1967 replies to this topic

[eteam]

Nixeldon, on Mar 14 2018 - 15:20, said:

I know. You can't seem to explain what it would do, you just "believe" it would be an obvious improvement.  EJC said that blowouts were obviously what was meant by crap matches none of you objected to his definition. Not one person has explained what a crap match is except for EJC. Even though his assertions are 100% incorrect, at least he offered that much.

 

Stop using the straw man allegation. You clearly don't know what it means.

 

It is precisely the point! Everyone, except for you, is claiming that SBMM will make the game play better. Short matches and blowouts are frequently recurring complaints that are attributed to a dislike of a random mm and those exact attributes are what are claimed to be alleviated by skill balancing, except by you. 

 

No. We all don't know what you mean by this. I spent the past few days reviewing my last 500 matches and installed XVM for reference. None of them fit your 6 tanks/4k hp within two minutes. I had a total of two matches that ended under 4 minutes. Using WN8 as the prediction(slightly better than win rate alone), we lost a 61% CTW to being capped at 11:34 and won a 54% CTW by 15-3 at 11:12.

 

A match is over when it is over and the rest of what you stated is garbage.

 

I never implied the MM picked any or all winners and losers. That is the domain of you, Stiffy, Shadora, eteam, et al.

 

I have tracked matches, I have seen various data types and none of it supports what you are claiming.

 

You first have to understand what is fact. Your tracking did nothing but demonstrate the accuracy of the Chance To Win calculation. Instead of attempting to understand why the prediction was off by almost 6% you jumped to a conclusion you hadn't tested. 6% is quite significant as far as player win rate distribution is concerned.

 

1) XVM was underrating your team weight based on your account when you started tracking. You had a lower xTE rating in the M6 and a lower win rate. You had a low average tier and a lower account WN8. Assuming your performance normalized with enough remaining battles to allow for random factors and stat correction, then XVM CTW was very inaccurate in the range it should have been most accurate.

 

Yes. I have a record and I have screen shots. There is no correlation in my data to anything you have ever posted.

 

​Maybe you should read individual posts and not lump everyone together with your assumptions and pat responses.

You know and have been told that SBMM would balance skill a very simple concept really.  If you don`t understand that can`t help you.

And for some of us it would be an obvious improvement.

As for crap matches they are the ones where the skill balance is very skewed and as a result the MM has more influence on the outcome.  (As demonstrated by Neatoman`s graph)

 

[NeatoMan]

EmperorJuliusCaesar, on Mar 14 2018 - 10:54, said:

I tracked them via pen and paper because that's easier than tabbing out of the game and risk a client crash.  My results were between 41-42%, slightly more than 40%.

He's shown you his data and it clearly shows that 40% are crap pointless battles that are driving people away from the game.

How did you ascertain "crap" and "pointless"?   Without knowing your methodology or logic I can only guess as to what you mean by "crap". Without being able to compare our data we might as well be comparing apples and oranges.   I don't platoon, seal club, gold spam, or pad my stats.  I simply play the game, and I don't seem to get as many crap games as you do.

 

Block Quote

As for crap matches they are the ones where the skill balance is very skewed and as a result the MM has more influence on the outcome.  (As demonstrated by Neatoman`s graph)

 The problem is everyone who picks that 40% value has absolutely no idea what kind of team imbalance leads to those win chances.  "very skewed" is not a 60% win chance, or 65% or even 70%.   Every example given of heavily skewed teams are examples of games that are 80% win chances or worse.  I agree, those are stacked teams, but those are also rare.

Edited by NeatoMan, Mar 14 2018 - 17:24.

[Nixeldon]

eteam, on Mar 14 2018 - 11:07, said:

​Maybe you should read individual posts and not lump everyone together with your assumptions and pat responses.

You know and have been told that SBMM would balance skill a very simple concept really.  If you don`t understand that can`t help you.

And for some of us it would be an obvious improvement.

As for crap matches they are the ones where the skill balance is very skewed and as a result the MM has more influence on the outcome.  (As demonstrated by Neatoman`s graph)

 

What assumptions did I make, especially with no basis?

 

SBMM, by definition, is supposed to balance skill. Presenting tautologies doesn't make it more apparent.

 

You claim it is an obvious improvement solely because numbers/colors match, same as budha. The colors are basically arbitrary and you can change them to any color you want. You don't need SBMM to do that. How would it be an obvious improvement? That is the core question. You keep talking in circles like TrueBudha.

 

His graph demonstrated that HE had an influence on his matches.

 

What do you think SBMM does!?! It doesn't remove MM influence, it removes player influence. SBMM increases MM influence by default.

 

 

[eteam]

NeatoMan, on Mar 14 2018 - 16:18, said:

How did you ascertain "crap" and "pointless"?   Without knowing your methodology or logic I can only guess as to what you mean by "crap". Without being able to compare our data we might as well be comparing apples and oranges.   I don't platoon, seal club, gold spam, or pad my stats.  I simply play the game, and I don't seem to get as many crap games as you do.

 

 The problem is everyone who picks that 40% value has absolutely no idea what kind of team imbalance leads to those win chances.  "very skewed" is not a 60% win chance, or 65% or even 70%.   Every example given of heavily skewed teams are examples of games that are 80% win chances or worse.  I agree, those are stacked teams, but those are also rare.

This provided by Nixeldon's post.  Whether or not I interpret correctly or not is another matter.

Average Win Rate Smoothing

 

The reason we balance average win rate is due to statistics gathered on the difference win rate makes in the outcome of a match. We discovered that even a small advantage of 2 or 3 percentage points resulted in much higher probability of winning. The below chart is based on the result of over 30,000 matches before the smoothing algorithm was added.

 

 

Prior to implementing the win rate smoothing, only 40% of matches were within +/- 1 percentage point, 30% were +/- 2 percentage points, and 30% +/- 3 or more. After implementing the smoothing algorithm, over 99% of matches are now within +/- 1 percentage point (based on a sample of over 140,000 matches).

[Nixeldon]

eteam, on Mar 14 2018 - 11:44, said:

This provided by Nixeldon's post.  Whether or not I interpret correctly or not is another matter.

Average Win Rate Smoothing

None of us are arguing that better players don't win more often. That alone says nothing about match quality.

 

You conveniently left out this statement from their developers:

 

"The meaning of this distribution is open to interpretation, but games typically have around 7-9 survivors while 10+ survivors (28% of matches) is actually a normal result of the game. This supports the "snowball effect" theory that once a team has a slight advantage -both in terms of numbers and available HP- they have more opportunities to quickly overwhelm the opposing team."

 

And this:

 

"It could very well be that a purely random distribution of players across teams (while respecting tier and vehicle) is simply the fairest way to go about team creation." 

 

"I do not think you would notice a drastic difference even if we disabled the influence skill has on Matchmaking which is certainly a point in favor of removing the mechanic all together. The goal was to try and eliminate that underlying random element of which players landed randomly on which team and I believe we accomplished that with the system. But if succeeding in that goal isn't providing an overall better experience for everyone, then the system doesn't need to stay."

[NeatoMan]

eteam, on Mar 14 2018 - 11:44, said:

This provided by Nixeldon's post.  Whether or not I interpret correctly or not is another matter.

~~~snip~~~

Those are from AW and are based on differences in average team win rates.   You have to get out to +/-3% in order to reach ~70/30% win probability (2:1 odds).  In order to get that kind of difference you'd need five dark green players on one team, and five red players on the other (with the rest being average yellow players).  +/-2% (~62/38% win chance) is three dark greens on one team and three reds on another (the rest yellow).  +/-1% (~56/44% win chance) is two greens on one team and two reds on another (the rest yellow).  Neither is close to all green/blue vs all red/orange.  Plus, If I don't see any change in game play out to 2:1 odds, then I see no need to rebalance those games.

[Rolling_Pig]

These fn idiots at WOT wont do it.  I quit 5 years ago because of the bs mm not being based on skill.  I have not won one time today and all blow-out losses.   Ready to quit again and everyone in my clan 5000-6500 plus ratings for most say the same thing... getting to the point where its not fun and worthless to play anymore when a team has no chance to win....  Server was down to12000 the other day.  When I used to play it never went below 35000.... when its down to 2000 then maybe they will do something but it will be too late.....

[eteam]

Nixeldon, on Mar 14 2018 - 17:01, said:

None of us are arguing that better players don't win more often. That alone says nothing about match quality.

 

You conveniently left out this statement from their developers:

 

"The meaning of this distribution is open to interpretation, but games typically have around 7-9 survivors while 10+ survivors (28% of matches) is actually a normal result of the game. This supports the "snowball effect" theory that once a team has a slight advantage -both in terms of numbers and available HP- they have more opportunities to quickly overwhelm the opposing team."

 

And this:

 

"It could very well be that a purely random distribution of players across teams (while respecting tier and vehicle) is simply the fairest way to go about team creation." 

 

"I do not think you would notice a drastic difference even if we disabled the influence skill has on Matchmaking which is certainly a point in favor of removing the mechanic all together. The goal was to try and eliminate that underlying random element of which players landed randomly on which team and I believe we accomplished that with the system. But if succeeding in that goal isn't providing an overall better experience for everyone, then the system doesn't need to stay."

 

​The point that I am trying to make here is that the random MM influences the probability of winning.  You always suggest we can't explain why we want some form of skill balancing.  Well there it is.

[Nixeldon]

eteam, on Mar 14 2018 - 12:13, said:

​The point that I am trying to make here is that the random MM influences the probability of winning.  You always suggest we can't explain why we want some form of skill balancing.  Well there it is.

You are putting your cart before the horse. This isn't a reason "we want". Several are saying win rates won't be affected, which is a contradictory statement. Why didn't you say from the start you wanted to equalize win rates? SBMM will easily contract win rates.

 

 

 

 

[eteam]

NeatoMan, on Mar 14 2018 - 17:03, said:

Those are from AW and are based on differences in average team win rates.   You have to get out to +/-3% in order to reach ~70/30% win probability (2:1 odds).  In order to get that kind of difference you'd need five dark green players on one team, and five red players on the other (with the rest being average yellow players).  +/-2% (~62/38% win chance) is three dark greens on one team and three reds on another (the rest yellow).  +/-1% (~56/44% win chance) is two greens on one team and two reds on another (the rest yellow).  Neither is close to all green/blue vs all red/orange.  Plus, If I don't see any change in game play out to 2:1 odds, then I see no need to rebalance those games.

 

​Thanks for your response.  So at the risk of asking you to maybe repeat what you have already posted.  In your experience what percentage of games played would fall into the 2:1 odds category?

[eteam]

Nixeldon, on Mar 14 2018 - 17:20, said:

You are putting your cart before the horse. This isn't a reason "we want". Several are saying win rates won't be affected, which is a contradictory statement. Why didn't you say from the start you wanted to equalize win rates? SBMM will easily contract win rates.

 

 

 

 

 

I understand that SBMM will contract win rates.

That is not the motivation for wanting some form of skill balancing.  My motivation for some form of skill balancing is to simply eliminate those games that have a huge skill disparity. In my opinion these games result in an increase in bots, yolo's, drownings, and camping etc.  These for me are some to the attributes that lessen the game experience and I believe have an influence on players leaving the game.

 

 

[spud_tuber]

eteam, on Mar 14 2018 - 11:44, said:

 

I understand that SBMM will contract win rates.

That is not the motivation for wanting some form of skill balancing.  My motivation for some form of skill balancing is to simply eliminate those games that have a huge skill disparity. In my opinion these games result in an increase in bots, yolo's, drownings, and camping etc.  These for me are some to the attributes that lessen the game experience and I believe have an influence on players leaving the game.

 

 

While I would argue that skill baalnced MM will increase botting and yolo because the increased win rates will make those more effective means of grinding.  I'm less sure about drownings, but camping is a product of single death, imo, and likely won't be effected much by Skill balanced MM.

[Staz211]

EmperorJuliusCaesar, on Mar 14 2018 - 08:35, said:

 

If the teams are balanced, and you do 0 damage, you have a much lower chance of winning that match.  If however the match is, like 40% of our current matches, very imbalanced to the point it's over before it starts....then it's still very likely you'll be carried to the win.  Only an idiot couldn't understand that. 

Yeah, because 47% of your entire populace being on govt aid is a good thing.  Great job with "job creation" and the "land of opportunity".  Great country, no wonder so many with the means to live elsewhere.....do...

 

EmperorJuliusCaesar, on Mar 14 2018 - 08:41, said:

 

40% of battles(math has been shown) being over before they start is in no way fair, that's 40% of the time playing wasted. 

Unicoms OPENLY ADMIT in their youtube videos to gaming the current system, so no, that is not fair. 

It IS about stopping the loss of players, especially newer ones(would stop seal clubbing to a great extent), even WG has acknowledged it and said they will keep changing MM until they get it right.  Nice theory you have there at the end, too bad it's all wrong and extremely selfish, you'd rather keep losing players until there is no NA server rather than fight fair battles, we get it.

 

da_Rock002, on Mar 14 2018 - 08:59, said:

 

 

Completely random? 

Do you have a record of your scores?   Keep one for a while and let us know how often you are on the losing side of 15-7 and worse outcomes.   Check out how often you're on the winning side versus the losing side.   You'll see WoT random at it's finest.       You might notice that over ~95% of the 15-7 and worse scores have one thing in common.   One team is half full of players with similar stats to you.   The other team has maybe one or two.   You'll also see the losing side is chock full of newbies.   Check out how often you see a blowout loss.        

 

If you think about it, that isn't completely random.    It's the signature of WoT/WG's daily ration of battles.   40% of them suck.   And not a bit of 'completely' in that random.

 

 The MM is broken.   Why?    Who cares why?    Oh wait, the people who benefit from it care big time for an obvious reason.   They are being rewarded with wins they think they deserve.   And don't want the golden goose to quit laying their golden eggs.    So they can 'earn' some more stature in broken matchups.

 

Talk about 'something you did not earn'......    Early on in your career, you DID EARN your WR.   And congratulations on that.   Apparently WoT hasn't handed out the crap of today from the beginning.   But TODAY it feeds newbies an entirely different diet it feeds the blues.    When the 30 players in a match are about 3/4 tomatoes and newbies, and 1/2 of them are on one side, that 1/2 is going to see mostly bad-broken matchs.   The 1/4 is going to see GOOD-broken for them.       So it makes sense you demand there is nothing wrong.  

And what is really happening here is blues demanding that nothing needs fixing...  that random really does work the way WoT/WG spits it out...  that the people reporting the breakage want something given to them.   (They really want real random given to everyone.)

 

What they want is for WoT to stop TAKING from new players and giving it to the group that needs it the least.

 

This might be the most hopelessly daft group of people I have ever conversed with. You honestly cannot grasp the simple, fundamental difference between equality of opportunity and equality of outcomes, and the basic nature of a bell curve. 

 

I can only hope that you guys aren't fully grown men. 

[Nixeldon]

eteam, on Mar 14 2018 - 12:44, said:

I understand that SBMM will contract win rates.

That is not the motivation for wanting some form of skill balancing.  My motivation for some form of skill balancing is to simply eliminate those games that have a huge skill disparity. In my opinion these games result in an increase in bots, yolo's, drownings, and camping etc.  These for me are some to the attributes that lessen the game experience and I believe have an influence on players leaving the game.

But several others claim it won't affect win rates, which is an issue. If not for the emphasis on winning matches, the rare moments of team cooperation would be nonexistent in WOT.

 

Some bot programs are better than the average player, many players grief, etc. and do so regardless of their teams. There is no evidence that would change. All SBMM would do is incentivize more of that same behavior. Wouldn't a more logical approach be to address the pernicious teammates?

 

The key aspect to consider from AW is they attempted a tighter MM than is being proposed in these threads, even addressing platoons before they scrapped it altogether and their SIMM made little difference.  Aside from super platoons running MBTs(terrible balancing), the win rate smoothing was successful and that was all that was accomplished.

 

 

 

 

 

 

 

[Jer1413]

EmperorJuliusCaesar, on Mar 14 2018 - 15:54, said:

 

I tracked them via pen and paper because that's easier than tabbing out of the game and risk a client crash.  My results were between 41-42%, slightly more than 40%.

He's shown you his data and it clearly shows that 40% are crap pointless battles that are driving people away from the game. 

The CRAP slot machine MM is bad for everyone except those that admit to gaming the system to inflate their stats.  When it's admitted to, that says a lot, when people play lower tiers(seal clubbing) to pad their stats, they know what they are doing, they are gaming the system to inflate their stats and look "good" at a 13+ GAME.  It's no wonder why some don't want it change, they don't want a fair fight.

 

You do realize that your standard bearer for the SBMM cause is about the biggest seal clubber/stat padder in the game, right?

[NeatoMan]

eteam, on Mar 14 2018 - 12:23, said:

​Thanks for your response.  So at the risk of asking you to maybe repeat what you have already posted.  In your experience what percentage of games played would fall into the 2:1 odds category?

When I adjust for tier (which actually helps the SBMM proponents, by moving more games out to higher/lower win chance).  I get ~25% of all my games are greater than 2:1 odds (beyond 70/30% win chance).  However, my skill pushes a few more games beyond 2:1 odds when the curve is shifted higher.  For the average player it's ~20% of games that fall outside 2:1 odds.

 

I use 30/70% win chance because that's the range beyond which I start to see faster games and more blowouts.

Edited by NeatoMan, Mar 15 2018 - 00:16.

[eteam]

NeatoMan, on Mar 14 2018 - 23:15, said:

When I adjust for tier (which actually helps the SBMM proponents, by moving more games out to higher/lower win chance).  I get ~25% of all my games are greater than 2:1 odds (beyond 70/30% win chance).  However, my skill pushes a few more games beyond 2:1 odds when the curve is shifted higher.  For the average player it's ~20% of games that fall outside 2:1 odds.

 

I use 30/70% win chance because that's the range beyond which I start to see faster games and more blowouts.

 

​Thanks for your constructive post.

[eteam]

Nixeldon, on Mar 14 2018 - 17:01, said:

None of us are arguing that better players don't win more often. That alone says nothing about match quality.

 

You conveniently left out this statement from their developers:

 

"The meaning of this distribution is open to interpretation, but games typically have around 7-9 survivors while 10+ survivors (28% of matches) is actually a normal result of the game. This supports the "snowball effect" theory that once a team has a slight advantage -both in terms of numbers and available HP- they have more opportunities to quickly overwhelm the opposing team."

 

And this:

 

"It could very well be that a purely random distribution of players across teams (while respecting tier and vehicle) is simply the fairest way to go about team creation." 

 

"I do not think you would notice a drastic difference even if we disabled the influence skill has on Matchmaking which is certainly a point in favor of removing the mechanic all together. The goal was to try and eliminate that underlying random element of which players landed randomly on which team and I believe we accomplished that with the system. But if succeeding in that goal isn't providing an overall better experience for everyone, then the system doesn't need to stay."

 

​Here are the links to the WT developer articles on SBMM provided by Nixeldon.

Fair comment above when I did not include the whole link.

Well worth the read if you didn`t the first time round.

https://aw.my.com/us/news/general/developer-diary-matchmaking-20

https://aw.my.com/us/news/general/developer-diary-matchmaker-and-skill

https://aw.my.com/us/news/general/myths-about-matchmaker-part-2-losing-streaks

[AMSRocker]

What chapter are we on in writing this book now?

Page 91 and counting...

[SoTrue]

Staz211, on Mar 13 2018 - 21:51, said:

 

You have proven nothing here. All you've done is state that sometimes you do well and lose, and other times you do poorly and win. That's not match rigging, that's the nature of a random system where each match is the sum of 30 independent, individual inputs. I have had 6k damage losses and 0 damage wins. Were those matches rigged? No, not at all. Were those matches decided before the match even began? No, not at all. One set of 15 players, who were randomly thrown together, outplayed another random set of 15 players who were randomly thrown together. If you're a good player, your input to the match has greater impacts than that if a bad player. .

 

 

Your SBMM idea would, as a matter of fact, force everyone towards a 50% WR. Without even getting deep into the math behind it, better players can't tend more towards 60 because the competition pool is not random, so their abilities will never be matched against the rest of the player base as a general whole. Bad players will not tend towards 40 because, the worse they play, the better the good players they will get matched with. This drags the bad player towards 50%, and weighs the good player down towards 50%. The better you play, the worse your teammates. The worse you play, the better your teammates. Will everyone have exactly a 50% WR? No, but the range of possible WRs will be much more narrow.

 

This entire thread is a microcosm of equality of opportunity Vs equality of outcome, perfectly reflective of debates that exist today and throughout human history. As per usual, the equality of outcome camp fails to see the second, third, and fourth order effects of their proposed policies that will attempt to artificially create "fairness." They fail to see that these policies will, inevitably, break and collapse the entire system.

 

That, and they resort to insults and baseless, dismissive assertions of being correct when confronted with a counter argument. It's hilariously reflective of larger society.

 

wow, shown obvious proof and still denies it.  Sad.  I totally proved it's possible.  Where is your data? Oh that's right, all you have is your opinion.  Which means nothing.

spud_tuber, on Mar 13 2018 - 22:29, said:

You know what immediately jumps out at me about that chart?

 

Average chance to win: 50.68%

Actual win rate:    56.30%

 

Win chance calculations are therefore inaccurate, and the 40-60 chance on that chart may not be representative of actual skill balanced MM.  Do you still have the raw data?  If so, try manually adjusting  all the win chances upward such that the average win chance becomes approximately the same as actual win rate and redo the above chart.

 

Quick and dirty adjustment would be to simply add the difference between the two averages to every win chance value.  More accurate methods I'm sure exist as well, as a doubt the battles with extreme win chances on either end actually need that large of an adjustment, but I can't think of a reasonable one off the top of my head.

 

 

Wow, the actual data does not confirm my opinion, so he asks 'can you adjust the data?'.  Sad.  That's the problem with both these guys.  Actual data proves sbmm WILL NOT force a player to 50% - AND THEY STILL TRY TO SAY IT WILL.  Sad.

sparango, on Mar 14 2018 - 00:32, said:

What happens to platoons that have lets say players of various skill levels. Do you just take the average skill or does the guy with the lowest skill have to step it up against statically better opponents?

 

Toons are something that would need to be worked out.  Making them a 'block' and taking the average skill is one way.  Another would be to take the highest player, and make the whole toon have that skill level.  Or you could still average them individually.  It would have to be tested to see which ends up being the best way to factor them into balanced teams.

NeatoMan, on Mar 14 2018 - 02:52, said:

Your definition of crap is totally subjective.  Why should I believe you when I can see for myself just how bad they really are?  Especially when you've never really showed an understanding of how XVM works.

 

Your comment above about a red player not being more likely to have more red players on his team proves this.  It's why they shift their curve lower.  He guarantees one extra red player on his team every game

 

And you still don't get it.  That there is an imbalance is the only reason any player can 'shift' anything.  If every battle was between balanced teams, there would be no shift.  Every battle would be fair.  That there are imbalanced teams is a loophole that allows players to shift their bell curve.  It shouldn't be possible.