Jump to content


Steamroll wins/losses - Do they have to be over quickly and other misconceptions...


  • Please log in to reply
68 replies to this topic

SimplyPzB2 #1 Posted Aug 15 2019 - 06:25

    First lieutenant

  • -Players-
  • 106 battles
  • 795
  • Member since:
    05-26-2016

So this Topic is intended to discuss the game mechanic that is the matchmaker.  Specifically the effect on gameplay that occurs because the MM does not factor in "Team Skill Balance". 

-

I want to point out right up front I have no issue with any of the 30 players who were in the following sample battle.  They all appeared to play hard, and I'm glad they were here as they add to the playerbase.  This Topic will only speak to the "Team Dynamic" that is "Team Skill". 

-

It's my opinion that it's a good thing for the game that the current MM balances both Tiers and Tank Class.  I think if it didn't, it would really put a damper on the quality of gameplay.  Right in line with this opinion is the fact that the current MM DOES NOT factor in team skill.  This post is a humble request to the WG powers that be to consider changing the MM to include skill - specifically balance the skill between both teams.  To be clear, I'm not asking for "perfect" balance.  For one thing, I thinks that's an impossible ask.  But, I do think it's possible to get the skill between both teams "reasonably close".

-

I'm going to provide one sample battle as a reference point.  Know that these battles happen hourly.  And based on many many other player's data, happen about for about 40% of all battles.  (20% in your favor, 20% not in your favor).  So it's the same "level" of an issue as being "bottom tier" used to be back in 3/5/7.   It's just that you can't "see" it happening as easily.  So please, spare us the 'small sample size comments'.  We all know I could post thousands of these (you know if I had THAT kind of free time...)

-

I've chose this particular Steamroll because a very common misconception is that "to be a blowout" a battle has to be "over in under 'x' minutes".  You can't arbitrarily pick "a time" that defines a blowout.  Too many other factors.  Big map?  Lots of slow heavies?  Lots of camping TDs?  Brawling map?  Sniper map?  Etc...

-

Pre-Battle Team Stats: The Enemy team has ONE player with a win rate at 50%. Just ONE!!!!

-

The enemy team "gets campy" early. They didn't know the map meta, so they just drove up to certain point and camped. The two enemies that "tried" to scout, died super quick. But camping on this map with a lot of TDs make it hard to be dug out. There was never any question this was going to be a win, but due to the map and tank spread (lots of TDs) it was going to "take a while" for the steamroll to happen.

-

Steamroll you say? What Steamroll? While this "gives the appearance" it's close. It's not. Look at the mini-map. We control the south, mid, and north. Again, they are camping hard, but in the end it's just delaying the inevitable.

-

WHAAAAT? They are in the lead? (At least in kills...). Again, appearances are deceiving. We've lost our scouts and a couple of "orange skill" players who pushed to hard into a camping team. Note the time, 8 minutes and we are down one kill.

-

"One minute later..." We are now up two kills and 3,000hp. Told you, appearances were deceiving. The game was over before it started. The enemy was never ever going to push into us effectively (had they, we would have simply mowed them down as they crossed the open map). THAT they were camping and it took more than 5 minutes to dig them out DOES NOT mean this wasn't a steamroll.

-

Another 90 seconds and we're up 5 kills and 4,000hp. At this point my team was just crawling all over each other to squeak in what little extra damage/experience we could.

-

Boom. There you have it. Won by 8 tanks and 5,000hp. Steamroll. Sure, it took over ten minutes, and we didn't "pull away" till late. But this win was more inevitable than Thanos....

-

FYI, I didn't need to be here for this win. Had I gone 100% afk, we still would have won. As for "my personal skill", I didn't really use any. Sniped south for the first half (never spotted), then when that dried up, moved north and sniped (never spotted). Only time I was spotted happened at the very end as I was racing a couple teammates to get to the last enemy alive. I got a couple shots in before one of my teammates took him out.

-

Note that even though the enemy was waaay less skilled, most of them managed to get a little damage. But that's the difference between a skilled team vs a much less skilled team. The top third of my team did over 1,000+ damage. Whereas the enemy team only had two players who managed to break 1,000+ damage. Like shooting fish in a barrel. - Why? Because the "random MM" didn't factor in team skill. The MM put ALL the skill on one team. I mean COME ON, the enemy had ONE PLAYER with a 50% win rate. Our team had four green-skill and one blue-skill player. OUR WORST players had 49% winrate. - Battles like this should never ever be allowed to happen.

-

So what could SBMM have done to make this battle competitive?  It's simple, swap a few players from the already selected teams.  As long as the 'swaps' occur 'horizontally', tier and tank class balance will be maintained.  An easy example of 'horizontal only' swapping would be you can only swap the top tank with the top tank of the other team.  So for this battle, you could only swap the Somua with the Emil.  Top tier tank for top tier tank/ heavy for heavy. 

-

To keep this simple, all MM would have to do is make sure both teams have 'about' the same number of "green-skill", "yellow-skill", "orange-skill", and "red-skill" players.  Platoons could only be swapped if the proposed swapped worked for all the tanks in the toon.

-

To keep this simple part II, let's just look at the "green-skilled" players.  There are a total of (5) "green-skilled" players.  "Ideally", there should be (3) on one team, and (2) on the other team.  So for this battle there are several options, I'm indifferent as to which is done:

-

Option 1: Swap Somua and E25 to enemy team.  This would leave (3) green-skill on Team 1, and put (2) green-skill on Team 2.

-

Option 2: Swap Somua, and Toon 2 (the T-50-2 and AMD).  This would leave (2) green-skill on Team 1, and put (3) green-skill on team 2.  Note how Toon 2 CAN BE swapped horizontally as the AMX and Type 64 are not in "conflicting" toons.  Tank tier and class are maintained.

-

Once this is done for the "green-skill", the SBMM would do this "as needed" down the skill line.  "Yellow-skill" blanced next, then "Orange-skill", etc.

-

-

Another misconception is this would somehow "punish" the higher skilled players.  This is patently false.  None of the 30 players had anything to do picking the teams.  It's 100% "luck of the draw".  So some will say "..but but you are moving some of the good players to a worse team".  This CAN'T be true.  Once a couple of good players are moved to the "worse" team, it's NO LONGER a "worse team".  This is a key component of SBMM.  THERE ARE NO BAD TEAMS, AND THERE CAN'T BE. 

-

-

The final misconception is that SBMM will somehow reduce the "Variety" that "supposedly" comes from unbalanced teams facing off.   That simple fact of the matter is that no algorithm is every going to "perfectly balance" 30 human players.  And there will still be a HUGE variety of skill in SBMM.  Look at this sample battle, only (6) out of (30) are the higher green-skill players.  In the next battle there might be (16/30) green-skilled player, or there might be (2/30).  You will get the EXACT SAME VARIETY of skilled players under sbmm, the only difference will be that BOTH teams will have about the same amount of skill. 

-

This actually might be the most important point.  Let's say each team has just (1) green-skilled player.  Which one of them is going to carry the 14 orange/red-skilled players?  That's a serious challenge.  Or what if there are (13) green-skilled players on each team?  Which team is going to coalesce better and pull out the win?  That's a serious challenge. 

-

This sample battle just wasn't a challenge...  at all...  These battles shouldn't exit. 

-

Please WG, get on this. 

-

(Cynical side note:  Look at it this way WG, if you DO balance the skill...   Tank class, Tier, AND Team Skill will be balanced.  No one will be able to come to the forums and complain they got the short end of the stick....  That should make the effort totally worth it right there...)

-

Peace, Love, and Long Live WOT....

 



NeatoMan #2 Posted Aug 15 2019 - 06:45

    Major

  • Players
  • 28756 battles
  • 21,249
  • Member since:
    06-28-2011

View PostSimplyPzB2, on Aug 15 2019 - 00:25, said:

I'm going to provide one sample battle as a reference point.  Know that these battles happen hourly.  And based on many many other player's data, happen about for about 40% of all battles.  (20% in your favor, 20% not in your favor).  So it's the same "level" of an issue as being "bottom tier" used to be back in 3/5/7.   It's just that you can't "see" it happening as easily.  So please, spare us the 'small sample size comments'.  We all know I could post thousands of these (you know if I had THAT kind of free time...)

About your reference battle.

team 1 total: 87336

team2 total: 56970

win chance: >90% {we'll call it 90% since that's where it seems to max out}

 

your reference battle happens once every 20 battles, or 5% of the time.  2.5% for and 2.5% against.   just though I'd let everyone know.

 

Block Quote

I've chose this particular Steamroll because a very common misconception is that "to be a blowout" a battle has to be "over in under 'x' minutes".  You can't arbitrarily pick "a time" that defines a blowout.  Too many other factors.  Big map?  Lots of slow heavies?  Lots of camping TDs?  Brawling map?  Sniper map?  Etc...

 IOW, a blowout is a blowout is a blowout.

 

 

after battle report shows that:

the skorpian (a yellow player) did most of the heavy lifting. 

you, (a green player) did quite a bit as expected.

the somua (green) did OK

two more orange players did OK

 

 

did the overwhelmingly skilled players really cause that blowout?


Edited by NeatoMan, Aug 15 2019 - 06:52.


SimplyPzB2 #3 Posted Aug 15 2019 - 07:00

    First lieutenant

  • -Players-
  • 106 battles
  • 795
  • Member since:
    05-26-2016

View PostNeatoMan, on Aug 15 2019 - 06:45, said:

About your reference battle.

team 1 total: 87336

team2 total: 56970

win chance: >90% {we'll call it 90% since that's where it seems to max out}

 

your reference battle happens once every 20 battles, or 5% of the time.  2.5% for and 2.5% against.   just though I'd let everyone know.

 

 IOW, a blowout is a blowout is a blowout.

 

 

after battle report shows that:

the skorpian (a yellow player) did most of the heavy lifting. 

you, (a green player) did quite a bit as expected.

the somua (green) did OK

two more orange players did OK

 

 

did the overwhelmingly skilled players really cause that blowout?


Um, you yourself said that these types of battles happen +/-40% of the time (keeping in mind that half the time/20% they are in your favor, the other half/20% they are not in your favor).  Do I really need to go grab the dozen other screenshots of similar battles I'VE POSTED THIS WEEK?  I think the forum users get it, that these battles happen ever single hour of gameplay. 10/90%, 20/80%, 30/70%... every hour of every day...

-

Super disingenuous for you to suggest otherwise....

-

Double super disingenuous for you to try and "muddy the waters" by looking at "individual" tanks.  Everyone can see this battle was won before it even started.  Team 1 would have had to "Team Yolo" ™ or "Team Drown" ™ to get a loss out of this one.

-

Considering the concepts of "fair play", "competitive play", etc, it's IMPOSSIBLE to defend this battle.  Unless fair play and competitiveness are concepts you don't think are important to quality gameplay.   I for one, prefer to actually EARN MY WINS....



NeatoMan #4 Posted Aug 15 2019 - 07:14

    Major

  • Players
  • 28756 battles
  • 21,249
  • Member since:
    06-28-2011

View PostSimplyPzB2, on Aug 15 2019 - 01:00, said:

Um, you yourself said that these types of battles happen +/-40% of the time

No i didn't say that 90/10 battles happen 40% of the time.   A 90/10 battle is quite different from a 70/30 battles BY A LOT.   You simply lump them together and then throw out the most extreme example as if it's representative of the whole bunch.  nice try... but no



Nixeldon #5 Posted Aug 15 2019 - 07:20

    Major

  • Players
  • 61771 battles
  • 2,416
  • Member since:
    10-30-2011
BudhaMathTM

SimplyPzB2 #6 Posted Aug 15 2019 - 07:46

    First lieutenant

  • -Players-
  • 106 battles
  • 795
  • Member since:
    05-26-2016

View PostNeatoMan, on Aug 15 2019 - 07:14, said:

No i didn't say that 90/10 battles happen 40% of the time.   A 90/10 battle is quite different from a 70/30 battles BY A LOT.   You simply lump them together and then throw out the most extreme example as if it's representative of the whole bunch.  nice try... but no


Not really, 40-60%-ish chance to win can be considered 'balanced'.  So 70/30 is out of balance, as is 80/20 and 90/10.   Add these all up and you get 40% of all your battles are un-blanced....



NeatoMan #7 Posted Aug 15 2019 - 07:49

    Major

  • Players
  • 28756 battles
  • 21,249
  • Member since:
    06-28-2011

View PostSimplyPzB2, on Aug 15 2019 - 01:46, said:

Not really, 40-60%-ish chance to win can be considered 'balanced'.  So 70/30 is out of balance, as is 80/20 and 90/10.   Add these all up and you get 40% of all your battles are un-blanced....


but the player makeup of a 70/30 battle is far different than a 90/10 battle, so it's a bit disingenuous putting out only the rare extreme ones and then claiming "this happens 40% of the time"



SimplyPzB2 #8 Posted Aug 15 2019 - 08:10

    First lieutenant

  • -Players-
  • 106 battles
  • 795
  • Member since:
    05-26-2016

View PostNeatoMan, on Aug 15 2019 - 07:49, said:


but the player makeup of a 70/30 battle is far different than a 90/10 battle, so it's a bit disingenuous putting out only the rare extreme ones and then claiming "this happens 40% of the time"


Nope, not at all.  40-60% is a 'solid' range for 'balanced' battles.  Could it be 39-61%?  Sure.  How about 42-58%. Sure.   Point is, it doesn't have to be and exact number.  Like the number of heavies in a battle.  There is no limit (that I'm aware of), but WG has decided that one team can only have (1) more heavy than the other - you know for balance.  So we pick a reasonable range, 40-60%, and go with it.  It's that simple....



BlaqWolf #9 Posted Aug 15 2019 - 08:15

    First lieutenant

  • Players
  • 22774 battles
  • 577
  • [-W-W-] -W-W-
  • Member since:
    04-21-2011
What're you complaining about, OP?  At least (from the images above) you seem to have been on the 'skill heavy' team most of the time...

But yes, WG has stacked teams like this since it was released (it became apparent once XVM came on the scene).  I had to give up on the rainbow years ago because I couldn't take the pain seeing an entire team of red on my side versus a team with maybe three reds on the other side.
Nor did I particularly like it when *my* team had all of the blues/purples because I knew I was going to be able to contribute little to nothing while they mashed the potatoes.  I couldn't even add butter before the game was over.

BALANCE (not segregated queues) needs to be implemented a bit more homogeneously than it has been the last decade.  Just weigh by PR - which is not the most accurate measure of skill, and it can be manipulated - if the manipulator is willing to have their all-important XVM color get downgraded).  A variance of +/- 15% between the teams would go a LONG way.  If one team gets a high PR platoon the rest of their team will probably suck while the opposing team may be rather even overall - they'll still probably get streamrolled by the purples, anyway, but at least they have a chance.

NeatoMan #10 Posted Aug 15 2019 - 08:21

    Major

  • Players
  • 28756 battles
  • 21,249
  • Member since:
    06-28-2011

View PostSimplyPzB2, on Aug 15 2019 - 02:10, said:

Nope, not at all.  40-60% is a 'solid' range for 'balanced' battles.  Could it be 39-61%?  Sure.  How about 42-58%. Sure.   Point is, it doesn't have to be and exact number.  Like the number of heavies in a battle.  There is no limit (that I'm aware of), but WG has decided that one team can only have (1) more heavy than the other - you know for balance.  So we pick a reasonable range, 40-60%, and go with it.  It's that simple....


totally missed the point.  I'm not arguing about the range you chose as unbalanced.  It's the choice of the most extreme example and claim that it is representative of 40% of battles:

 

"I'm going to provide one sample battle as a reference point.  Know that these battles happen hourly.  And based on many many other player's data, happen about for about 40% of all battles."

 

It is not even close to representing what happens in 40% of battles.  It only represents what happens at the very extremes of unbalance in 5% of all battles.


Edited by NeatoMan, Aug 15 2019 - 08:26.


oldewolfe #11 Posted Aug 15 2019 - 08:36

    Major

  • -Players-
  • 24993 battles
  • 3,683
  • [1STBN] 1STBN
  • Member since:
    11-08-2014

I see the Game start to Snowball, I'm getting a New Team.....     No If's, And's or But's, it's YOLO to the Front, Grab a Few Damage points and Off to the Garage....

 

You all can't be Counted in but for ine Thing, a Loss....      And I have no trouble helping Red out with a Win...



13Jake55 #12 Posted Aug 15 2019 - 08:55

    Captain

  • -Players-
  • 21968 battles
  • 1,079
  • Member since:
    02-02-2017
In the first place there is no single reason why blowouts happen. Skill has some effect but if your shells fly in the dirt due to randomness no amount of skill will help, your a purple playing like a tomato.  I think what weighs more on a battle than skill not the  tank class but the tanks themselves. I've seen where one side is top heavy with op tanks and the other side has butter tanks. Its a foregone conclusion that the op side wins because their op tanks are impervious to the other sides shells.

An_old_slow_guy #13 Posted Aug 15 2019 - 08:57

    Captain

  • Players
  • 63863 battles
  • 1,280
  • [ELVIS] ELVIS
  • Member since:
    08-10-2012
I've been in more battles lately than I can count where the skill in-balance was extreme.  I think 40% is pretty accurate, but it does vary based on time of day and day of week - in other words current server population makes a difference too. 

RIA1911 #14 Posted Aug 15 2019 - 09:09

    First lieutenant

  • Players
  • 6633 battles
  • 958
  • Member since:
    08-14-2011

View Postoldewolfe, on Aug 15 2019 - 02:36, said:

I see the Game start to Snowball, I'm getting a New Team.....     No If's, And's or But's, it's YOLO to the Front, Grab a Few Damage points and Off to the Garage....

 

You all can't be Counted in but for ine Thing, a Loss....      And I have no trouble helping Red out with a Win...


This right here is part of the imbalance that they battles can develop, impatience.  Seen a lot of battles that are on the edge and people don't have the patience to give it time and see what happens. 

The good players are going to do their thing .  The bad players just kinda take their gun out of the battle by rushing into the front or not being patient if the game has slowed down at the moment.

 

I've seen teams almost come back and even win when people don't waste themsevles.

 

I got fuss at by a player earlier tonight for not helping him, he rushed out to battle and got reduced by arty and others tanks down to 10% , before my slower tank could even get to where he went to. At this point i setup to defend, he was already lost. his impatience in a fast tank cost him.

 

 

 


Edited by RIA1911, Aug 15 2019 - 09:11.


Cpt_Zapp_Brannigan #15 Posted Aug 15 2019 - 11:17

    Captain

  • Players
  • 35569 battles
  • 1,350
  • Member since:
    07-07-2013
ah, that awkward moment when you are only down to HE rounds in your E25

Guido1212 #16 Posted Aug 15 2019 - 12:41

    Community Contributor

  • Players
  • 83047 battles
  • 9,093
  • [GFLC] GFLC
  • Member since:
    06-11-2011

View PostCpt_Zapp_Brannigan, on Aug 15 2019 - 10:17, said:

ah, that awkward moment when you are only down to HE rounds in your E25

 

Which should have been fixed way back in the garage before the big red battle button was pushed.  Which speaks to player skill and knowledge.  Which provides context to a great many things.



ColonelShakes #17 Posted Aug 15 2019 - 12:55

    Captain

  • -Players-
  • 16469 battles
  • 1,714
  • [CDN] CDN
  • Member since:
    01-17-2016

TLDR , same garbage logic as usual.  

 

Good news !

 

Random battles are to remain random.  WG has chimed in on that quite a few times.   

 

Your 'skill based MM' exists in a format called 'ranked battles'.  Not sure when that will be back.  

 

Please post from your main account as well.  Hiding behind an alt account to bypass forum moderation - not cool.  Bet you are permabanned.



UnturnedLeaf #18 Posted Aug 15 2019 - 13:39

    Staff sergeant

  • -Players-
  • 32402 battles
  • 311
  • Member since:
    03-05-2017

WG should eliminate player skill from being a factor in it's MM. Perhaps they can experiment with a more random approach and see if that improves gameplay and reduces the lopsidedness in their current format. 

 

Of course, that would negatively affect their bottom line. 



Jer1413 #19 Posted Aug 15 2019 - 13:46

    Captain

  • Players
  • 49043 battles
  • 1,642
  • [RR13] RR13
  • Member since:
    02-24-2013

View PostUnturnedLeaf, on Aug 15 2019 - 08:39, said:

WG should eliminate player skill from being a factor in it's MM. Perhaps they can experiment with a more random approach and see if that improves gameplay and reduces the lopsidedness in their current format. 

 

Of course, that would negatively affect their bottom line. 

 

No need to eliminate anything. MM is already blind to skill, about as random as you can get.



Kliphie #20 Posted Aug 15 2019 - 14:04

    Major

  • Players
  • 33652 battles
  • 6,030
  • [GFLC] GFLC
  • Member since:
    07-20-2012
15 - 7 is considered a blow out now?  




1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users