Jump to content


Basing tank balancing on arithmetic means is wrong

balance

  • Please log in to reply
4 replies to this topic

Sengell #1 Posted 23 September 2018 - 07:22 PM

    Corporal

  • Players
  • 3974 battles
  • 28
  • Member since:
    11-21-2017

Wargaming explained on Youtube that they strive for arithmetic means in wins and losses in order to balance tanks. The assumption attaining thos arithmetic means would equal a balanced game is wrong. How the majority fares in a tank is just a limited indicator for balance which is not sufficient enough for those decisions.

 

Imagine a Formula 1 race car. The majority would crash it in the attempt to drive it. They would do considerably worse in it compared to more forgiving or userfriendly cars. But no one would argue that the Formula 1car is objectively worse than other cars based on the mere fact that the majority could not handle it well.

 

This applies to tanks in WOT aswell. Looking at the majority in the middle of the distribution does not promote any reliable conclusion. Conclusive data for further balancing of tanks has to be derived from proficient tankers at the upper end of the distribution. Therefore arithmetic means and how the majority fares in a tank are useless for balancing decisions.

 

So given your disclosure how you make balancing decisions based on those statistics, your assumptions are wrong and are detrimental for balancing decisions. I hope this was helpful.

 



korbendallas-01 #2 Posted 23 September 2018 - 07:45 PM

    Major

  • Players
  • 39963 battles
  • 6,171
  • [GER_1]
  • Member since:
    05-15-2016

View PostSengell, on 23 September 2018 - 08:22 PM, said:

Imagine a Formula 1 race car. The majority would crash it in the attempt to drive it.

 

Well, wrong actually. The vast majority would not be able to start it.

 

Your point still stands though.

 

View PostSengell, on 23 September 2018 - 08:22 PM, said:

This applies to tanks in WOT aswell. Looking at the majority in the middle of the distribution does not promote any reliable conclusion. Conclusive data for further balancing of tanks has to be derived from proficient tankers at the upper end of the distribution. Therefore arithmetic means and how the majority fares in a tank are useless for balancing decisions.

 

Completely agree there is more than one type of scale to balance a tank. Closely related, I said the other day that a tank can be op in two ways: The first allows beginners to make more points than they would usually do (usually heavily armored tanks like the Minuteman or the Mercenary premiums), and the second allows experienced players to completely dominate games (V/IV, T2).



sledge6301 #3 Posted 23 September 2018 - 09:25 PM

    Captain

  • Players
  • 45960 battles
  • 1,244
  • [PIRAI]
  • Member since:
    02-20-2017

View PostSengell, on 23 September 2018 - 07:22 PM, said:

Wargaming explained on Youtube that they strive for arithmetic means in wins and losses in order to balance tanks. The assumption attaining thos arithmetic means would equal a balanced game is wrong. How the majority fares in a tank is just a limited indicator for balance which is not sufficient enough for those decisions.

 

Imagine a Formula 1 race car. The majority would crash it in the attempt to drive it. They would do considerably worse in it compared to more forgiving or userfriendly cars. But no one would argue that the Formula 1car is objectively worse than other cars based on the mere fact that the majority could not handle it well.

 

This applies to tanks in WOT aswell. Looking at the majority in the middle of the distribution does not promote any reliable conclusion. Conclusive data for further balancing of tanks has to be derived from proficient tankers at the upper end of the distribution. Therefore arithmetic means and how the majority fares in a tank are useless for balancing decisions.

 

So given your disclosure how you make balancing decisions based on those statistics, your assumptions are wrong and are detrimental for balancing decisions. I hope this was helpful.

 

 

You are on topic but wrong in assumtion.

 

A Formula one car is before even  the first lap simulated in every angel and by collected data from previous runs by the same driver it is a big task to just gain 1/10 of a second .

 

The developers never change more then one thing at a time never because if they did they cant have back to back data and that is the most important thing.

 

Drivers are never involved in the direct developmeant they cant be there job is to try to add the feel for the changes by adding human factor.

 

By changing tryin to balance tanks around playerstats is never going to be the right tool the only way is by using tank stats.

 

I think in many ways the devs shoot over the target and we can clearley see that they have no clue what they are doing.

 

Now we lost more then 6 months due to ignorance about the real factors playing the key roll in this game for me it does not matter much but for the competetive players it is devestating.

 

Competition is a hard and hash enviormeant but this game is not built for that it is a multiplayer free client based game shooting tanks.

 

WG dont listen to players do your homework go by tankstats test by AI make the game fun to play and if you want to build a competetive game mode exclude that from RNG.



BrendenF-99 #4 Posted 23 September 2018 - 11:39 PM

    Corporal

  • Players
  • 28257 battles
  • 55
  • [-ALF-]
  • Member since:
    07-07-2016
What they should be doing is only looking at the stats to figure out which tanks need to be further looked at, then playing that tank in game to figure out why the stats are as good as they are, making changes based on the overall stats is pretty much assuming that everyone who plays the tank all have the same skill level, they should know that some tanks are harder to play well and are less forgiving than other tanks, and those tanks like the AMX 30 B are usually played by people who know how to play the tank well because the lack of armor makes it less forgiving than other tanks, so less skilled players would have a harder time playing that tank so they probably won't be playing it too often. so when the majority of players in a tank seem to know what they are doing of course the overall stats on the tank are going to be higher than most of the others.

sledge6301 #5 Posted 25 September 2018 - 02:33 PM

    Captain

  • Players
  • 45960 battles
  • 1,244
  • [PIRAI]
  • Member since:
    02-20-2017

View PostBrendenF-99, on 23 September 2018 - 11:39 PM, said:

What they should be doing is only looking at the stats to figure out which tanks need to be further looked at, then playing that tank in game to figure out why the stats are as good as they are, making changes based on the overall stats is pretty much assuming that everyone who plays the tank all have the same skill level, they should know that some tanks are harder to play well and are less forgiving than other tanks, and those tanks like the AMX 30 B are usually played by people who know how to play the tank well because the lack of armor makes it less forgiving than other tanks, so less skilled players would have a harder time playing that tank so they probably won't be playing it too often. so when the majority of players in a tank seem to know what they are doing of course the overall stats on the tank are going to be higher than most of the others.

 

Skill level is more about the knowledge of the game then specific tanks my guess it is also about experience and battle count in a tank that fits your playstyle and funfactor.

 

This game has a very high treashold to get good at but i see a lot of players running there favorit over and over again that helps.

 

Minimap awerness and playing the tank to its objective meaning not sniping in a heavy supportive in a medium and spotting keeping enemie tanks spooted long enough to get TD..s Heavies to take them out and asking for artillary support is a key but  forgotten by most.







Also tagged with balance

1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users