Jump to content


Tier10 ruined


  • Please log in to reply
28 replies to this topic

ChristOfTheAbyss #21 Posted 18 September 2019 - 05:07 PM

    Warrant Officer

  • Player
  • 251 battles
  • 893
  • Member since:
    07-22-2019

View PostDorander, on 18 September 2019 - 04:59 PM, said:

 

We're all 6.67% of a team, or half that of influencing the overall battle outcome. This means that if everything would be even, we'd expect an inactive player to lose 3.33% of their winrate compared to average winrate, which being 49% means we'd expect an idle player to end up around 45.66%.

 

Problem is everything isn't even, not having an active gun on your side or even a person blocking a flank by sheer presence means the enemy has more opportunities and better angles to fire from and will wipe out the idle player's team faster than the other way around, losing more guns and thus losing even more tanks more quickly, aka the "snowball effect". Now the snowball effect is difficult to quantify but at the very least we'd expect idle players to end up lower than 45.66%, not higher. Your own account matches that theory. 

 

I'd offer the explanation that most normal accounts who have lower winrates than yours have a normal activity profile which means they progress in tiers. At tier 1, we don't expect to see a lot of skill, mostly new players with "random" behaviour by which I mean lacking the more tactical behaviour of experienced players. The ability to capitalize on enemy mistakes and the ensuing advantages (such as outgunning them) is expected to be higher amongst experienced players, and subsequently, higher among the higher tiers. The most likely explanation for you being "not THAT low" is that you never ventured beyond tier 1 on this account.

 

Welp, didnt think it that far. Explains a bit. Aint venturing to higher tiers though. Tests like this are fine at t1 where no-one really plays, not so fine at t8 or something where people really try.



jabster #22 Posted 18 September 2019 - 05:30 PM

    Field Marshal

  • Beta Tester
  • 12823 battles
  • 26,964
  • [WSAT] WSAT
  • Member since:
    12-30-2010

View PostDorander, on 18 September 2019 - 03:59 PM, said:

 

We're all 6.67% of a team, or half that of influencing the overall battle outcome. This means that if everything would be even, we'd expect an inactive player to lose 3.33% of their winrate compared to average winrate, which being 49% means we'd expect an idle player to end up around 45.66%.

 

Problem is everything isn't even, not having an active gun on your side or even a person blocking a flank by sheer presence means the enemy has more opportunities and better angles to fire from and will wipe out the idle player's team faster than the other way around, losing more guns and thus losing even more tanks more quickly, aka the "snowball effect". Now the snowball effect is difficult to quantify but at the very least we'd expect idle players to end up lower than 45.66%, not higher. Your own account matches that theory. 

 

I'd offer the explanation that most normal accounts who have lower winrates than yours have a normal activity profile which means they progress in tiers. At tier 1, we don't expect to see a lot of skill, mostly new players with "random" behaviour by which I mean lacking the more tactical behaviour of experienced players. The ability to capitalize on enemy mistakes and the ensuing advantages (such as outgunning them) is expected to be higher amongst experienced players, and subsequently, higher among the higher tiers. The most likely explanation for you being "not THAT low" is that you never ventured beyond tier 1 on this account.


I don’t think it works like that as you compare the ‘combat power’ of team vs. team. So let’s say that’s a d100 and one player always subtracts 10 from that result. On average you’re going to lose 10% more rolls and not 5%. Well I think so anyway.



pecopad #23 Posted 18 September 2019 - 06:12 PM

    Captain

  • Player
  • 30909 battles
  • 2,344
  • [UGN] UGN
  • Member since:
    09-04-2015

View Postjabster, on 18 September 2019 - 05:30 PM, said:


I don’t think it works like that as you compare the ‘combat power’ of team vs. team. So let’s say that’s a d100 and one player always subtracts 10 from that result. On average you’re going to lose 10% more rolls and not 5%. Well I think so anyway.

 

I don't think this logic explains the result, although I never thought of the problem on the bad or good player position.

 

I usually think of this problem for an average player, with no impact on the outcome,so the win or loss should be determined by both the skill level of his team and the skill level of the other team. Now on average, and with random skill MM, he should have the same chance of landing in a good team vs a bad team. So my intuition is that he should end with a 50% WR. 

 

Now it all depends on what we consider more important, having high skill players or not having bad players in the team, which one has more impact?

 

For the average player my intuition is that it would be indifferent, good players defend that the most important is the high skill player who can biase results by carrying games, and that's why they have good WR's,       

 

Never saw an argument that defends that WR is determined by bad players....         


Edited by pecopad, 18 September 2019 - 06:13 PM.


Dorander #24 Posted 18 September 2019 - 07:26 PM

    Lieutenant General

  • Player
  • 21025 battles
  • 6,042
  • Member since:
    05-07-2012

View Postjabster, on 18 September 2019 - 04:30 PM, said:


I don’t think it works like that as you compare the ‘combat power’ of team vs. team. So let’s say that’s a d100 and one player always subtracts 10 from that result. On average you’re going to lose 10% more rolls and not 5%. Well I think so anyway.

 

If you pretend that the snowball effect doesn't actually do anything but everything remains even other than the personal rating loss of the player, your remaining teammembers would still carry you to victory half the time, as in a situation where everyhing is even our base winrate is 50%, not 100%.

 

Then again I'm not the most amazing statistician so I might be wrong here, but if the measured loss ought to be 6.67% then we ought to be seeing way worse winrates across the board because the predicted winrate before snowball effect would be 42.33%.

 

View Postpecopad, on 18 September 2019 - 05:12 PM, said:

 

Never saw an argument that defends that WR is determined by bad players....         

 

Of course winrates are also determined by bad players. We beat each other in this game, so winrates are a reflection of relative skill, not absolute skill. If somebody gets good rates, somebody else is getting bad rates, so the winrate differences don't occur because people are good players or because people are bad players, but because some players are better than some other players. For any relative position you need two objects that can relate to each other and thus both are equally important in the measured result.

 

This is why the silly idea that you see here from time to time, that anyone who invests time into learning the game, can reach a 50% winrate minimum. It's mathematically impossible. It's also why it's silly to complain about bad players existing (by which "people with low winrates" are referred). Those people got those winrates because they were beaten by better players, who thus got their higher winrates. With a rating such as winrate, good players CREATE these bad players, just like these bad players create the good players.



Cannes76 #25 Posted 18 September 2019 - 09:56 PM

    Lieutenant

  • Player
  • 71613 battles
  • 1,904
  • [TAKE] TAKE
  • Member since:
    04-16-2011

View PostChristOfTheAbyss, on 18 September 2019 - 04:50 PM, said:

 

Just tested something while playing it. Can I get as bad winrate as some people here have, without intentionally hindering my teammates. I couldnt. I have no idea how people get 43% or worse winrates as I couldnt do it even if I tried.


Harder to do in low tiers than in high tiers. In high tier vehicles you generally have a higher impact as you are often top tier...



Bulldog_Drummond #26 Posted 18 September 2019 - 10:40 PM

    Field Marshal

  • Player
  • 34862 battles
  • 11,525
  • [DRATT] DRATT
  • Member since:
    08-10-2014

View Posttrntbl2109, on 17 September 2019 - 08:27 PM, said:

So all the games are ending 0/1/2-15 or 15-2/1/0. Game is over in 5 minutes and you can tell the result after 2 minutes. You just cannot make any meaningful difference in game, it's instant win or loss. No excitement, no fun. What the heck happened to tier 10? 

 

Oh noes



Tilly042 #27 Posted 19 September 2019 - 03:01 PM

    Staff Sergeant

  • Player
  • 37478 battles
  • 368
  • [A_S_S] A_S_S
  • Member since:
    09-08-2012

View PostNekoPuffer_PPP, on 18 September 2019 - 03:28 PM, said:

 

Well it's true! It doesn't get any worse for you, but for everyone else! :great:

 

Arty is always the answer.

I had about 200 games in Arty total until a few weeks ago.  because I'm rubbish at lights and TDs, I've been on mission 15 for the T55 for both for ages.  I then thought I'd get myself a couple of SPGs to go through the T55 missions and whacked 14 of them out in no time.  Now I'm stuck on mission 15 as well! :(



Kartoshkaya #28 Posted 19 September 2019 - 03:05 PM

    Warrant Officer

  • Player
  • 28205 battles
  • 974
  • [HAPY-] HAPY-
  • Member since:
    01-01-2015

View PostTilly042, on 19 September 2019 - 03:01 PM, said:

I had about 200 games in Arty total until a few weeks ago.  because I'm rubbish at lights and TDs, I've been on mission 15 for the T55 for both for ages.  I then thought I'd get myself a couple of SPGs to go through the T55 missions and whacked 14 of them out in no time.  Now I'm stuck on mission 15 as well! :(

 

Cool



SlyMeerkat #29 Posted 19 September 2019 - 03:57 PM

    Colonel

  • Player
  • 19569 battles
  • 3,626
  • [-RLD-] -RLD-
  • Member since:
    01-29-2013
I just avoid tier 10 altogether, worse tier of them all IMO




1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users