Jump to content


wotbstars first rating system attempt plans


  • Please log in to reply
392 replies to this topic

voyager35 #161 Posted 05 February 2015 - 06:55 PM

    Ilha Formosa

  • Players
  • 24428 battles
  • 403
  • [TWNO1]
  • Member since:
    06-29-2014

I totally agree with you. As what I said " If a guy can do 65% win rate without gold and platooning, I will pay him more respect than the one spam gold and do 90% platooning but has the same WR." Because I know he earned it by himself, fighting his own way out in the potato jungle. 

However, what I was trying to say is there is no good way to penalized people for platooning. People don't like to see their numbers got nerfed. Put more weight on the solo play results will be a better way, as mentioned before. 

Also, the damage should more important than WR. A potato dont get significant higher damage when they platooning. And a good player, ever when they lose, still contribute to the team. 


Edited by voyager35, 05 February 2015 - 06:57 PM.


rolleiflex #162 Posted 05 February 2015 - 07:18 PM

    Senior Sergeant

  • Players
  • 29271 battles
  • 519
  • [SKILL]
  • Member since:
    07-04-2014

The solution is simple:  Break out platooned and non-platooned stats separately:

 

Win Rate:   Platooned and non-Platooned

Avg Damage:  Platooned and non-Platooned

Survival Rate:  Platooned and non-Platooned

Tanks Spotted:  Platooned and non-Platooned

etc...

 

This should be done overall, by individual tank, and by tier.

 

Trying to create one personal number or metric to represent overall skill level is not that useful, as the amount of "normalization" that would be required would insert all kinds ways one could "game" or manipulate their numbers to create a false impression.

 

 

 


Edited by rolleiflex, 05 February 2015 - 07:19 PM.

 

US: T110E5    RU: T-62A, IS-7, IS-4, OBJ 268, OBJ 140    DE: E-100, E-50M   UK: FV215b, FV4202   JP: STB-1


SkiFletch #163 Posted 05 February 2015 - 07:24 PM

    First Sergeant

  • Players
  • 30801 battles
  • 5,911
  • [SRT]
  • Member since:
    07-07-2014

View Postrolleiflex, on 05 February 2015 - 07:18 PM, said:

The solution is simple:  Break out platooned and non-platooned stats separately:

 

Win Rate:   Platooned and non-Platooned

Avg Damage:  Platooned and non-Platooned

Survival Rate:  Platooned and non-Platooned

Tanks Spotted:  Platooned and non-Platooned

etc...

 

This should be done overall, by individual tank, and by tier.

 

Trying to create one personal number or metric to represent overall skill level is not that useful, as the amount of "normalization" that would be required would insert all kinds ways one could "game" or manipulate their numbers to create a false impression.

 

 

 

 

:facepalm: :arta:

 

Read the discussion dude.  I don't want to, but I must do it this way:  THE API DOES NOT GIVE YOU ANY OF THE DATA YOU NEED TO CALCULATE PLATOON STATISTICS.

 

There is only platoon victories.  No platoon defeats, no platoon damage, no platoon survival, etc etc.  It's just not there.



Unicorn143 #164 Posted 05 February 2015 - 07:35 PM

    First Sergeant

  • Players
  • 34842 battles
  • 2,383
  • Member since:
    04-29-2012

View Postvoyager35, on 05 February 2015 - 06:55 PM, said:

I totally agree with you. As what I said " If a guy can do 65% win rate without gold and platooning, I will pay him more respect than the one spam gold and do 90% platooning but has the same WR." Because I know he earned it by himself, fighting his own way out in the potato jungle. 

However, what I was trying to say is there is no good way to penalized people for platooning. People don't like to see their numbers got nerfed. Put more weight on the solo play results will be a better way, as mentioned before. 

Also, the damage should more important than WR. A potato dont get significant higher damage when they platooning. And a good player, ever when they lose, still contribute to the team. 

 

If a guy can do 65% win rate without gold and platooning

 

>>If a guy can do 65% win rate without gold and platooning

Spamming or not spamming gold is everyone personal choice. Spam gold but compromise your credit income. This is your decision as of individual players. We respect your decisions and we encourage to do the best of what they can to perform better on the game. If it is spamming gold, reading forums, playing the right hours, taking rest after a certain number of sequential losses. It is all up to you as long as YOU DO IT ALONE. When you bring another player into equation your stats become distorted by a presence of another player. We NEED YOUR STATS not another player stats to determine who you are and how good is your gameplay. When negating some of the platoon victories we are putting more focus on your solo performance. Your platoon victory is worth only 0.85 of what your solo victory because you did not do it yourself.

 

 

 


Edited by face4stas, 05 February 2015 - 07:35 PM.


rolleiflex #165 Posted 05 February 2015 - 07:35 PM

    Senior Sergeant

  • Players
  • 29271 battles
  • 519
  • [SKILL]
  • Member since:
    07-04-2014
Sorry about that.  I must have missed that detail.  In that case, you can't really say anything certain about the effect of platooning, other than we suspect it helps an already above average player achieve an even higher win rate by reducing the uncertainty of the match making.

 

US: T110E5    RU: T-62A, IS-7, IS-4, OBJ 268, OBJ 140    DE: E-100, E-50M   UK: FV215b, FV4202   JP: STB-1


Asiaticus #166 Posted 05 February 2015 - 07:40 PM

    First Sergeant

  • Players
  • 16340 battles
  • 2,871
  • Member since:
    07-19-2014

Now that the system has a large sample, there is one calculation I'd be interested in: correlation between Platoon/total victory % and WR. This would need to be calculating using the whole sample, not per player, but it would use data revealed by the API. The correlation is almost certainly positive, but I wonder exactly how much it contributes.

 

(For those of you who don't know stats, correlation measures how two sets of numbers vary together. It ranges from -1.0 to +1.0, with 0 meaning the sets don't relate to each other, negative numbers indicating that as the first goes up the second goes down, and positive indicating that as the first goes up the second goes up.)



Unicorn143 #167 Posted 05 February 2015 - 07:44 PM

    First Sergeant

  • Players
  • 34842 battles
  • 2,383
  • Member since:
    04-29-2012

View PostAsiaticus, on 05 February 2015 - 07:40 PM, said:

Now that the system has a large sample, there is one calculation I'd be interested in: correlation between Platoon/total victory % and WR. This would need to be calculating using the whole sample, not per player, but it would use data revealed by the API. The correlation is almost certainly positive, but I wonder exactly how much it contributes.

 

(For those of you who don't know stats, correlation measures how two sets of numbers vary together. It ranges from -1.0 to +1.0, with 0 meaning the sets don't relate to each other, negative numbers indicating that as the first goes up the second goes down, and positive indicating that as the first goes up the second goes up.)

 

Yes I will try to do something to compare average performance of the players who have

0-10% of platoon victories

10-20% of platoon victories

etc

will see how those groups stay against each other


Edited by face4stas, 05 February 2015 - 07:45 PM.


voyager35 #168 Posted 05 February 2015 - 07:48 PM

    Ilha Formosa

  • Players
  • 24428 battles
  • 403
  • [TWNO1]
  • Member since:
    06-29-2014

 View Postface4stas, on 05 February 2015 - 01:35 PM, said:

 

If a guy can do 65% win rate without gold and platooning

 

>>If a guy can do 65% win rate without gold and platooning

Spamming or not spamming gold is everyone personal choice. Spam gold but compromise your credit income. This is your decision as of individual players. We respect your decisions and we encourage to do the best of what they can to perform better on the game. If it is spamming gold, reading forums, playing the right hours, taking rest after a certain number of sequential losses. It is all up to you as long as YOU DO IT ALONE. When you bring another player into equation your stats become distorted by a presence of another player. We NEED YOUR STATS not another player stats to determine who you are and how good is your gameplay. When negating some of the platoon victories we are putting more focus on your solo performance. Your platoon victory is worth only 0.85 of what your solo victory because you did not do it yourself.

 

 

 

I'm not saying you do it wrong, I agree we should have different ways to evaluate the performance of solo and platoon fighting. 

My point is making solo play 15% more valuable is better than making platoon play 15% less. 

To be honest, this is your work, so no one, including me, have a right to tell you what to do. Just do whatever you think is the right thing to do. For the one disagree with you, they can come up with their system.



SkiFletch #169 Posted 05 February 2015 - 07:51 PM

    First Sergeant

  • Players
  • 30801 battles
  • 5,911
  • [SRT]
  • Member since:
    07-07-2014

View Postrolleiflex, on 05 February 2015 - 07:35 PM, said:

Sorry about that.  I must have missed that detail.  In that case, you can't really say anything certain about the effect of platooning, other than we suspect it helps an already above average player achieve an even higher win rate by reducing the uncertainty of the match making.

 

Sorry about the caps :).  It's like the umpteenth time It's been brought up.  You're right though that we only suspect an effect on stats.  I still think normalized statistics would be accounting for both the potato tooners and the unicum tooners and even out.  I'll keep manually collecting the data I can regarding this.


sheriffsherman #170 Posted 05 February 2015 - 07:52 PM

    First Sergeant

  • Players
  • 24534 battles
  • 4,674
  • Member since:
    10-06-2014
Ok you guyz are creepin me out a little.  I mean, i knew their were really smart ppl on these forums, but sheesh!  This is bordering on 'big bang theory' type stuff:)  i have a degree from a UC system school, cough cough and i'm feeling really dumb reading this thread:)

FIRST TO 200 TANKS ON BLITZ!!!

I'm a gen 1 gamer, been crushing bad guys since the 1970s


mbrcmf #171 Posted 05 February 2015 - 09:06 PM

    Senior Sergeant

  • Players
  • 23268 battles
  • 669
  • Member since:
    12-22-2014

View Postface4stas, on 05 February 2015 - 07:44 PM, said:

 

Yes I will try to do something to compare average performance of the players who have

0-10% of platoon victories

10-20% of platoon victories

etc

will see how those groups stay against each other

And once you have the correlation (if there is one) ... An equation can also be calculated to determine if it is linear /non - linear etc, and give you the offset for wr% (platoon, non). Realizing that wr‰ is not the only variable in your "BPI" - Blitz performance indicator. 



mehliveat #172 Posted 05 February 2015 - 09:51 PM

    First Sergeant

  • Players
  • 300 battles
  • 3,196
  • Member since:
    08-05-2014

View Postoriginalmadkilla, on 06 February 2015 - 04:55 AM, said:

Well said! I think that If I could duplicate myself and have two of myself in every match, then I have no doubt that my win rate would increase significantly, probably anywhere from 15%-30%, and the same goes for many other solo only players.

 

Certain platooners here are simply not being objective or honest with themselves.

 

Personally, I don't platoon that much, only occasionally when another forumer ask me.  But I have found that some people you work really well together, while others its just like a random game.  If I kept on playing with people that I work well with, then I can see my WR jump.  If its the other group then it will stay around the same.  Therefore I agree with Asia below

 

View PostAsiaticus, on 06 February 2015 - 05:40 AM, said:

Now that the system has a large sample, there is one calculation I'd be interested in: correlation between Platoon/total victory % and WR. This would need to be calculating using the whole sample, not per player, but it would use data revealed by the API. The correlation is almost certainly positive, but I wonder exactly how much it contributes.

 

(For those of you who don't know stats, correlation measures how two sets of numbers vary together. It ranges from -1.0 to +1.0, with 0 meaning the sets don't relate to each other, negative numbers indicating that as the first goes up the second goes down, and positive indicating that as the first goes up the second goes up.)

 

We need to find the correlation between platoon/total victory and WR.  Even if its just a quick plot of Platoons wins vs WR on a graph.  Then we can clearly see the general trend.

 

View Post1991Syclone, on 05 February 2015 - 11:32 PM, said:

All this talk of platooning.  Look at the existing WN8 information for the desktop version - http://wiki.wneffici....net/pages/WN8  Notice that nowhere does it mention taking platooning into account.  Yes, platooning can increase your win rate, but I think the majority of players in this game do not platoon. 

 

Just because WN8 doesn't take platooning into account, doesn't mean we shouldn't here.  The two games are quite different.  For one a platoon of 2 in a 7 player game is 28.5% of the team's firepower, while a platoon of 3 in a 15 player game is 20% of the team's fire power.  Thats quite a difference.


Edited by mehliveat, 05 February 2015 - 09:55 PM.


GandalfTehGray #173 Posted 05 February 2015 - 10:33 PM

    Sergeant

  • Players
  • 1884 battles
  • 261
  • Member since:
    05-10-2014
On xbox we have had five man platoons for a few months now in normal mm, never have I seen anyone want to degress anything stat related. My dc/r there has stayed steady basically since I joined a clan while before it was rising. My survival ratio is also dropping because I don't have to play as conservatively due to having reliable teammate,  but even with 33% it is not a auto win button that magically inflates my k/d and dc/r, reliable platoonmmates most times makes it harder to keep good ratios high.

A wizard is never late Frodo Baggins, nor is he early. He arrives precisely when he means to.

Current Grinds: Comet ,IS, T29, M4E8  , Jackson, T-34-85

Premium Tanks: T34 , IS6 , T25 , Ltp , and Box Tank

Seal Clubbers: T18, Hetzer, KV1, Micro Maus


mehliveat #174 Posted 05 February 2015 - 10:47 PM

    First Sergeant

  • Players
  • 300 battles
  • 3,196
  • Member since:
    08-05-2014

View PostGandalfTehGray, on 06 February 2015 - 08:33 AM, said:

On xbox we have had five man platoons for a few months now in normal mm, never have I seen anyone want to degress anything stat related. My dc/r there has stayed steady basically since I joined a clan while before it was rising. My survival ratio is also dropping because I don't have to play as conservatively due to having reliable teammate,  but even with 33% it is not a auto win button that magically inflates my k/d and dc/r, reliable platoonmmates most times makes it harder to keep good ratios high.

 

My point was mainly to highlight that the games are different, and we should not be taking WN8 as gospel.

 

Another difference is the type of players you get on each platform.  Mobile platforms will attract a lot more pubs than XBox or PC, therefore it is likely the difference between platoon players and random players is not as much as on Blitz.  

 

Going back to what Asiaticus suggested previously, we should try to see if there is a correlation between WR and platoon wins.  Then we can definitively say one way or the other about the effects of platooning on WR.  My guess is that there is a positive correlation, but how significant?  We won't know till the data is out.



warko_1 #175 Posted 05 February 2015 - 11:47 PM

    Senior Sergeant

  • Players
  • 19518 battles
  • 571
  • [SRT_]
  • Member since:
    07-09-2014
Isn't there a statistic that tells the number of games played platooned? I'm guessing no, otherwise we wouldn't be having this debate. If we had that statistic then it would be just a matter of WR platooned vs WR solo. Problem solved. WG must have that data. They get every possible statistic. They should make it available. 

Edited by warko_1, 05 February 2015 - 11:48 PM.


LtRaseiniai #176 Posted 06 February 2015 - 12:04 AM

    First Sergeant

  • Players
  • 76452 battles
  • 4,417
  • [EWENS]
  • Member since:
    06-29-2014
Everyday of the week someone posts a version of the question, how can i increase my win rate.  And the answer the comes backeverday of the week is platoon, platoon, platoon, platoon, platoon, platoon, platoon.  It's not true of course.  Some people become very good players and never platoon.  All of the platooning, clans, and teamspeaking that has gone on has been for the purpose of inflating win-rates.  And inflate them it has done.    .3 was the right rate to correct for this inflation.

johndd1_winner #177 Posted 06 February 2015 - 12:42 AM

    Sergeant

  • Players
  • 14756 battles
  • 496
  • [GRIM]
  • Member since:
    07-08-2014

View Postthieldke, on 06 February 2015 - 12:04 AM, said:

Everyday of the week someone posts a version of the question, how can i increase my win rate.  And the answer the comes backeverday of the week is platoon, platoon, platoon, platoon, platoon, platoon, platoon.  It's not true of course.  Some people become very good players and never platoon.  All of the platooning, clans, and teamspeaking that has gone on has been for the purpose of inflating win-rates.  And inflate them it has done.    .3 was the right rate to correct for this inflation.

 

You made a decent point, right up until the last couple of sentences.

All the platooning, clanning and TSing, as you put it, has NOT been for the purpose of WR inflation. I'd argue it is largely done for enjoyment. Solo pubbing in the current player base environment...just...sucks. Platooning, with or without TS provides an extra layer of enjoyment. If I have to play tank x, which I dislike playing, I might as well play it with someone else.

 

No one that I know in this game has ever said (to me) "I want to platoon to get my WR up." It's always been more of a misery loves company approach.

 

No doubt, platooning has been put forward as one of the best way to increase your WR. But, two taters in a platoon are still two taters in a platoon. You must still be good to have a good WR, platooned or not. Further, platoons don't fight in a vacuum. There are 7 players in the other team, and 5 more on yours, who all have a say in the outcome of the battle. Even when tooned, you must still play well. 

 

Finally, since we have no idea how much impact, inflation as you call it, platooning has had on WR, I'm curious to know how ".3" is, in fact, the "right" rate to correct for an unknown number?

 

I solo pub FAR more than I platoon, so I really don't have a dog in this fight, personally. I just fail to see the logic in applying arbitrary "corrections" to an unknown value. Especially when the "inflation" of that value is available for all the players to enjoy. It's not like the ability to platoon is only open to "good" players. Anyone can do it. If you think it makes THAT much of a difference, then platoon 100% of your games. 

 

Even doing so, many of us will fail to increase our WR by any significant number.


You're just a bag of particles obeying the laws of physics

 

 

 


acrisis #178 Posted 06 February 2015 - 01:41 AM

    Captain

  • Players
  • 21028 battles
  • 12,626
  • [III-C]
  • Member since:
    12-27-2014

Had the luxury of having a super unicum (per blitz rating) on my team, and he carried (solo)!

His (assuming) personal xp rating the same as his several thousand battle count! 

1/3 platoon victories ...

 

Impossible to quantify without more data released by WG, but still that's a whole lot of win without platooning. Win rate per tank higher than anyone I've seen and even perfect. In two months time! Holy smokes.

 

I need some of what he's eating and drinking ;)
 


 

Forum essentials:  > Desktop mode 
Check recent threads. Check WG staff threads. Use search.
    > WG Bugs and current update     > Feedback     > Vehicle Bay     > Guides     > Videos 

 

     BE  (+)   

 [ NEW > ]  Lag, Packet Loss, Ghost Shells, Networking and more demystified usingPingPlotter (free)  
Blitz Community Coalition | Looney Tooners | Triarii
Blitz University | Basic Training  |  Mentor  |  The Pingplotter Guy

 


johndd1_winner #179 Posted 06 February 2015 - 02:18 AM

    Sergeant

  • Players
  • 14756 battles
  • 496
  • [GRIM]
  • Member since:
    07-08-2014

 

Any stats analysis is only as good as the data available. No rating system will be perfect, and there will always be ways people attempt to pad the system.

 

Is a 70% WR player really that good, or has he simply played 5000 games in the T18? Is that 2.5 damage ratio player that good, or does he only play the Hetzer? I'd doubt anyone on these forums would ever judge a players skill based only on a single stat. You would review that players entire profile, taking all the numbers in together to form an opinion.

 

For the same reason, WN8, or Star1, or any other rating system will never tell the whole story. But WN8 is the culmination of a series of attempts to implement a rating system. Is WN8 perfect? No, but it's about the best we can do with the data available.

 

My recommendation would be to simply use the data that is available, without feeling the need to try to correct for what you/we do not know. When you release your rating system, simply acknowledge the possible issues with the data so the users can make their own judgement on the systems value. "This is the system I developed. These are the formulas I used, and this is my reasoning behind it".

 

That's about the best you can do, and reasonable people will understand that. Making things up, such as platoon correction will only invalidate your efforts in many players eyes. 


You're just a bag of particles obeying the laws of physics

 

 

 


acrisis #180 Posted 06 February 2015 - 02:29 AM

    Captain

  • Players
  • 21028 battles
  • 12,626
  • [III-C]
  • Member since:
    12-27-2014

I agree to focus on determining the " regular " number first. That'll be a lot of work in itself to come up with the formula.  

 

In the case of the super unicum I saw active, he was very well rounded player mostly active in the top tiers. Least of activity in the lowest tiers. 

As opposed to the sealclubbers. I've seen a few of those and they'll have an equally high xp rating in the few or several thousands, but 3-4-... times the battle count of their rating; and they'll drive one or two TD in low tier, maybe one in mid tier.  


 

Forum essentials:  > Desktop mode 
Check recent threads. Check WG staff threads. Use search.
    > WG Bugs and current update     > Feedback     > Vehicle Bay     > Guides     > Videos 

 

     BE  (+)   

 [ NEW > ]  Lag, Packet Loss, Ghost Shells, Networking and more demystified usingPingPlotter (free)  
Blitz Community Coalition | Looney Tooners | Triarii
Blitz University | Basic Training  |  Mentor  |  The Pingplotter Guy

 





3 user(s) are reading this topic

0 members, 3 guests, 0 anonymous users