 - 10:45 AM, said:

You aren't getting the point of the 40:40:20 illustration. It references the spread of win rates in the population, which for simplicity i used 20% (ie between 40%ers and 60%ers). Under an sbmm that spread will be lower, hence the model would likely reduce to something like 49:49:2 - where the vast majority of the player pool will be between 48% and 50% (assuming mean stays at 49% given draws). Remember, under a sbmm, a 40%er is going to get no longer going to be vs an average player, they'll be vs a shitter. Their win rate will be close to 50%. And a good will no longer be vs an average player, they will be vs a good player, so their win rate will also be close to 50%.

 

Or to use the terms used in the model but for a shitter, you will be able to be shit and you'll win 49% of the time. And you can be shit and lose 49% of the time. While for a good player, you can be good and win 49%, or you can be good and lose 49% of the time. If this doesn't send off alarm bells for any of the sbmm people, then i suspect their alarm is broken.

 

 

If it referenced the spread of win rates it should be more like 40:20:40. I've always seen it discussed as impact on outcome so that's the context I used it.

 

I addressed earlier if the win-rates narrow. I don't see it as a big deal. It's just a new normal if it happens. The granularity will then be that the decimal gets important e.g. 49.2% vs 50.1% It's just a number, not a meaning of itself. There's nothing that makes the current 40-60% special of itself. It just describes the situation (approximately).

 

View PostEzz, on 12 September 2020 - 10:45 AM, said:

 

Tougher challenge or less challenge. For a good player it will be almost impossible to achieve the performance they do currently. 60%ers won't exist. While for a shitter, it will be near impossible to do as badly as they do now. They could play even worse than they do now and still win more than present. And yes, this is assuming rewards aren't rebalanced.

 

As i've suggested multiple times in the the previous sbmm discussions, the concept could work if wg completely overhaul rewards. They'd first need a skill metric other than win rate. Doesn't matter what it is but as everyone will be winning about the same amount, it clearly can't be win rate. Then for those that are considered good under that metric, they would need considerably more rewards every battle - eg a 3x multiplier. Average players don't need a tweak, while bads would get a third as much as they do now (assumes a 3 bucket skill system - naturally it would need to be more granular). Hence the incentive to actually play well and improve would be retained. But let's be honest, while it would be hilarious the sorts of threads that would come out due to such a system, it would not be a step forward for the game.

 

Finally, there is this underlying assertion that sbmm will make battles closer and stop steam rolls. The closest things to sbmm we have currently show no evidence that this will be the case. If anything, they show the opposite. Maps, teams, tanks just simply don't interact for there ever to be a balance like that. Consider an sbmm where same tanks are driven by same skill players (which as noted would be a shit fight for the mm trying to find). Next those players would have to go and fight over the same part of the map in a 1v1. And then, once they finished fighting, they must not then contribute to any of the other 1v1s. Then, and only under those farcical limitations, would we see battles that would end up fairly equal all the time.

 

Or to put it another way, steamrolls are rarely due to full team skill imbalances. They are mostly due to one team playing one area, starting the snow ball and from that point it's gg. 3 tanks early turns into 5 tanks when that mob show up somewhere else. Finish off the battle and it's a 7 tank or more lead. It's just how battle momentum works.

 

Here, we somewhat agree. As said earlier I don't see the ranked battles mode as an analogue. Also, in your example, that may well be that you have one team who just have no idea.

 

I would only like to see this trialled, say on sandbox. I'm not arguing I know it's going to be better or nirvana, but I just think things could be better than they are, and it does seem like it would improve. And the main issue I brought it up again was I hadn't seen a definitive reason against it, although there'd been lots of comments saying 'it's been discussed before and a bad move' without any justification. I think this coaxed out that it isn't true, but as said earlier, this is hypothetical as WG have shown no desire to change so it's not like there is any likelihood of this happening anway. 

 


Edited by The_Big_M, 12 September 2020 - 02:28 PM.


CardinalMite #56 Posted 12 September 2020 - 04:11 PM

    Captain

  • Member
  • 32109 battles
  • 1,389
  • [PBKAC] PBKAC
  • Member since:
    04-18-2014

View PostThe_Big_M, on 12 September 2020 - 06:03 AM, said:

 

Most likely PR. 

 

Remember, we're not trying to achieve perfection and don't expect it to be perfectly 50/50, just get rid of the rough edges which often exist- 40:40 as suggested by some.

Great platitudes, lots of vague statements. Doesn't address the issue I was pointing out that it won't achieve what you say it will.

 

And it doesn't matter which stat system you pick, recents, overall W/R. PR, winrate in selected tank. Using it to "balance" teams is a fools errand that will not "get rid of rough edges," it will be flat out wrong more often than it will be right.

 

Or to put it in your terms, "it will make the rough edges worse in practise."


“Holiday ops is balanced for Asia because a good player opening 11 boxes will get just as many decorations as a noob opening 75..”—.Murazor new head of global festive events.

 

"It's like you literally dont read your own posts, let alone others, make up instead what you wished others said, then put a non-sensical counter argument to that imagined post."

 

"Never argue with an idiot, they drag you down to their level then beat you with experience. But then if I limited myself to only arguing with non-idiots on these forums I'd literally never post"


CardinalMite #57 Posted 12 September 2020 - 04:23 PM

    Captain

  • Member
  • 32109 battles
  • 1,389
  • [PBKAC] PBKAC
  • Member since:
    04-18-2014

View PostThe_Big_M, on 12 September 2020 - 06:09 AM, said:

 

Yet it was stated previously that Ranked battles are SBMM... 

 

As said above, PR is the obvious metric that comes to mind. It's a generally good indication of general skill and that's what we're talking about. Not having teams of reds+ oranges with 2 yellows vs teams with 6 greens, 8 yellows and 1 blue.

What makes you think I would hold Ranked Battles as any good example of SBMM?

 

Whatever the stated purpose of Ranked Battles is. One thing it is not is a more enjoyable way of playing random battles. In fact it is the absolute poster child of how it would make Random Battles significantly worse, less enjoyable, more toxic etc .etc. if something like its match up mechanism was employed in randoms.

 

As for PR, it is just another artificial metric offered as an alternative measure to win rate. It enhances understanding of player capabilities not one bit vs other measures and is just as prone to distortion and padding as those.

 

It also would not work on a granular level of per match, per player, in a particular tank.

 

Still it might mean that in your colour example more of the players will look similar and give you a faux warm fuzzy feeling that it is all better matched now.... at least before the match actually happens and you realise none of those players rated as "similarly capable " actually weren't at all.

 

Mission accomplished. 


Edited by CardinalMite, 12 September 2020 - 04:24 PM.

“Holiday ops is balanced for Asia because a good player opening 11 boxes will get just as many decorations as a noob opening 75..”—.Murazor new head of global festive events.

 

"It's like you literally dont read your own posts, let alone others, make up instead what you wished others said, then put a non-sensical counter argument to that imagined post."

 

"Never argue with an idiot, they drag you down to their level then beat you with experience. But then if I limited myself to only arguing with non-idiots on these forums I'd literally never post"


Ezz #58 Posted 12 September 2020 - 05:13 PM

    How many flipping posts do I need to get past Major?

  • Beta-Tester
  • 74444 battles
  • 36,903
  • [PBKAC] PBKAC
  • Member since:
    07-17-2012
The summary is... could the game be made more balanced from a stats of players perspective - most likely yes. Would that lead to more enjoyable gameplay - mostly likely no.

Edited by Ezz, 12 September 2020 - 05:26 PM.

Who the [edited] are you? Get Spoofed! "wouldn't be a proper WG balance change if they didn't [edited] something up after all "

>9000 cynicism brought to you by P2W, Balance TM and the Cartoon Connection

R. Pubbie: "why are all PBKAC players so rude, arrogant and nasty? and why do Mods favor them?"


afoxq #59 Posted 12 September 2020 - 05:47 PM

    Ultimate LUCK Killstealer

  • Member
  • 38544 battles
  • 1,634
  • [LUCK] LUCK
  • Member since:
    10-08-2012

View PostThe_Big_M, on 12 September 2020 - 02:03 PM, said:

 

Most likely PR. 

 

Remember, we're not trying to achieve perfection and don't expect it to be perfectly 50/50, just get rid of the rough edges which often exist- 40:40 as suggested by some.


Lol, as soon as he said use PR, I was like this should have ended the discussion.
 


Survived the Bread & Butter Rektage of 2015/16

Survived the DPS Bus Apocalypse of 2014

Winning was is LUCK™


AlexTheKid72 #60 Posted 12 September 2020 - 08:12 PM

    Sergeant

  • Member
  • 28850 battles
  • 199
  • [SABRE] SABRE
  • Member since:
    05-28-2017

View PostCardinalMite, on 12 September 2020 - 07:43 AM, said:

So you played the tank 1 game, it was a landslide win. Your recents in that tank are 100% win rate. You play the tank 9 more times, you are blessed by winning 6 of those, you didn't do all that much or contribute all that much to the wins, but hey everyone gets win and loss streaks right? You are a 70% win rate unicum in the tank after 10 battles. Say your overall account win rate is 48%, you play another 90 battles in the tank, only because you are a 48% player on average, and the tank is higher tiers where you average more like 44%, as the number of games increases towards that hundredth game your win rate for the tank drops to 45%. MM would have matched you completely wrongly for most of those matches. As it would have done for most of the other players on the teams, most of the time. 

 

Picking teams as a kid worked well because human beings don't measure the capability of their fellows by statistics, they don't need to because they generally know all the other kids being picked and how good a player they are without particular measures. MM doesn't know one player ID from another by any means other than statistics.

 

Point is the upshot is that your skill matching per match is meaningless, the stats only work over statistical samples, even there you get anomalies. How do you compare someone who played a hundred battles in a tank that they ground from completely stock, versus someone that played 100 battles but free XP'd all the modules? The kind of matching you talk about is only worth doing if it is accurate, if it isn't then it literally would be flat out worse even by the goals you set for MM in introducing that type of matching.

I said pick a stat ... anything is better than 15 awesome players in similar tanks to your team on the other team. 
Picking teams as a kid worked well because the best got picked first, then the other team picked the second best ..... and so on. Everyone had a gg, win or lose you knew it was fair coz the picks were 1 per team.
I AM NOT TALKING SKILL MM, JUST SKILL DISTRIBUTION!


Edited by AlexTheKid72, 12 September 2020 - 08:13 PM.

signature.png




1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users