Jump to content

Need someone good at maths to start WoT conspiracy

  • This topic is locked This topic is locked
6 replies to this topic

Stokkolm #1 Posted 27 August 2013 - 11:18 PM


  • Player
  • 13532 battles
  • 1,528
  • Member since:
It has become commonly accepted on forums that the vehicle global win ratio is the best indicator of a tank's balance. Of course, some people have figured out that certain tanks are played by better players, while others are common noob magnets. That's how OPMeter was created, and it's a superior indicator.
But both of them have a fundamental flaw. They don't take into account the number of battles played. Most people on the forums agreed KV-1S is OP yet it's win ratio did not usually stand out. But something else stands out: KV-1S is played almost as much as all other tier 6 tanks combined. Could this be dragging it's win ratio down?
In my opinion yes. But here comes the tricky part, proving it from a mathematical point of view, and I'm not really good at that. What we know is that if KV-1S was played by a single player on the whole server it's win ratio could be anywhere from exactly 0% to 100%. But if there are 2 players and they meet in one match, one will win and the other will lose, resulting in the global win ratio of KV-1S being dragged closer to 50%. If 100% of the players on the server would play only KV-1S than the global win ratio would sit at exactly 50% (not taking into account draws for simplicity).
So the conclusion is the more popular one tank is, the closer it's global win ratio will get to 50%. In practice that means a good tank which is less played can have 60% win ratio, while a stronger tank which is very popular will have only 55%. If this is really true, we'd have to reconsider how we look at World of Tanks' balance.
So, anyone can help with the maths part? It would be embarrassing to launch such daring claims without backing with facts. Thank you.  :smile:

Edited by Stokkolm, 27 August 2013 - 11:20 PM.

GiveThemNothing #2 Posted 27 August 2013 - 11:30 PM

    Warrant Officer

  • Player
  • 8536 battles
  • 686
  • Member since:
Too much time?

Unch_04 #3 Posted 28 August 2013 - 12:54 AM


  • Player
  • 59210 battles
  • 272
  • [FE2O3] FE2O3
  • Member since:
and much more.

assuming everybody play for the record..............

soolerman #4 Posted 28 August 2013 - 06:43 AM

    Second Lieutenant

  • Player
  • 24366 battles
  • 1,218
  • Member since:
What you need is a model to describe the relationship of player type to probable game outcome variables and then to factor in the characteristics of the tanks that affect this. Yea good luck with that one mate.

What you are better off proving is that new players of a particular tank have a lower win ratio than experienced players of the same tank.

I would do this by looking at the sub set numbers. Say the first 20 games a player has in a new tank and what percentage of all players of that tank have 20 games and what is the win ratio. Then repeat that for games 100 to 120. Then for 200 to 220 games and so on. What you will find if your suspicions are correct is that the win ratio will go up in each successive sub set you look at but the number of players in each sub set will fall. This method will stop you having to crunch large numbers demonstrate your point and stop us all having to get a maths PHD to understand the end dater.  :smile:

Hydro69 #5 Posted 28 August 2013 - 09:03 AM


  • Player
  • 22843 battles
  • 55
  • [1STD] 1STD
  • Member since:
Don't forget to consider crew experience, skills, modules, consumables, gold, he shells... in the equation.

KawaGreen #6 Posted 28 August 2013 - 09:11 AM


  • Clan Diplomat
  • 13307 battles
  • 1,931
  • Member since:

View PostStokkolm, on 27 August 2013 - 11:18 PM, said:

But if there are 2 players and they meet in one match, one will win and the other will lose, resulting in the global win ratio of KV-1S being dragged closer to 50%.
And if they draw?
I got in total 153 draws, that's almost 2%.

Edited by KawaGreen, 28 August 2013 - 09:12 AM.

anonym_kL7qtn3e52MB #7 Posted 28 August 2013 - 10:10 AM

    Lieutenant General

  • Player
  • 0 battles
  • 6,815
  • Member since:
Ahhh, the old "conspiracy"  thread.

Without actual numbers, no chance  in hell to prove anything.


100 matches are not enough to see what is really happening.
1000 matches, not enough
10.000 matches, still not enough.

You know why it's never enough?
The Random Number Generator...
Even if it rolls " 1, 1, 1, 1, 1, 1, 2", it's random.

So forget about it.

1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users