Jump to content


Use MoEs instead of WR to determine game balance


  • Please log in to reply
17 replies to this topic

Sentience #1 Posted 10 September 2018 - 05:44 AM

    First lieutenant

  • Players
  • 24999 battles
  • 776
  • [ARROW]
  • Member since:
    01-02-2016

This is constructive criticism of the current method used to balance the game.

 

Based on various threads/posts on the topic, it is clear that WG uses a global winrate average (WR) to balance tanks within the game. This is a bad idea because:

 

1. Global population gameplay does not reflect gameplay in a competitive environment. (Noise)

2. Global winrate does not accurately reflect tank lethality (bloom,reload,alpha,etc) which most adjustments seem focused around.

3. Global winrate does not take into account packages/configurations of player tanks.

 

The first point is straight forward - if the game is to be taken seriously as an e-sport, it should be fair and balanced at its highest level. By using the global winrate, you are including noise, such as yolo tankers, AFKers, teammate killers, stock tanks, and complete beginners in the data. In recent competitions (e.g ESL on PS4), we saw (competent) teams rely heavily on IS-3s and it’s variants, with very few situational substitutions (Lowe, AMX, T32). If the game was truly balanced, this wouldn’t be the case. You would see more diverse team configurations, centered around different strategies/tactics/synergies.

 

Let’s use the following game as an example to illustrate the other two points:

 

Spoiler

 

In this game, my team won by capping early. What are the balance implications of this game? From a global winrate perspective, the Waffle E100 would actually be buffed, as a team with two waffles is clearly inferior to a team with only one. M48 Patton would receive a huge buff, since clearly having two on the losing team contributed to the loss. The Indien Panzer would receive a nerf, since clearly having one on your team improves you winrate. What gets lost here is the lethality of the tank - in the right hands, a Waffle is utterly devastating to an entire flank. Also lost is the loadouts/skills/equipment of the tanks involved.

 

Here’s what I propose to fix balancing - use MoE II and MoE III data only to balance your tanks.

 

1. Taking only the top 15% of players into account eliminates most sources of noise, and reflects actual gameplay of competitive minded players.

2. MoEs accurately reflect tank lethality (both direct and indirect/spotting) relative to others in its tier/class.

3. Top 15% of players generally use effective tank loadouts/configurations.

 

If the general population does not believe competitively balanced tanks to be fair, that is something that can be addressed through better tutorials and community engagement (both officially/unofficial). Also - some players in the general population are simply not interested in improving their tanking skills.

 


How to fix game balance in a nutshell: Click here.

How to fix Mercenaries in a nutshell: Click here.

How to keep history buffs happy: Click here.

How to make clans relevant again: Click here.


dzhey43 #2 Posted 10 September 2018 - 06:54 AM

    Sergeant

  • Beta Tester
  • 20219 battles
  • 212
  • [CRAKI]
  • Member since:
    11-14-2013

That’s bad idea. MoA players and WR players have different goals in mind, first think about damage, they do not fight for victory, loss is ok if they do more damage. And seconds think about victory. And the result of every battle is not about who made more damage, but who won or lost.

 

There is player efficiency on the profile, that could work maybe)



DwarfOnDrugs #3 Posted 10 September 2018 - 09:17 AM

    Captain

  • Players
  • 26439 battles
  • 1,425
  • [-UNL-]
  • Member since:
    03-12-2014

Or just work out what nerfs/buffs/rebalancings are needed by playing the game.

e.g. anyone would have said that the british meds and heavies needed a buff. anyone would have realised the 121/131 needed a bit of loving. anyone would have said the amx 30/b was fine in its pre-nerfed state. the majority of people agree the deathstar and waffle need a rebalancing.

Also when looking at how to buff/nerf tanks look at the amount of games played in the vehicle, if the vehicle isn't played that often there is most likely a reason for that. Either too new or not popular for a reason (Either sub-par stats or a very complex play-style that the average player can't play.


(And yes i do need a new banner i can't be bothered to change it yet)

https://www.youtube....GOV0Ca_kQVSX2-w


KilledByPing #4 Posted 10 September 2018 - 12:19 PM

    Captain

  • Players
  • 10922 battles
  • 1,600
  • Member since:
    02-12-2017

View PostSentience, on 10 September 2018 - 08:44 AM, said:

This is constructive criticism of the current method used to balance the game.

 

Based on various threads/posts on the topic, it is clear that WG uses a global winrate average (WR) to balance tanks within the game. This is a bad idea because:

 

1. Global population gameplay does not reflect gameplay in a competitive environment. (Noise)

2. Global winrate does not accurately reflect tank lethality (bloom,reload,alpha,etc) which most adjustments seem focused around.

3. Global winrate does not take into account packages/configurations of player tanks.

 

The first point is straight forward - if the game is to be taken seriously as an e-sport, it should be fair and balanced at its highest level. By using the global winrate, you are including noise, such as yolo tankers, AFKers, teammate killers, stock tanks, and complete beginners in the data. In recent competitions (e.g ESL on PS4), we saw (competent) teams rely heavily on IS-3s and it’s variants, with very few situational substitutions (Lowe, AMX, T32). If the game was truly balanced, this wouldn’t be the case. You would see more diverse team configurations, centered around different strategies/tactics/synergies.

 

Let’s use the following game as an example to illustrate the other two points:

 

Spoiler

 

In this game, my team won by capping early. What are the balance implications of this game? From a global winrate perspective, the Waffle E100 would actually be buffed, as a team with two waffles is clearly inferior to a team with only one. M48 Patton would receive a huge buff, since clearly having two on the losing team contributed to the loss. The Indien Panzer would receive a nerf, since clearly having one on your team improves you winrate. What gets lost here is the lethality of the tank - in the right hands, a Waffle is utterly devastating to an entire flank. Also lost is the loadouts/skills/equipment of the tanks involved.

 

Here’s what I propose to fix matchmaker - use MoE II and MoE III data only to balance your tanks.

 

1. Taking only the top 15% of players into account eliminates most sources of noise, and reflects actual gameplay of competitive minded players.

2. MoEs accurately reflect tank lethality (both direct and indirect/spotting) relative to others in its tier/class.

3. Top 15% of players generally use effective tank loadouts/configurations.

 

If the general population does not believe competitively balanced tanks to be fair, that is something that be addressed through better tutorials and community engagement (both officially/unofficial). Also - some players in the general population are simply not interested in improving their tanking skills.

 

 

Thats not going to fix anything since WG matchmaker put all the good players in same team against all the bad players.

It doesnt matter what will be used to determine which one of them are good. :trollface:



Sentience #5 Posted 10 September 2018 - 12:56 PM

    First lieutenant

  • Players
  • 24999 battles
  • 776
  • [ARROW]
  • Member since:
    01-02-2016

View PostDwarfOnDrugs, on 10 September 2018 - 04:17 AM, said:

Or just work out what nerfs/buffs/rebalancings are needed by playing the game.

e.g. anyone would have said that the british meds and heavies needed a buff. anyone would have realised the 121/131 needed a bit of loving. anyone would have said the amx 30/b was fine in its pre-nerfed state. the majority of people agree the deathstar and waffle need a rebalancing.

Also when looking at how to buff/nerf tanks look at the amount of games played in the vehicle, if the vehicle isn't played that often there is most likely a reason for that. Either too new or not popular for a reason (Either sub-par stats or a very complex play-style that the average player can't play.

 

I’m proposing a method to quantify the balance problem. It’s one one thing to say this tank isn’t balanced, it’s another to know by how much that tank is unbalanced. Simply playing the game isn’t good enough - you limit your sample size to your games, on your server, at your time of day, and your playstyle.

How to fix game balance in a nutshell: Click here.

How to fix Mercenaries in a nutshell: Click here.

How to keep history buffs happy: Click here.

How to make clans relevant again: Click here.


Sentience #6 Posted 10 September 2018 - 01:01 PM

    First lieutenant

  • Players
  • 24999 battles
  • 776
  • [ARROW]
  • Member since:
    01-02-2016

View Postdzhey43, on 10 September 2018 - 01:54 AM, said:

That’s bad idea. MoA players and WR players have different goals in mind, first think about damage, they do not fight for victory, loss is ok if they do more damage. And seconds think about victory. And the result of every battle is not about who made more damage, but who won or lost.

 

There is player efficiency on the profile, that could work maybe)

 

The nice thing about MoEs is that it accounts for indirect damage. This is why it is a superior marker than WN8 in determining tank effectiveness. Secondly, tanks should be balanced relative to each other, not on its ability to defend/capture objectives. Reading a map, deciding whether or not to cap, are decisions for the player to make.

How to fix game balance in a nutshell: Click here.

How to fix Mercenaries in a nutshell: Click here.

How to keep history buffs happy: Click here.

How to make clans relevant again: Click here.


dzhey43 #7 Posted 10 September 2018 - 02:06 PM

    Sergeant

  • Beta Tester
  • 20219 battles
  • 212
  • [CRAKI]
  • Member since:
    11-14-2013

View PostSentience, on 10 September 2018 - 01:01 PM, said:

 

The nice thing about MoEs is that it accounts for indirect damage. This is why it is a superior marker than WN8 in determining tank effectiveness. Secondly, tanks should be balanced relative to each other, not on its ability to defend/capture objectives. Reading a map, deciding whether or not to cap, are decisions for the player to make.

 

If one goes to mark the tank he will never cap) This players are as bad as noob players, they won’t risk if his team is loosing, while he seats undetected in the bushes no matter tank type and there enough enemies to shoot he will do nothing for the win, damage is more needed than to win. Btw wn8 is based on damage mostly. Wn8 players are exactly the same as MoE player. The tank efficiency is the number on your profile on official site. I don’t know how it is calculated (never was interested before).

DwarfOnDrugs #8 Posted 10 September 2018 - 02:22 PM

    Captain

  • Players
  • 26439 battles
  • 1,425
  • [-UNL-]
  • Member since:
    03-12-2014

View PostSentience, on 10 September 2018 - 12:56 PM, said:

 

I’m proposing a method to quantify the balance problem. It’s one one thing to say this tank isn’t balanced, it’s another to know by how much that tank is unbalanced. Simply playing the game isn’t good enough - you limit your sample size to your games, on your server, at your time of day, and your playstyle.

 

How about getting a group of 10 prople from the active community ranging in skill level. And get all their opinions on the vehicles WG wants to change before they change it at all.

Someone from (Rough guide);

IMTLS/UNL

CDR

N/NL

KMD

SQUAD

BEARD/GUYS

XGN

WGW

 


(And yes i do need a new banner i can't be bothered to change it yet)

https://www.youtube....GOV0Ca_kQVSX2-w


HopHound83 #9 Posted 11 September 2018 - 11:12 AM

    Corporal

  • Players
  • 8141 battles
  • 99
  • Member since:
    10-04-2014

The problem with WR is if you're or were a .500 player and improving or even holding steady you get thrown in with the sub .500 crowd more often than not and well hey check the timer it's 2 minutes in and half of your team is dead. That's hard on the middle to get better WR's to move up the skill ladder. Lets be honest the middle is probably the largest group. However current matchmaking is creating basically two classes. The teams that always win by crushing margins and the grapes they ran over. 18 months ago I had a win rate in the upper 50"s, 57 point something if memory serves. Then it was day after day of terrible teams no matter how well I did or my platoon did we lost. Some of these games are over so fast its like the Green Berets versus the Boy scouts.

Every rare once in awhile I get a run of being on a decent team. It seems to depend on the day and time. Its like the WoT lottery, did I pick the right day and time.  I'm not a tanking god by any means, I'm competent, I account for myself do decent damage, get kills, and support my team the best I can. I do watch the online videos and check other players techniques to get better. But honestly I quit playing last November out of pure frustration only came back 6 weeks ago. Last year, the last month, sheesh! What a joke games where I would have 5 or more kills by myself and literally that would be the only kills on my team. 8 or more players on my team with goose eggs across the board. How do you even get to tier VII or VIII and not hit anything at least once in a match? I noticed the decline right after the first Motherland earn op. No idea what happened but it just got ridiculous fast after that. After 6 weeks of being back same same. I get kills, do damage, finish in the top 3, MoE"s, Ace Mastery Badges, and still....loss after loss.



I44I Warlock #10 Posted 11 September 2018 - 11:20 AM

    Major

  • Players
  • 32923 battles
  • 2,510
  • [-MLW-]
  • Member since:
    02-12-2014

This is not a good idea. 

Moe requirements are 3,2k for the E75, 4k for the E5 and 215b, thats pretty low but that doesn't mean it needs a buff.

(super-) Unicums can rise the requirements high even if the tank is not op, especially when the tanks playerbase is low.

 

A lot of needed changes can be discovered by taking a look at the tanks soft stats, or just by playing the game. 


     

                  

 

                Heute fette Beute!
                        
 
Nerf: 183, WT-E100 and Arty
Arty is OP and takes no skill.
 
Currently 3 Maked tanks: 62
               
                              
                        

Sentience #11 Posted 11 September 2018 - 02:22 PM

    First lieutenant

  • Players
  • 24999 battles
  • 776
  • [ARROW]
  • Member since:
    01-02-2016

View PostI44I Warlock, on 11 September 2018 - 06:20 AM, said:

This is not a good idea. 

Moe requirements are 3,2k for the E75, 4k for the E5 and 215b, thats pretty low but that doesn't mean it needs a buff.

(super-) Unicums can rise the requirements high even if the tank is not op, especially when the tanks playerbase is low.

 

A lot of needed changes can be discovered by taking a look at the tanks soft stats, or just by playing the game. 

 

I’m going to sound like a broken record here. When you make important decisions, you want to be able to quantify that decision. Game balancing is an important decision. When patch 4.6 was released, the developers were compelled to start listing the actual statistics they used to determine which tanks needed to be buffed or nerfed. We can argue about the stats they picked (WR versus anything else) but the decision to base it on hard data is the correct approach.

 

Its hard to justify a changing a tank based on “feel”. What you feel about that tank is based on your experience/playstyle/server/whatever. It is unique to you. Since experiences vary, how would one choose to valuate one’s feedback over another?

 

Hypothetical example:

 

Most people think the Waffle E 100 is overpowered. Ok - how do we quantify that? Do we nerf it a little? A lot? Just give up and remove it from the game?

 

By using hard data, from examining its WR/MoE/whatever with its peers, we gain an objective “feel”. We create a reference point from which we can see how effective our adjustment was. And that’s based on the “feel” of hundreds of players. Not just some dev holding a controller behind a desk. That’s where quality data matters, my proposal is simply a more precise method to gather it.


How to fix game balance in a nutshell: Click here.

How to fix Mercenaries in a nutshell: Click here.

How to keep history buffs happy: Click here.

How to make clans relevant again: Click here.


SHRiIVIP #12 Posted 11 September 2018 - 02:53 PM

    Major

  • Players
  • 16307 battles
  • 2,438
  • [-NL-]
  • Member since:
    02-18-2017

View Postdzhey43, on 10 September 2018 - 02:06 PM, said:

 

If one goes to mark the tank he will never cap) This players are as bad as noob players, they won’t risk if his team is loosing, while he seats undetected in the bushes no matter tank type and there enough enemies to shoot he will do nothing for the win, damage is more needed than to win. Btw wn8 is based on damage mostly. Wn8 players are exactly the same as MoE player. The tank efficiency is the number on your profile on official site. I don’t know how it is calculated (never was interested before).

 

Really? So that's why almost every single player with a high wn8 and many tanks with 3 MoE routinely have 60%+ win rates? I must be missing all these players you're talking about who have had sub 50% win rate with high wn8s and many 3 MoE tanks.

Former Account: HappeningShrimp

46 Tanks with 3 MoE

 Spoiler


Warthog II #13 Posted 11 September 2018 - 06:30 PM

    Captain

  • Players
  • 39359 battles
  • 1,145
  • [KMD]
  • Member since:
    05-11-2014

View PostDwarfOnDrugs, on 10 September 2018 - 09:17 AM, said:

Or just work out what nerfs/buffs/rebalancings are needed by playing the game.

e.g. anyone would have said that the british meds and heavies needed a buff. anyone would have realised the 121/131 needed a bit of loving. anyone would have said the amx 30/b was fine in its pre-nerfed state. the majority of people agree the deathstar and waffle need a rebalancing.

Also when looking at how to buff/nerf tanks look at the amount of games played in the vehicle, if the vehicle isn't played that often there is most likely a reason for that. Either too new or not popular for a reason (Either sub-par stats or a very complex play-style that the average player can't play.

 

131?

 I think you mean the 113 or perhaps that 133 mentioned in the extra patch notes?


  

Some people claim Ensk should be tier restricted.   I don't believe their lies. 


sledge6301 #14 Posted 11 September 2018 - 06:37 PM

    Captain

  • Players
  • 45084 battles
  • 1,170
  • [PIRAI]
  • Member since:
    02-20-2017

Where is the competetive gamemode ?

 

I cant find it?

 

Wot console never have or are you refering to multiplayer game mode?

 

The only one is war story player VS enviroment ( AI Boots ).

 

 



sledge6301 #15 Posted 11 September 2018 - 06:49 PM

    Captain

  • Players
  • 45084 battles
  • 1,170
  • [PIRAI]
  • Member since:
    02-20-2017

View PostSentience, on 11 September 2018 - 02:22 PM, said:

 

I’m going to sound like a broken record here. When you make important decisions, you want to be able to quantify that decision. Game balancing is an important decision. When patch 4.6 was released, the developers were compelled to start listing the actual statistics they used to determine which tanks needed to be buffed or nerfed. We can argue about the stats they picked (WR versus anything else) but the decision to base it on hard data is the correct approach.

 

Its hard to justify a changing a tank based on “feel”. What you feel about that tank is based on your experience/playstyle/server/whatever. It is unique to you. Since experiences vary, how would one choose to valuate one’s feedback over another?

 

Hypothetical example:

 

Most people think the Waffle E 100 is overpowered. Ok - how do we quantify that? Do we nerf it a little? A lot? Just give up and remove it from the game?

 

By using hard data, from examining its WR/MoE/whatever with its peers, we gain an objective “feel”. We create a reference point from which we can see how effective our adjustment was. And that’s based on the “feel” of hundreds of players. Not just some dev holding a controller behind a desk. That’s where quality data matters, my proposal is simply a more precise method to gather it.

 

Thumbs up to this.

 

btw waffle got a reload nerf to 60 sec now.

 

WR is one way and they have to choose one the problem was they did not say they did not inform us the player base about why and what if they did we would let this happen and looking forward to a follow up and WG monitoring to find out problems and correct them.

 

Now they have to trow everything away all hard work and we are left with what?

 

Patience is key in every change but monitoring and reacting is vital for the change to be supported by the end user .

 

WG why do you still after all this bashing still refuse to share your good intensions with your customers you have all the tools to have pools and email social media to take control of this there is still time you can fix this you can we want you to do you?

 

 



MortDeLaPeur #16 Posted 11 September 2018 - 07:53 PM

    Corporal

  • Players
  • 15739 battles
  • 31
  • [OMCS]
  • Member since:
    11-06-2016
Probably the biggest problem to match making is tank type to map type.  I don't know if there are stats per tank per map but if they used those the teams could be balanced better.  If you put 10 tanks on one team that do good on a particular map and 10 tanks that do bad on that map the other team there will probably be an unbalanced win.  I really believe that it is tank performance per map that is determining the wins.

DwarfOnDrugs #17 Posted 11 September 2018 - 08:04 PM

    Captain

  • Players
  • 26439 battles
  • 1,425
  • [-UNL-]
  • Member since:
    03-12-2014

View PostWarthog II, on 11 September 2018 - 06:30 PM, said:

 

131?

 I think you mean the 113 or perhaps that 133 mentioned in the extra patch notes?

 

113, was confusing my small brain with 121 and 113

(And yes i do need a new banner i can't be bothered to change it yet)

https://www.youtube....GOV0Ca_kQVSX2-w


steelheadclan #18 Posted 12 September 2018 - 02:29 PM

    Major

  • Players
  • 13741 battles
  • 2,547
  • Member since:
    03-05-2017

View Postdzhey43, on 10 September 2018 - 06:06 AM, said:

 

If one goes to mark the tank he will never cap) This players are as bad as noob players, they won’t risk if his team is loosing, while he seats undetected in the bushes no matter tank type and there enough enemies to shoot he will do nothing for the win, damage is more needed than to win. Btw wn8 is based on damage mostly. Wn8 players are exactly the same as MoE player. The tank efficiency is the number on your profile on official site. I don’t know how it is calculated (never was interested before).

Lol

I mainly play for marks and WN8 and I have less games than you and a higher win rate. 

 

My first 3 marked tank has 700+ games and a 67% win rate. 

 

Almost all my 2 and 3 marked tank are in the 58-73% range. 

 

Defending the cap is often the best way to farm hit points and I’ll never cap early or with an advantage so I and rest of team can get damage and the XP and silver that comes with it. 

 

 

 

 

 






1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users