This has to stop

[MEANN]

Yesterday I raged after 4 blow outs in a row. I will post the Monitor below as evidence. I had no less then 6 red or bad players on my team every game. the contrasting team would usually only have 3 or less.

The match below was the end of my rope for the day. my team mate was the only good rated player. I can't believe that WG has no ability to at least evenly spread the very bad players on each team. 

to put context on the skill level. 

The Personal Rating is measure of skill in World of Warships. It measures performance in every warship and compares it to certain expected values. Differences between actual and expected values are used to calculate final Personal Rating value.

Basic assumptions:

• values in the range of 0 - 3000,
• colors and values similar to the well-known World of Tanks WN8,
• WoWS is more strategic game than WoT so win rate should influence Personal Rating,
• number of destroyed planes should be ignored - in WoWS AA is automated so no skill is needed to destroy planes. Even with CVs killing planes is very easy (click and forget)
• elements in order of importance: damage dealt, warship kills, win rate,

Formula:
Personal Rating calculations are based on expected and actual values of wins, damage dealt and warships kills. To get expected values use expected values for warships (preview, json) and multiply them by number of battles played.

 

Skill	Range
Bad	0 - 750
Below Average	750 - 1100
Average	1100 - 1350
Good	1350 - 1550
Very Good	1550 - 1750
Great	1750 - 2100
Unicum	2100 - 2450
Super Unicum	2450 - 9999

 How do you think this game went? 

 

 

 

 

 

 

  

[tcbaker777]

i think WG should ban that mod, it only leads to toxic behavior from people that think theyre too good to be put on a team with teammates that arent as good as them

you people need to get the hell off whatever high horse you think you have and start being decent human beings, that or you can just stop playing the game, because your little superiority complexes have no place here

[BingBongBrian]

 Maybe all those reds are actually secretly good players, ya never know.

Anyway, you get good teams, you get bad teams, that's just how it is. Happens to everyone, just deal with it (or find some unicum divmates to carry you).

[Telastyn]

I love that your team had 1 whole karma combined.

 

Maybe WG’s matchmaker is better than we give it credit for?

[MEANN]

Just now, tcbaker777 said:

i think WG should ban that mod, it only leads to toxic behavior from people that think theyre too good to be put on a team with teammates that arent as good as them

it isnt a mod, it is a stand alone program. however, you think that the match making was in any way balanced?

 

[MEANN]

Just now, Telastyn said:

I love that your team had 1 whole karma combined.

 

Maybe WG’s matchmaker is better than we give it credit for?

yeah they made sure the one team was not just a lil better but no chance at winning 

[Captain_Rawhide]

I suggest they take all the purple players and only let them play with other purple players.  If they have to wait ten minutes for a game, so be it.  Or else or you could just go play Clan battles, which is where you  should be if you want to be 'protected' from playing with the other 80% of the player base.

[YouSatInGum]

I have logged sample sizes exceeding 300 looking at this exact thing.

5% spread is bad and depending on the tier and absolute skill of the top team* will likely result in a 96 - 100% chance of loss for the bad team. 

 

 

*- Once 5% is reach to only teams to achieve an upset were either at mid tiers or the "better" was still only average or slightly better than.  If your team is average and the other team is good and unicum then there were zero upsets.  It supports the theory that as the WR goes up the skill level scales in an exponential fashion.

[Mykawa]

2 minutes ago, Captain_Rawhide said:

I suggest they take all the purple players and only let them play with other purple players.  If they have to wait ten minutes for a game, so be it.  Or else or you could just go play Clan battles, which is where you  should be if you want to be 'protected' from playing with the other 80% of the player base.

This would be nice but after a few hundred games the purple players would mostly be blue then they would be back with us lowlifes. 

[TheChattaHooligan]

22 minutes ago, BingBongBrian said:

 Maybe all those reds are actually secretly good players, ya never know.

Anyway, you get good teams, you get bad teams, that's just how it is. Happens to everyone, just deal with it (or find some unicum divmates to carry you).

^^^ This

[YouSatInGum]

22 minutes ago, BingBongBrian said:

Maybe all those reds are actually secretly good players, ya never know.

Anyway, you get good teams, you get bad teams, that's just how it is. Happens to everyone, just deal with it (or find some unicum divmates to carry you).

You argument would hold water in ranked... the problem is once you allow the divisions that you referenced, it give the player an tool to control distribution of skill among the players.  It is well known among players that have been here a while that many players in upper or unicum clans play most of their games in this fashion stacking the deck in their favor.... Of course this is to the detriment of players that can't or won't do this.

24 minutes ago, tcbaker777 said:

i think WG should ban that mod, it only leads to toxic behavior from people that think theyre too good to be put on a team with teammates that arent as good as them

you people need to get the hell off whatever high horse you think you have and start being decent human beings, that or you can just stop playing the game, because your little superiority complexes have no place here

That is only your interpretation of how he views this.  How do you know why he feels the way he does?

You are advocating banning a mod that maybe some players misuse, but other use in a positive way and punishing everybody.... isn't that the same selfish feeling you just railed against?

BTW  -  there are no real good ways of banning this mod without affect other stats viewing platforms... that would mess with wows.numbers especially

 

[Helstrem]

17 minutes ago, MEANN said:

Bunch of nonsense.

MM has practically nothing to do with blowouts.  No matter how many times you post this complaint that won't change and you'll still be barking up the wrong tree,

So, do you want to reduce blowouts or do you want to reduce the matches with a pretty forgone winner and loser?  The tools to do so are not the same.  A skill based MM can do the later, your MMM won't be able to tell you the likely outcome, but blowouts will still happen at nearly the same rate.  Or if blowouts are your pet peeve then the rules of the game need to be changed so that there are effective comeback mechanisms.  Or if you want both items then both changes need to happen.

[Soshi_Sone]

Although I haven't worked any statistics myself, and although I do see my fair share of blowouts (both sides), I still have a good number of nail biting games that are super intense and (even with a loss) are very enjoyable and fun.  And I HATE to lose.

And yes, it is the very nature of "randomness" that one can expect a series of bad (and good) runs.  The difference is in those matches where one has the skill to throw what should be part of a bad run (in a purely random sequence) to the good side...or the lack of skill that gives up a good run (in a purely random sequence) to the bad side.

I can pretty much say that EVERYONE here has had a run of four or more games where they were saddled with a poor team.

Also, be careful about divisions and clans.  A three player purple division GUARNATEES at least three unicum players on one side AND makes it more likely (greater than 50% chance) that a random solo player will find themselves on the OTHER team.  So divisioning with good players skews the statistics in favor of the division.  There is a reason WG limits divisions to three players.  It's the magic number that allows a team to skew the MM statistic in one's favor, but not so much that the randomness of the MM in general won't "randomly" place that division at a disadvantage that they can't overcome, even with skill. 

[YouSatInGum]

6 minutes ago, Mykawa said:

This would be nice but after a few hundred games the purple players would mostly be blue then they would be back with us lowlifes. 

Not exactly.  For players with a few thousand games it would take much much longer but still lets talk in the long run.

The real effect in the long run would be the better players would still have higher WR averages but relative to their skill they would just be clustered closer to the mean.  In other words the normal curve would be shifted up and the relative skill would be grow exponentially as you moved further away from the mean.

I do not advocate this type of matchmaking but there are alternatives that don't do this.  There are easy to implement formulas where adjustments are only made to the original group the MM made and players are swapped in the event of a large skill disparity such as the one the OP posted.  No increase in wait times happen this way, unicum division tampering is compensated for, and long run stats effects are minimized.  It's really a win win for the majority of players.  The only losers are those using divisions to manipulate skill distributions and those interested in having more blowout games (WG?).

[Torenico]

30 minutes ago, MEANN said:

Yesterday I raged...

 

There, found your problem, raging at a videogame. Just turn it off and play something else, really.

[MEANN]

1 minute ago, YouSatInGum said:

  There are easy to implement formulas where adjustments are only made to the original group the MM made and players are swapped in the event of a large skill disparity such as the one the OP posted.  No increase in wait times happen this way, unicum division tampering is compensated for, and long run stats effects are minimized.  It's really a win win for the majority of players.  The only losers are those using divisions to manipulate skill distributions and those interested in having more blowout games (WG?).

I agree, just balance the skill set. in my game posted above. just have the 11 bad players 5 on the side of the none unicum player and the 4 on the less skilled team

[nhf]

32 minutes ago, MEANN said:

Yesterday I raged after 4 blow outs in a row. I will post the Monitor below as evidence. I had no less then 6 red or bad players on my team every game. the contrasting team would usually only have 3 or less.

The match below was the end of my rope for the day. my team mate was the only good rated player. I can't believe that WG has no ability to at least evenly spread the very bad players on each team. 

to put context on the skill level. 

The Personal Rating is measure of skill in World of Warships. It measures performance in every warship and compares it to certain expected values. Differences between actual and expected values are used to calculate final Personal Rating value.

Basic assumptions:

• values in the range of 0 - 3000,
• colors and values similar to the well-known World of Tanks WN8,
• WoWS is more strategic game than WoT so win rate should influence Personal Rating,
• number of destroyed planes should be ignored - in WoWS AA is automated so no skill is needed to destroy planes. Even with CVs killing planes is very easy (click and forget)
• elements in order of importance: damage dealt, warship kills, win rate,

Formula:
Personal Rating calculations are based on expected and actual values of wins, damage dealt and warships kills. To get expected values use expected values for warships (preview, json) and multiply them by number of battles played.

 

Skill	Range
Bad	0 - 750
Below Average	750 - 1100
Average	1100 - 1350
Good	1350 - 1550
Very Good	1550 - 1750
Great	1750 - 2100
Unicum	2100 - 2450
Super Unicum	2450 - 9999

 How do you think this game went? 

 

 

 

 

 

 

  

WOW uses MM and RNG in a failed attempt to "fix" their broken meta.  Making a more profitable game is WOW's goal, not a better game.  Deal with it or find a new game.

[Ducky_shot]

2 minutes ago, MEANN said:

I agree, just balance the skill set. in my game posted above. just have the 11 bad players 5 on the side of the none unicum player and the 4 on the less skilled team

And then you'd be back here complaining about how MM gave all the good consumable ships to one team and not the other and asking how to play a DD against 5 radar and your team has 0.

[YouSatInGum]

7 minutes ago, Helstrem said:

MM has practically nothing to do with blowouts.  No matter how many times you post this complaint that won't change and you'll still be barking up the wrong tree,

So, do you want to reduce blowouts or do you want to reduce the matches with a pretty forgone winner and loser?  The tools to do so are not the same.  A skill based MM can do the later, your MMM won't be able to tell you the likely outcome, but blowouts will still happen at nearly the same rate.  Or if blowouts are your pet peeve then the rules of the game need to be changed so that there are effective comeback mechanisms.  Or if you want both items then both changes need to happen.

Incorrect.  Stats directly contradict your statement.

Teams with WR that exceed a differential of around 2.5% are statistically more likely to win to a significant level.  This signal because ever stronger as the differential is increased.

This isn't to say that blowouts don't happen with even teams....of course they do because of the game mechanics you are referring to.... that is one of the factors.  However, in decent percentage of games (1/4 to 1/3 depending on differential used), if the factor of skill difference appears, that factor has a high correlation with a win or loss.

[YouSatInGum]

11 minutes ago, Soshi_Sone said:

Although I haven't worked any statistics myself, and although I do see my fair share of blowouts (both sides), I still have a good number of nail biting games that are super intense and (even with a loss) are very enjoyable and fun.  And I HATE to lose.

And yes, it is the very nature of "randomness" that one can expect a series of bad (and good) runs.  The difference is in those matches where one has the skill to throw what should be part of a bad run (in a purely random sequence) to the good side...or the lack of skill that gives up a good run (in a purely random sequence) to the bad side.

I can pretty much say that EVERYONE here has had a run of four or more games where they were saddled with a poor team.

 

I would generally agree, and you appear to understand the issue here.... however....

12 minutes ago, Soshi_Sone said:

Also, be careful about divisions and clans.  A three player purple division GUARNATEES at least three unicum players on one side AND makes it more likely (greater than 50% chance) that a random solo player will find themselves on the OTHER team.  So divisioning with good players skews the statistics in favor of the division.  There is a reason WG limits divisions to three players.  It's the magic number that allows a team to skew the MM statistic in one's favor, but not so much that the randomness of the MM in general won't "randomly" place that division at a disadvantage that they can't overcome, even with skill. 

I would disagree about the magic 3 player number WG limits for divs.  I'd be more likely to claim the number was picked in a more arbitrary manner or somebody's personal feeling of what would be tolerated.  I have noticed that at particular times you are much more likely to run into these purple divs that skew skill distribution such a prime times, weekend prime time especially (when CB isn't on).  Middle of the day during the week then you are less likely.  I seriously doubt WG put any real statistical analysis into the magic 3 number. 

It would take a lot of number collection and crunching but it wouldn't surprise me if a solo player who usually plays during these prime times had their WR depressed in a significant way.

[Helstrem]

8 minutes ago, YouSatInGum said:

Incorrect.  Stats directly contradict your statement.

Teams with WR that exceed a differential of around 2.5% are statistically more likely to win to a significant level.  This signal because ever stronger as the differential is increased.

This isn't to say that blowouts don't happen with even teams....of course they do because of the game mechanics you are referring to.... that is one of the factors.  However, in decent percentage of games (1/4 to 1/3 depending on differential used), if the factor of skill difference appears, that factor has a high correlation with a win or loss.

That isn't what I said.

MMM can predict which team will win.  It does not predict the quality of that win.  Unicum vs unicum teams regularly experience blowouts.  Potato vs potato teams regularly experience blowouts.  This is because blowouts are not significantly affected by skill.  Blowouts are caused by the game rules not having any significant anti-snowball effect tools included.

A skill based MM would make each match's winner/loser outcome harder to predict.  It would have very little (~5% reduction) effect on the number of blowouts.  To combat blowouts the rules of the game need to be changed to include tools that effectively combat the snowball effect.

[_GreyBeard_]

Oh look, it must be that time of the month again.

[Skuggsja]

I agree the frustration. 

There a days that have games so bad it makes you never want to play again. 

How to fix it, I'm not sure. It would be nice if there was a way to not have horrendous days though. 

[Walrussy]

      While I do agree with your wish for slightly more balanced MM. However, at the same time nobody would want to play well. MM that waits for a player of equal skill could make certain players wait an unreasonable amount of time for a player of “equal skill” to hop into queue.

For example:

      My favorite ships for randoms are Marceau and Smaland. I’m ranked 7th and 4th on the  server in those ships respectively. If I wanted to play my favorite ships I’d have to wait for a equivalent super unicum player to be in queue. Top 10 players in most ships are mostly the same 30-40 people in a relative handful of clans. There could be nights where there are only a few of these players online. I’d honestly quit the game if I had to wait 5 mins+ for one of 30 or so people to “balance” the MM so I could play a ship I enjoy.

      Random battles are called random for a reason. The fun comes from making the most of the teams you have. So while every game having balanced teams is a nice wish having that wish is much more complicated than one would like to think.

[Asym_KS]

25 minutes ago, Helstrem said:

MM has practically nothing to do with blowouts.  No matter how many times you post this complaint that won't change and you'll still be barking up the wrong tree,

So, do you want to reduce blowouts or do you want to reduce the matches with a pretty forgone winner and loser?  The tools to do so are not the same.  A skill based MM can do the later, your MMM won't be able to tell you the likely outcome, but blowouts will still happen at nearly the same rate.  Or if blowouts are your pet peeve then the rules of the game need to be changed so that there are effective comeback mechanisms.  Or if you want both items then both changes need to happen.

I am 100% sure our host can not afford to even try "Combat Effectiveness (CE)" as a metric.....  First, the game isn't, in any way shape or form, skill based nor even a meritocracy !  Skill is some vague metric that somehow involves winning and losing........that is statistically not relevant; because, this is a PVP platform where other players choices create "your W/L" stats.....  Too much chaos in the system to create a metric.

And, if we ever even attempted a CE metric to use with a 1D MM............well, the screaming would start at minute one of deployment and the quitting the game at moment two......  CE is a seriously dangerous metric.  In the AI prototype games I've demo'd these past two years, a few combat based games have created CE as envisioned by the military in earlier SIMs......  The "testing staff" of experienced video game players usually quit.......  They can't handle the reality of what CE is and more importantly, what CE actually says about how really "GuD" you are.....  In fact, experienced soldiers and extreme video gamers experience a "skill dissonance" in these games......to the point of quitting.  And, in at least one of the Prototypes, an eight year old boy was an absolute death sentence to the enemy............why>?  Because he didn't come into the test with all sorts of "cultural baggage;" and, simply listened to what the game had to say about CE and, went out and "did it" as the game said.......  To the absolute frustration of the rest of the testing staff.......  CE is all about the metric's "expectations" based on what that "game culture" envisions as GuD........and that, makes CE a "gray metric" and not a hardened state of being....  CE won't work in this game no matter what our host can't afford.........