Jump to content


MOE Calculation Explanation plus 4th Mark Difficulty

MOE 4th Mark MOE Explained

  • Please log in to reply
20 replies to this topic

natedaishmaster #1 Posted 01 October 2021 - 09:21 PM

    Captain

  • Players
  • 30426 battles
  • 1,935
  • [IMTLS]
  • Member since:
    07-29-2014

This is partially an explanation of how MOEs work and partially an explanation of why they’re so damn hard right now. If I get anything wrong on how MOEs work, I would appreciate a correction so it’s accurate please.

 

The following is my interpretation of how the target values for damage standing are calculated based on MisterSPGs descriptions in the Discord. Max Chaos has a description for how individual average damage is calculated using an EMA with this link and is still correct and doesn’t need further explanation and is linked below. Ironically, the explanation provided at the beginning of that thread for target damage is almost completely correct, except it left out the legendary “handicap” that has been the topic of discussion recently and seems to have been deliberately avoided in the thread (Max Chaos avoided questions relating to how it worked)

http://forum-console...57#entry2169257

 

At it’s most simple the formula for your percentage is your EMADamage / TargetDamage. If that’s above 95%/85%/65% then you get 3/2/1 MOE

 

The calculation takes into account every game played in a vehicle in the last 2 weeks and puts them ordered in a list. For this example let’s take the list to be 100 total games in the last two weeks. Without a handicap, the formula takes the highest damage game and sets it as the 100% requirement.

 

An interesting quality is time. Your last 100 battles as a player apply to your EMA regardless of when you played them but only battles in the last 2 weeks apply to the Target Damage. The Target Damage a month ago has nothing to do with the current target damage. Moving on 95% of the best game is the 3 MOE requirement and so on. This is obviously ridiculous as nobody can average 95% of their best game in a tank.

 

This is where the handicap comes in. WG describes it as a map-based handicap and is different for every class. The logic behind it is that different tanks do better or worse on different maps and it has to be accounted for so that the MOEs are possible (the E100 obviously does better on Himmelsdorf than Malinovka on average). In essence this handicap is a fudge factor so WG can make the MOEs the appropriate difficulty.

 

The way the handicap works is by reducing the game used for the 100% requirement from the highest damage game in the set of 100 to the 10th highest damage game in the set (disclaimer, WG won’t disclose the actual handicaps so the numbers are made up but the math is fine).

 

This eliminates the ridiculous outlier games and makes the MOEs the difficulty that WG thinks that they should be. This is the value that WG changed in August. The change effectively changed the target game from selecting the 10th best game to the 6th best game and increased MOE requirements by ~5%.

 

I’ll put a description of some of the practical implications here because it’s important later. It’s a good assumption that better players have better games than worse players. The consequence of this with the formula is that the higher percentage of games played by good players in a tank will cause the target damage for that tank to be very high. It is why stuff like the 113 is difficult to get MOEs on. It is very popular for good players but not popular overall so the requirements are high. An inverse might be some of the giveaway tanks being easier to get MOEs such as the M60 or obj268v5 or T95e6 because they are free so bad/new players make up a larger portion of the games played

 

WG claims that the reason that they increased the handicap is that there was less difference between the best and worst maps for a lot of tanks. This makes sense, the reduced count and prominence of open maps, plus the accuracy and pen buffs all might contribute to maps being more equal. If you go back to the formula, the bottom 50 games getting better doesn’t actually increase the requirement since it’s just the 10th best game. The 10th best game stayed at the same required damage while players had higher EMA damages because there weren’t as many bad maps. WG’s change to the 6th (again, these numbers are just examples) best game was meant to correct for that because WG saw more players getting MOEs than they were before 6.0. I don’t know what WG’s benchmark is but I’d assume it’s more or less 5% of players with >100 games in a tank have 3 MOE, 15% 2 MOE, and 35% 1 MOE since that’s the official description

 

Here is the link to the Discord conversation where most of it was explained. Be warned, it is pretty long

https://discord.com/channels/630895439520202783/801656276228374568/892485326285709383


3 MOE: Hellcat, Lycan, Sp1C, T20, M41 Bulldog, Motherland, STA 1, T95E2, Pershing, AMX 50 100, AMX 13 90, T54ltwt, T49, RU251, M46 Patton, Type 61, Leopard PTA, T54, E50, Batchat AP, Skoda T50, Centurion 7/1, Sharpshooter, T10, Conquerer, AMX 13 105, Sheridan, T-100LT, RHM PZW, M60, Chisel, M48 Patton, Leopard 1, E50M, Obj 140, Batchat 25t, TVP 50/51, AMX 50B, Valour, IS7, Super Conqueror, The Machine, Chieftain, FV215B, T110E4

 

MOE Calculation Explanation: http://forum-console...ark-difficulty/


natedaishmaster #2 Posted 01 October 2021 - 09:21 PM

    Captain

  • Players
  • 30426 battles
  • 1,935
  • [IMTLS]
  • Member since:
    07-29-2014

I saved speculation and personal opinion for the comments but it is essentially a continuation of the post. This brings us to now where MOEs are more difficult than they’ve ever been. It is a combination of things causing the high mark requirements.

 

The first is obviously WG’s handicap adjustment since it directly raised target damage by ~5%

 

The second is the 4th mark itself. MisterSPG revealed that most tanks have seen a ~10% increase in target damage on top of the 5% caused by the handicap change, in total most tanks are 15-20% harder to mark than they were 2 months ago

 

This is due to how the MOE calculation works and the habits of good players. Many good players, myself included, will play a lot of tanks until we get 3 MOE (or X battles to appear on 3rd party leaderboards) and move on to the next tank to do the same. With the 4th mark being announced, all of these good players have reason to pick up the tanks that they stopped playing and sweat for the 100%. Now that a lot of good players are playing games in a lot of different tanks all at once for the 4th mark, the proportion of good players in a lot of tanks has increased so the target damage has increased (from MisterSPG, stuff like the SConq, IS7, and bat all have higher % of good players playing it now than before). This has happened in the past where stuff like the TVP had ridiculous requirements on release because a lot of good players played it, but it has never happened across the board like it has today. It’s one thing to have a single very difficult tank but never has it been so hard for almost every tank in the game (specifically tier X). This would have happened even without the handicap change that WG made.

Here comes some of the hardcore speculation. I think that the handicap compounded the inflation that was going to happen for 4th MOEs more than just the two would increase alone. Higher requirements caused good players to have to play more games with more degenerate tactics (platooning with arty to farm assisted damage to inflate personal average damage) while prefarming 100% standing on a lot of tanks. More games by good players means a higher proportion of games by good players which means higher target damage. I think that this effect was entirely unanticipated by WG. They probably could’ve delayed the handicap change until after the 4th MOE inflation went down because the inflation would have a similar effect of increasing the target damage that they wanted

 

MOEs now are objectively harder than they’ve ever been. There’s no way 35% of players playing most given tanks have 1 MOE, 15% with 2 MOE, etc. I easily consider myself top 1% player and most of my tier Xs are between 90-95% (I had most of them >95% before 6.0) which means that whatever WG is using for tracking for the number of players with MOEs must be showing that not enough player can reach the requirements as WG would normally like. WG thinks (probably correctly) that MOEs will eventually settle down to normal difficulty in some amount of time, probably 1-4 months depending on how quickly good players either get their 4th mark or give up. I personally (and a lot of players based on Kellen’s poll) am frustrated by it and think that WG should do something to temporarily lower requirements to a less impossible benchmark. This has 2 benefits. First it would give a more consistent expectation for marking tanks rather than the hyperinflated state it’s in now. Second, it would get the best players their 4th mark quicker and cause the target damage to go back to a more normal value quicker than it would naturally. WG has shown some resistance to this as nothing is going “wrong”, it is all happening within the system and the system should correct itself. It just kinda sucks as a player to be locked out of doing something due to an unanticipated combined effect of 2 things inflating MOE requirements on almost every tank to ridiculous level (honestly the fact that it’s game wide is the real issue, we were ok with a random tank or two being extremely challenging but it being on almost every tank is ridiculous)


Edited by natedaishmaster, 01 October 2021 - 09:33 PM.

3 MOE: Hellcat, Lycan, Sp1C, T20, M41 Bulldog, Motherland, STA 1, T95E2, Pershing, AMX 50 100, AMX 13 90, T54ltwt, T49, RU251, M46 Patton, Type 61, Leopard PTA, T54, E50, Batchat AP, Skoda T50, Centurion 7/1, Sharpshooter, T10, Conquerer, AMX 13 105, Sheridan, T-100LT, RHM PZW, M60, Chisel, M48 Patton, Leopard 1, E50M, Obj 140, Batchat 25t, TVP 50/51, AMX 50B, Valour, IS7, Super Conqueror, The Machine, Chieftain, FV215B, T110E4

 

MOE Calculation Explanation: http://forum-console...ark-difficulty/


I44I Warlock #3 Posted 01 October 2021 - 10:19 PM

    Major

  • Players
  • 47338 battles
  • 3,723
  • [I66I]
  • Member since:
    02-12-2014

View Postnatedaishmaster, on 01 October 2021 - 10:21 PM, said:

 

The way the handicap works is by reducing the game used for the 100% requirement from the highest damage game in the set of 100 to the 10th highest damage game in the set (disclaimer, WG won’t disclose the actual handicaps so the numbers are made up but the math is fine).

 

 

First of all, thanks for taking the time and explaining it as well as your efforts on Discord. 

 

 

But i think you got the calculation for the requirements wrong (or maybe i get your explaination wrong). The way i understood is: 

>They take the highest damage game on the best map and the highest damage game on the worst map.

>Now they list all matches which are between these 2 ascending after combined damage. 

>now they take the median of these matches 

>this median is 100% damage standing. Means it is not the Xth highest damage game. 

>That's The way the requirements are calculated. 

 

To keep the requirements more realistic, they had a negative offset/handicap that reduced the threshold to get the 65%/85%/95%/100% achievement. 

This negative offset/handicap has been recalculated now, as you already mentioned. 


 

                             

 

               - Execute Order 66 -
                        
 
 3 marked tanks: 101
 
>Arty is op, takes no skill and is a toxic game mechanic. 
>WG fix your server connection 
                              
                        

natedaishmaster #4 Posted 01 October 2021 - 10:30 PM

    Captain

  • Players
  • 30426 battles
  • 1,935
  • [IMTLS]
  • Member since:
    07-29-2014

View PostI44I Warlock, on 01 October 2021 - 02:19 PM, said:

 

First of all, thanks for taking the time and explaining it as well as your efforts on Discord. 

 

 

But i think you got the calculation for the requirements wrong (or maybe i get your explaination wrong). The way i understood is: 

>They take the highest damage game on the best map and the highest damage game on the worst map.

>Now they list all matches which are between these 2 ascending after combined damage. 

>now they take the median of these matches 

>this median is 100% damage standing. Means it is not the Xth highest damage game. 

>That's The way the requirements are calculated. 

 

To keep the requirements more realistic, they had a negative offset/handicap that reduced the threshold to get the 65%/85%/95%/100% achievement. 

This negative offset/handicap has been recalculated now, as you already mentioned. 

 

The way I understood it from SPG is that what you're saying is how they came up with the handicap in the first place and doesn't actively factor into the calculation. From what I understood from SPG (and my own limited coding knowledge) the map and player data gets stripped away from the result fairly early in the MOE calculation since it's a lot easier to just store and access a 1xN array that only has numbers. Practically both methods will look and act the same if the handicap is modified correctly so it's more of a nuance and doesn't distract from how the calculation works

Edited by natedaishmaster, 01 October 2021 - 10:30 PM.

3 MOE: Hellcat, Lycan, Sp1C, T20, M41 Bulldog, Motherland, STA 1, T95E2, Pershing, AMX 50 100, AMX 13 90, T54ltwt, T49, RU251, M46 Patton, Type 61, Leopard PTA, T54, E50, Batchat AP, Skoda T50, Centurion 7/1, Sharpshooter, T10, Conquerer, AMX 13 105, Sheridan, T-100LT, RHM PZW, M60, Chisel, M48 Patton, Leopard 1, E50M, Obj 140, Batchat 25t, TVP 50/51, AMX 50B, Valour, IS7, Super Conqueror, The Machine, Chieftain, FV215B, T110E4

 

MOE Calculation Explanation: http://forum-console...ark-difficulty/


Kaptain Proton #5 Posted 01 October 2021 - 11:07 PM

    First lieutenant

  • Players
  • 57835 battles
  • 538
  • [47R]
  • Member since:
    03-07-2016

Well this is more complex than I thought.   I got my first 4th mark my favorite tank.  The VK 30.01D.   I just looked at my combined damage after each game and see how it affected the percentage.  That gave me a target number.   
 

I am going for two other tanks but it has been difficult.   Each time I play I try to get .5%.   Then quit.   It gets a little to stressful being in full try hard mode.   



atwilliams07 #6 Posted 01 October 2021 - 11:40 PM

    Captain

  • Players
  • 45185 battles
  • 1,406
  • Member since:
    03-25-2017
Great post OP. Thanks

Tempest fox3 #7 Posted 02 October 2021 - 02:21 AM

    Major

  • Beta Tester
  • 27217 battles
  • 25,633
  • [JOCO]
  • Member since:
    08-22-2013

I was under the impression one of the things they changed which made MOEs change a while back. Was they weighted your results on various maps differently. IE if you played a map that was historically unfavorable to the tank your score would get rated higher than the same score on a map that was 'favorable' to the tank.

 

But they removed/modified that weighting system a few updates ago. Unless I misunderstood something. 


War is a Democracy and the enemy always gets a vote.

3 MoE's: E-25, M41 Walker Bulldog, E-50, Snakebite, E-75, T32, Tiger 131, Skoda T25, Boilermaker, Sentinel AC4 Experimental, M41B Brazil, Obj.416, Cromwell Knight, M551 Sheridan (Cold war) - In order obtained

Spoiler

PhonicKitty8825 #8 Posted 02 October 2021 - 08:39 AM

    Major

  • Players
  • 27764 battles
  • 5,317
  • [USC4N]
  • Member since:
    04-10-2019
It's just messed up. Why make it THAT hard? I hope those 10+ people with gold marks are enjoying themselves.

4th MoE/Gold Moe  3rd MoE: IS-7, IS-3, M.O.B.A.T., M24 Chaffee '53, T-44A, T-72AV, M113, M551 Sheridan, M46A1 Patton


FluffyColt12271 #9 Posted 02 October 2021 - 12:35 PM

    Staff sergeant

  • Players
  • 24507 battles
  • 465
  • Member since:
    09-01-2019

View Postnatedaishmaster, on 01 October 2021 - 09:21 PM, said:

This is where the handicap comes in. 

.....

 

 I don’t know what WG’s benchmark is but I’d assume it’s more or less 5% of players with >100 games in a tank have 3 MOE, 15% 2 MOE, and 35% 1 MOE since that’s the official description

 

 

First, thanks for setting this out. 

 

Repeating some of what I put on Discord, just for the record, but adding some points.

 

1. The map-based handicap is an arbitrary adjustment. I use the example of what would happen if there was only 1 map. There would be no map-based adjustment, so the 100% MOE requiremt would be set at 100% of the single best game from the past fortnight. Or if there were two maps with the minutest difference in games, then the handicap would be near zero. 

 

I am not saying that there hasn't been a narrowing of the map gap but if I understand how it works it means that applying the formula with updated numbers does not give the "correct" answer if the goal is to have the correct difficulty. Think about the one map example and let me know if you agree or disagree?

 

2. You assume the requirement is 5% have 3 moe - I'd say that's a big assumption and goes against what misterSPG said on Discord. He said there that having marks that are impossible to get if too many unis play the tank is a *feature* not a bug. 

 

I would actually be amazed if the formula worked so that 5% of players had three MOE because there is nothing in the formula disclosed so far that makes that the outcome. Unless that map delta was set so correctly, but based on what misterSPG has set out I don't see how that is a part of the calculation. 

 

Good stuff, I've been enjoying this saga. 



natedaishmaster #10 Posted 02 October 2021 - 03:17 PM

    Captain

  • Players
  • 30426 battles
  • 1,935
  • [IMTLS]
  • Member since:
    07-29-2014

View PostFluffyColt12271, on 02 October 2021 - 04:35 AM, said:

 

First, thanks for setting this out. 

 

Repeating some of what I put on Discord, just for the record, but adding some points.

 

1. The map-based handicap is an arbitrary adjustment. I use the example of what would happen if there was only 1 map. There would be no map-based adjustment, so the 100% MOE requiremt would be set at 100% of the single best game from the past fortnight. Or if there were two maps with the minutest difference in games, then the handicap would be near zero. 

 

I am not saying that there hasn't been a narrowing of the map gap but if I understand how it works it means that applying the formula with updated numbers does not give the "correct" answer if the goal is to have the correct difficulty. Think about the one map example and let me know if you agree or disagree?

 

2. You assume the requirement is 5% have 3 moe - I'd say that's a big assumption and goes against what misterSPG said on Discord. He said there that having marks that are impossible to get if too many unis play the tank is a *feature* not a bug. 

 

I would actually be amazed if the formula worked so that 5% of players had three MOE because there is nothing in the formula disclosed so far that makes that the outcome. Unless that map delta was set so correctly, but based on what misterSPG has set out I don't see how that is a part of the calculation. 

 

Good stuff, I've been enjoying this saga. 


1. I don’t think the “handicap” is specifically for maps. It’s seems like a general use modifier that can be adjusted for anything, maps were just the justification this time. It will always be there cuz statistically you can’t take the 99th or 100th percentile game because they’re outliers so you always need to go at least to the 95-97 range

 

2. Yeah I doubt it works as it says in game. WG obviously has some target for % of players getting MOEs but I don’t know what it is. This was an easy example that most people would understand since it’s how people interpret MOEs in general.

 

MisterSPG says “feature” but is more like it’s not unintended and an expected response from the code, even if it can cause some issues. The main thing that we as players are arguing about is that the “feature” of being really difficult when a lot of unis play tanks is ok/not too bad when it’s one new tank but really annoying when it’s almost every tank in the game like it is now


3 MOE: Hellcat, Lycan, Sp1C, T20, M41 Bulldog, Motherland, STA 1, T95E2, Pershing, AMX 50 100, AMX 13 90, T54ltwt, T49, RU251, M46 Patton, Type 61, Leopard PTA, T54, E50, Batchat AP, Skoda T50, Centurion 7/1, Sharpshooter, T10, Conquerer, AMX 13 105, Sheridan, T-100LT, RHM PZW, M60, Chisel, M48 Patton, Leopard 1, E50M, Obj 140, Batchat 25t, TVP 50/51, AMX 50B, Valour, IS7, Super Conqueror, The Machine, Chieftain, FV215B, T110E4

 

MOE Calculation Explanation: http://forum-console...ark-difficulty/


FluffyColt12271 #11 Posted 02 October 2021 - 04:36 PM

    Staff sergeant

  • Players
  • 24507 battles
  • 465
  • Member since:
    09-01-2019

View Postnatedaishmaster, on 02 October 2021 - 03:17 PM, said:


1. I don’t think the “handicap” is specifically for maps. It’s seems like a general use modifier that can be adjusted for anything, maps were just the justification this time. It will always be there cuz statistically you can’t take the 99th or 100th percentile game because they’re outliers so you always need to go at least to the 95-97 range

 

 

Agree you need to have something less than the best game to set as the 100% target else no-one would have a chance of even approaching that value.

 

But what I'm reading from misterSPG's explanation is that *if* there was no best or worst map then the handicap would be zero - and the target would be 100% of that one dude's unicorn game. 

 

We all agree it shouldn't be left as 100%, but that doesn't mean that the correct level should be the map delta. Honestly I cant really see why that would be the right factor, as illustrated in my 1 map example.

 

The factor could be anything - like what difference you get if you use just one best game for each player in your sort - but there is no universal truth that the map delta uncovers which makes it "right".



natedaishmaster #12 Posted 02 October 2021 - 05:01 PM

    Captain

  • Players
  • 30426 battles
  • 1,935
  • [IMTLS]
  • Member since:
    07-29-2014

View PostFluffyColt12271, on 02 October 2021 - 08:36 AM, said:

 

Agree you need to have something less than the best game to set as the 100% target else no-one would have a chance of even approaching that value.

 

But what I'm reading from misterSPG's explanation is that *if* there was no best or worst map then the handicap would be zero - and the target would be 100% of that one dude's unicorn game. 

 

We all agree it shouldn't be left as 100%, but that doesn't mean that the correct level should be the map delta. Honestly I cant really see why that would be the right factor, as illustrated in my 1 map example.

 

The factor could be anything - like what difference you get if you use just one best game for each player in your sort - but there is no universal truth that the map delta uncovers which makes it "right".

I think it’s just simpler to think of it as a “game variance” handicap. If there was no variance, you could do your max damage every time. But you can’t cuz of MM, RNG, and maps. Even if there was one map, MM and RNG would still apply so you’d still have the handicap. In this case, map selection played a huge role in the “game variance” so it was used as a reason to change the handicap. For example, if +1/-1 was permanent, WG might have to make marks harder because there would be more consistency between games but the best games wouldn’t get better


3 MOE: Hellcat, Lycan, Sp1C, T20, M41 Bulldog, Motherland, STA 1, T95E2, Pershing, AMX 50 100, AMX 13 90, T54ltwt, T49, RU251, M46 Patton, Type 61, Leopard PTA, T54, E50, Batchat AP, Skoda T50, Centurion 7/1, Sharpshooter, T10, Conquerer, AMX 13 105, Sheridan, T-100LT, RHM PZW, M60, Chisel, M48 Patton, Leopard 1, E50M, Obj 140, Batchat 25t, TVP 50/51, AMX 50B, Valour, IS7, Super Conqueror, The Machine, Chieftain, FV215B, T110E4

 

MOE Calculation Explanation: http://forum-console...ark-difficulty/


Haukkis #13 Posted 02 October 2021 - 06:35 PM

    Major

  • Players
  • 41207 battles
  • 6,550
  • Member since:
    08-05-2015

View PostPhonicKitty8825, on 02 October 2021 - 09:39 AM, said:

It's just messed up. Why make it THAT hard? I hope those 10+ people with gold marks are enjoying themselves.

 

You're having some hard time trying to get those shiny stars on your IS-7, Kitty?  ;)

jujubugs #14 Posted 03 October 2021 - 12:38 AM

    Captain

  • Players
  • 23985 battles
  • 1,511
  • Member since:
    04-14-2016
It's so hard.My Tiger 1 has been on 95% forever.I managed to get it to 96% after a ton of matches.My Borsig has been stuck on 90% and now has moved down to 85%.This is painful.:(

Kaptain Proton #15 Posted 03 October 2021 - 01:15 AM

    First lieutenant

  • Players
  • 57835 battles
  • 538
  • [47R]
  • Member since:
    03-07-2016
I wonder how many there are now.    I imagine there will be at least one for each tank.   Maybe start up a 4th mark thread.  

FluffyColt12271 #16 Posted 03 October 2021 - 06:00 AM

    Staff sergeant

  • Players
  • 24507 battles
  • 465
  • Member since:
    09-01-2019

View Postnatedaishmaster, on 02 October 2021 - 05:01 PM, said:

I think it’s just simpler to think of it as a “game variance” handicap. If there was no variance, you could do your max damage every time. But you can’t cuz of MM, RNG, and maps. Even if there was one map, MM and RNG would still apply so you’d still have the handicap. In this case, map selection played a huge role in the “game variance” so it was used as a reason to change the handicap. For example, if +1/-1 was permanent, WG might have to make marks harder because there would be more consistency between games but the best games wouldn’t get better

There are lots of reasons we can't do our all-time best game every time we roll out that go way beyond MM, RNG and maps!

 

So I strongly *agree with you* that the standard for 3 marks shouldn't be for your average damage to be 95% of the absolute best game anyone played. 

 

But I disagree that "map delta" is the correct adjustment. 

 

Apparently marks were easier after 6.0 and map delta was lower but the coefficient in the handicap formula was unchanged. So handicap was updated and, hey presto, problem solved. 

 

But I'm simply saying "not so fast". Now we understand a bit more about how moe targets are set - and it is all relative to some percentile (70? 80? 90? we don't know) of the best game in the last fortnight - who is to say that the widespread abandonment of the game by good players wasn't the reason that marks became easier for those players that are left? Thus the reported "marks are easier" phenomenon? Or at least that being a significant explanatory variable.

 

Are you certain that the loss of so many good players who were routinely putting up top games didn't impact the "marks are easier" experience? I am not convinced this has even been tested. By July, were there fewer top players putting out top games? Instead, we have linked cause (map delta) and effect (easier marks) perhaps erroneously.

 

I know its easy to dismiss my "one map" example as theoretical and, in misterSPG's words, "meh", but what we have here is a basically flawed premise, that if the worst map is 20% worse than the best map (and I think it is this calculation) then the correct percentile for setting moe is 80%. 

 

I just don't think there is anything correct in this. My 1 map example is just supposed to highlight that that would produce a calculation where your EMA would have to be 100% of the best game to get 4 marks which is sort of impossible. 

 

The fact it is so extremely impossible (aside from rolling out a nerfed tank but it would have to be a mega nerf!) in that one case means that it is generally wrong in the multi-map case.  

 

It could be - by fluke - that the handicap produced by the formula is right and makes the stars align on the goldilocks difficulty. But it could also be that it is way too hard...or too easy!

 

One way of solving this would be if the formula compared your damage standing to the damage standing of players that have rolled that tank out in the past fortnight and awarded marks based on whether it was better than 65%, 85% or 95% of players. But that would be crazy!! (And would be as it is described in-game, and it wouldn't be possible to be "better than 100.0% of players" but hey ho.)



natedaishmaster #17 Posted 03 October 2021 - 04:34 PM

    Captain

  • Players
  • 30426 battles
  • 1,935
  • [IMTLS]
  • Member since:
    07-29-2014

View PostFluffyColt12271, on 02 October 2021 - 10:00 PM, said:

But I'm simply saying "not so fast". Now we understand a bit more about how moe targets are set - and it is all relative to some percentile (70? 80? 90? we don't know) of the best game in the last fortnight - who is to say that the widespread abandonment of the game by good players wasn't the reason that marks became easier for those players that are left? Thus the reported "marks are easier" phenomenon? Or at least that being a significant explanatory variable.

 

Are you certain that the loss of so many good players who were routinely putting up top games didn't impact the "marks are easier" experience? I am not convinced this has even been tested. By July, were there fewer top players putting out top games? Instead, we have linked cause (map delta) and effect (easier marks) perhaps erroneously.

 

One way of solving this would be if the formula compared your damage standing to the damage standing of players that have rolled that tank out in the past fortnight and awarded marks based on whether it was better than 65%, 85% or 95% of players. But that would be crazy!! (And would be as it is described in-game, and it wouldn't be possible to be "better than 100.0% of players" but hey ho.)

You may be completely right. Good players could’ve left and made it easier. WG would call that “working as intended” though and wouldn’t require a change. There was almost certainly a large difference in map performance but you may be right that they attributed too much of the change to that and not change in player skill overall (we don’t have the data so it’s all speculation)

 

As for the calculation, the current method is excessively simple. Load a value from an array then do division. Changing it to something complex isn’t the best solution. Plus I’m not sure changing the calculation is the correct way to go anyways since it would change the value of the marks. 
 

The general problem is that the easiest thing for WG to do is just stick it out cuz the marks will eventually drop anyways, it just result in a frustrating few months for anybody going for any mark


3 MOE: Hellcat, Lycan, Sp1C, T20, M41 Bulldog, Motherland, STA 1, T95E2, Pershing, AMX 50 100, AMX 13 90, T54ltwt, T49, RU251, M46 Patton, Type 61, Leopard PTA, T54, E50, Batchat AP, Skoda T50, Centurion 7/1, Sharpshooter, T10, Conquerer, AMX 13 105, Sheridan, T-100LT, RHM PZW, M60, Chisel, M48 Patton, Leopard 1, E50M, Obj 140, Batchat 25t, TVP 50/51, AMX 50B, Valour, IS7, Super Conqueror, The Machine, Chieftain, FV215B, T110E4

 

MOE Calculation Explanation: http://forum-console...ark-difficulty/


PhonicKitty8825 #18 Posted 03 October 2021 - 04:50 PM

    Major

  • Players
  • 27764 battles
  • 5,317
  • [USC4N]
  • Member since:
    04-10-2019

View PostHaukkis, on 02 October 2021 - 06:35 PM, said:

 

You're having some hard time trying to get those shiny stars on your IS-7, Kitty?  ;)

"Hard time" is a simple way of putting it. Was stuck at 80% and got back to 85.4% by having five 8k combined games in a row.  Now I had a bad game and I'm at 84.8%. 


4th MoE/Gold Moe  3rd MoE: IS-7, IS-3, M.O.B.A.T., M24 Chaffee '53, T-44A, T-72AV, M113, M551 Sheridan, M46A1 Patton


Pontiac Pat #19 Posted 03 October 2021 - 09:38 PM

    Major

  • WoTC Community Ambassador
  • 31524 battles
  • 8,494
  • [47R]
  • Member since:
    08-11-2014

View Postnatedaishmaster, on 01 October 2021 - 01:21 PM, said:

An interesting quality is time. Your last 100 battles as a player apply to your EMA regardless of when you played them but only battles in the last 2 weeks apply to the Target Damage. The Target Damage a month ago has nothing to do with the current target damage. Moving on 95% of the best game is the 3 MOE requirement and so on. This is obviously ridiculous as nobody can average 95% of their best game in a tank.

Great write-up Nate.  However, the underlined point is incorrect.  A given tank's EMA incorporates every battle played since they started tracking EMA data.  There is no 'drop off' of old games.  There is only the adjustment applied from the newest game.


Edited by Pontiac Pat, 03 October 2021 - 09:38 PM.

If you want to be one of the 47 Ronin check out our recruitment post and then send a message to Prof935 (xbox) or dimka935 (ps4).

Estimated Marks of Excellence thresholds - https://wotclans.com.br/tanks/moe


natedaishmaster #20 Posted 03 October 2021 - 11:06 PM

    Captain

  • Players
  • 30426 battles
  • 1,935
  • [IMTLS]
  • Member since:
    07-29-2014

View PostPontiac Pat, on 03 October 2021 - 01:38 PM, said:

Great write-up Nate.  However, the underlined point is incorrect.  A given tank's EMA incorporates every battle played since they started tracking EMA data.  There is no 'drop off' of old games.  There is only the adjustment applied from the newest game.


Lol that’s exactly what I thought too until Mister SPG said that it was exactly 100. Kinda makes sense. It’s confirmed as 100 for PC and we would get the same thing most likely


3 MOE: Hellcat, Lycan, Sp1C, T20, M41 Bulldog, Motherland, STA 1, T95E2, Pershing, AMX 50 100, AMX 13 90, T54ltwt, T49, RU251, M46 Patton, Type 61, Leopard PTA, T54, E50, Batchat AP, Skoda T50, Centurion 7/1, Sharpshooter, T10, Conquerer, AMX 13 105, Sheridan, T-100LT, RHM PZW, M60, Chisel, M48 Patton, Leopard 1, E50M, Obj 140, Batchat 25t, TVP 50/51, AMX 50B, Valour, IS7, Super Conqueror, The Machine, Chieftain, FV215B, T110E4

 

MOE Calculation Explanation: http://forum-console...ark-difficulty/





1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users