Jump to content


Influx of bad players. Numerical results.


  • Please log in to reply
228 replies to this topic

LesterQuaestor #1 Posted Jun 15 2016 - 22:14

    Captain

  • -Players-
  • 11970 battles
  • 1,308
  • Member since:
    03-26-2015

One of the most common complaints, in the forums and in game chat, is that there has been an influx of bad players. This is often blamed on 5X weekends, weekends themselves, the end of the school year, or other real world events. This supposed influx is often said to be the cause of poor results and/or a string of loses.

 

But what would a true influx of bad players really look like? This can be simulated numerically with a very simple model and statistics from WotLabs.

 

First I defined a metric that stands as a player rating. I defined this on the same scale as WN8, since we are all familiar with this scale, and stats are available for this metric. In the case of the simulation, we assume that this metric is an absolute measure of quality rather than an estimation based on the WN8 calculation.

 

Secondly, I populated a player pool with 20000 entries based on the WotLabs statistics (i.e. 6% of players are below 300, 14% between 300 and 450, etc). At the high end there are 18 players between 2450 and 2900 and 2 players greater than 2900. A second pool was exactly the same as this pool, but 5000 players with a rating below 500 were added.

 

Two teams were assembled by picking players randomly from this pool. 14 players where picked for "our" team, and 15 picked for the enemy. The sum of the allied and enemy ratings (including our contribution) are called Ra and Re. The chance to win was calculated by the formula below. It is said online that this is the formula used by XVM, but I don't know. It produces a nice normal distribution around 50% and is as good as any formula.

 

    CW= 0.5+ (Ra/(Ra+Re)-0.5)*1.5

 

A total of 1000 sets of teams were sampled for each data point. The results are shown below.

 

 

 

Our WN8 is shown on the x axis. Our average chance to win is the y axis. The blue line represents the normal pool and the red line the pool including an influx of bad players. The bars represent one standard deviation.

 

Two points need to be observed. First, and mainly, is that the effect of an influx of bad players is small compared to the normal variation expected from random teams. Given the large error bars, it is unlikely that anyone could tell whether there had been an influx of bad players. Secondly, there is an effect of such an influx, but it is an _increase_ in expected win rate, and that increase gets greater with your own skill.

 

 In point of fact, there is no mechanism where an influx of bad players can result in a decrease in your win rate, under the assumption of random MM.

 

  There are several limitations to this work. For one, it assumes a linear contribution based on skill, and that is probably not true. But it wouldn't change the result, just make it more pronounced for players with high skill. Secondly, the win chance calculation is not well established. But again, using a different formula would not give you a different result. The win chance is merely a statement that better teams tend to win more. Also, I did not include any possibility of platoons, but this would probably have little effect except for those darned purple pedder platoons.

 

  I believe this result as shown is inevitable. If there really were an influx of bad players, all it would mean is that there are 15 chances for one of them to be on the enemy team, but only 14 chances to be on your team. Thus any such influx is a benefit to any player who is not absolutely terrible themselves.

 

TLDR: Stop blaming others.

 



Strike_Witch_Tomoko #2 Posted Jun 15 2016 - 22:16

    Major

  • Players
  • 15574 battles
  • 12,568
  • Member since:
    05-04-2013

bad players....

good players....

 

to me they both mean 1 thing.  players.   for me to kill >=D

 

 



Fbomb #3 Posted Jun 15 2016 - 22:27

    Captain

  • Players
  • 29927 battles
  • 1,800
  • Member since:
    05-10-2011
I know where all those dudes are. I've had five consecutive sub 50% sessions. If it's a tier X match and my T71 is the highest wn8 on the team, a loss cometh

EvilSeraph #4 Posted Jun 15 2016 - 22:31

    Corporal

  • Players
  • 6 battles
  • 61
  • [502BW] 502BW
  • Member since:
    06-27-2013

assuming mm is random



TheGhostCat #5 Posted Jun 15 2016 - 22:33

    Major

  • Players
  • 20380 battles
  • 6,678
  • [LAMDA] LAMDA
  • Member since:
    05-13-2011

This doesn't get around the fact that more 45% ers on your team greatly increases your changes of losings: These are people who CONSTANTLY make harmful decisions for their team.

 

The baddies are real.



Kekistani #6 Posted Jun 15 2016 - 22:35

    Captain

  • -Players-
  • 26365 battles
  • 1,748
  • Member since:
    02-24-2015
I did a great thing for myself weeks ago: deactivated XVM during battles. Not only the team tabs are not all horribly colorful anymore, I'm also expecting everyone to be on my "level" or better. I've done the mistake of not giving a crap about players with bad WN8 during the battle before and it turned out they were having a good battle and ended up causing more damage to me than others because I wasn't paying enough attention to them.

Fbomb #7 Posted Jun 15 2016 - 22:35

    Captain

  • Players
  • 29927 battles
  • 1,800
  • Member since:
    05-10-2011

View PostEvilSeraph, on Jun 15 2016 - 16:31, said:

assuming mm is random

 

Oooohhhhhh.....you can't say that

Cutthroatlemur #8 Posted Jun 15 2016 - 22:36

    Major

  • Players
  • 19675 battles
  • 10,477
  • Member since:
    08-24-2011
Excellent work as always LesterQuaestor,thank you and +1

Fhone #9 Posted Jun 15 2016 - 22:37

    Sergeant

  • -Players-
  • 3328 battles
  • 222
  • Member since:
    03-13-2016
Too many 40%ers in pubs

LesterQuaestor #10 Posted Jun 15 2016 - 22:42

    Captain

  • -Players-
  • 11970 battles
  • 1,308
  • Member since:
    03-26-2015

Block Quote

 This doesn't get around the fact that more 45% ers on your team greatly increases your changes of losings: These are people who CONSTANTLY make harmful decisions for their team.

 

  But the point is that there can be 15 of them on the enemy team, but only 14 on _your_ team. Unless you are a 45%er.

 



Darth_Tator #11 Posted Jun 15 2016 - 22:43

    First lieutenant

  • Beta Testers
  • 20866 battles
  • 768
  • [FGTVE] FGTVE
  • Member since:
    07-26-2010
A lot of work, but has doesn't look anything like what we all experience out there.  Theory is fine, postulating is fine...engagement with the baddie changes everything.

Scorpiany #12 Posted Jun 15 2016 - 22:47

    Major

  • Game Knowledge Expert
  • 36995 battles
  • 13,051
  • [N1NJA] N1NJA
  • Member since:
    06-27-2013

Many players have a tendency to complain about summer resulting in an insult in bad players, thus somehow lowering their ability to carry battles and win. This is extremely ironic when you realize that in general, even if there is an influx of bad players, there will be an equal increase in bad players (or decrease of skill) on both teams. This means that whilst your own team may be a bit worse, so is the enemy team - Which means it is easier for you to defeat the enemy, with less interference from your own team in dealing damage and carrying - Thus if the general skill level declines, you're more likely to have even better games than before, as well as winning more battles.

 

An example of this effect, is the same as comparing a drop in Tier and its affect on your teams, battle performances and Win Rates. The lower in Tier you go, the worse the skill level on average is which you will encounter on both teams. As such, at lower Tiers, it is easier to get higher damage results in comparison to your tank's HP, more epic medals, higher average XP values, higher Win Rates, etc.

 

If you have an average Tier of 8, and are a competent player, if there is a legitimate decrease in average skill during the summer months, then your stats in terms of Win Rate and Tier-for-Tier battle performance should improve in the same way they would as though you dropped down to an average Tier of 7.



PD4_Banana_Fish #13 Posted Jun 15 2016 - 23:44

    Captain

  • Players
  • 21344 battles
  • 1,520
  • Member since:
    06-22-2013
The difference between a good player and a bad player is all in your perspective.  The actual difference is pretty small.  My advice to the OP: find something else to cry about.

TheGhostCat #14 Posted Jun 15 2016 - 23:47

    Major

  • Players
  • 20380 battles
  • 6,678
  • [LAMDA] LAMDA
  • Member since:
    05-13-2011

View PostLesterQuaestor, on Jun 15 2016 - 22:42, said:

 

  But the point is that there can be 15 of them on the enemy team, but only 14 on _your_ team. Unless you are a 45%er.

 

 

This haven't seen that battle...just the one where I get 10-12 on my team, and 2 on the enemy...

More to the point though - if you have 2 of them in your top 7 tanks, and you enemy doesn't - you've probably lost the match.



24_inch_pythons #15 Posted Jun 15 2016 - 23:55

    Major

  • Players
  • 29929 battles
  • 5,690
  • [HSOLO] HSOLO
  • Member since:
    03-28-2012

View PostBulletFarmer, on Jun 15 2016 - 16:44, said:

The difference between a good player and a bad player is all in your perspective.  The actual difference is pretty small.  My advice to the OP: find something else to cry about.

 

Ah ha ha, omg, no.

LesterQuaestor #16 Posted Jun 16 2016 - 00:01

    Captain

  • -Players-
  • 11970 battles
  • 1,308
  • Member since:
    03-26-2015

Block Quote

 but has doesn't look anything like what we all experience out there.

 

  It doesn't look _anything_ like what you observe? Pardon me, but can I ask what it is you observe?

 

In fact, what the results show is inevitable. If a baddie throws a game, it is more likely that he is on the other team than on your team. By a ratio of 15:14. That's not a lot, but there is no means for it to be the other way round.

 



vbluguitar #17 Posted Jun 16 2016 - 00:04

    Captain

  • Players
  • 18799 battles
  • 1,363
  • [CS7AO] CS7AO
  • Member since:
    01-21-2012

View PostBulletFarmer, on Jun 15 2016 - 17:44, said:

The difference between a good player and a bad player is all in your perspective.  The actual difference is pretty small.  My advice to the OP: find something else to cry about.

 

You obviously didn't read the original post.

your_Master #18 Posted Jun 16 2016 - 00:06

    First lieutenant

  • Players
  • 15525 battles
  • 743
  • Member since:
    07-22-2013

View PostBulletFarmer, on Jun 15 2016 - 23:44, said:

The difference between a good player and a bad player is all in your perspective.  The actual difference is pretty small.  My advice to the OP: find something else to cry about.

 

this quote makes sense when you realize this dude is a 41%er

 

anyway yeah summer teams are no worse than normal teams. if you are losing and you need someone to blame go find a mirror. 



vbluguitar #19 Posted Jun 16 2016 - 00:17

    Captain

  • Players
  • 18799 battles
  • 1,363
  • [CS7AO] CS7AO
  • Member since:
    01-21-2012

View Postyour_Master, on Jun 15 2016 - 18:06, said:

 

this quote makes sense when you realize this dude is a 41%er

 

anyway yeah summer teams are no worse than normal teams. if you are losing and you need someone to blame go find a mirror. 

 

I have carried a 57.7 percent win rate over the last 700 battles - Mostly due to my amazing platoon mate, Black Chicken ;)

Sunserrae #20 Posted Jun 16 2016 - 04:19

    Major

  • -Players-
  • 15962 battles
  • 2,486
  • Member since:
    07-19-2014

View PostBulletFarmer, on Jun 15 2016 - 22:44, said:

The difference between a good player and a bad player is all in your perspective.  The actual difference is pretty small.  My advice to the OP: find something else to cry about.

 






1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users