Jump to content
GloatingSwine

New global rating system. Pad hard!

Recommended Posts

4702 here.

 

As I said on the WoT forum, the rating actually seems to be considerably better than expected (though, I went in with very low expectations, so that's damning with faint praise).  I've noticed some oddities looking at comparison scores, but I suspect we'll need to see a comparison on a larger scale than looking at a few isolated numbers.

 

It is unfairly brutal on low battle accounts, which is going to limit it severely.  At this point, I would say "better than expected, worse than the WNx system", but beyond that? Time will tell.

Share this post


Link to post
Share on other sites

Its not that the new rating uses battles to increase the number, but rather uses battles to decrease it.

 

Awesome players with few battles = shitfucked

Crap players with battles  above the negative coefficient = godmode compared to person above

 

awesome players with many battles = godmode

crap players with uber many battles = crap compared to person above. stats likely still inflated somewhat due to battle count. im sure there is a slight boost, but it should be negligible after whatever number they have set. probably 10%, like wn counts wins.

 

 

The meeting point might something around 10k? After that WGs system stops dragging your number down due to low battle numbers and it actually uses the rest of the stats

 

It's like the low tier detriment works with wn6/7 and will be further increased in wn8. Only wg used battle count instead of tiers (they love them seal clubbers)

 

Im just amused that they refuse to allow the use of spotting damage. Are they actively trying to make scouting pointless? Seems to be given how games play out lately

Share this post


Link to post
Share on other sites

The ideal for using battle count would be to use it as a confidence predictor.  Decide what the averages for each stat spit out in your formula to work out what the rating of the average 49%er is, give every starting player that number as a starting rating, and use the number of battles to define a maximum deviation away from the average rating (up to, say 3k battles at which it uncaps). 

 

Players with low battle counts are basically treated as average by default because their stats are based on too low a sample to decide how good they actually are, and besides they're still learning how to play.

Share this post


Link to post
Share on other sites

High number of battles prevents from rerolling -> that means a good player that plays with he's first account has to deal with he's crappy stats from the beginning, where rerolls don't. So the battle number is perhaps a good thing to avoid this.

 

Dud still not taken into account >:|

 

And what about average xp per battle ? Is it based on total/nb battles which means if you play a lot of different tanks you can earn more xp due to x2 x3 or x5 for first win ? raw or premium xp ?

Does someone have these values to verify the avg xp ?

 

 

 

 

BTW : I'm happy, I achieved my first 11 kills in a game with my t49 =D

Share this post


Link to post
Share on other sites

Yea my rating under this WG system is pretty bad, because of my low number of battles. Even though I am not a re-roll, just an account that didnt play for a good while and is just now starting to play again, and not even alot these days.

 

Even though Im a decent player, I am considered bad by the rating.

 

A system that counts:

  • Battles
  • Victories
  • Games survived

 

Is generally flawed towards players who havent played alot of battles. What it should count is:

 

  • Winrate (%)
  • Victories (%)
  • Battles Survived (%)

Not mean numbers, but percentages would make it more accurate.

Share this post


Link to post
Share on other sites

They've released the formula for this now.

 

The variables are:

 

Number of Battles (this is a penalty factor by which your whole score is multiplied, the penalty reduces exponentially with increasing battles, 5k battles gives a 50% penalty, 10k gives a 20%, 15k gives roughly 7%, seems to exist to fight rerolls)

 

Win rate  

 

Survival rate (only counts battles where you win and survive, the arty padder's friend)

 

Hit rate

 

Average XP (non-premium, will only count XP since 8.8 release)

 

Average Damage

Share this post


Link to post
Share on other sites

They've released the formula for this now.

 

The variables are:

 

Number of Battles (Im screwed)

 

Win rate  (Good)

 

Survival rate (Meh, dieing for a win never hurt my feelings before. Now it will hurt my WGR)

 

Hit rate (LOL)

 

Average XP (Good)

 

Average Damage (Good)

 

Share this post


Link to post
Share on other sites

Looked up an upper-medium volume shit player, 2333

Looked up a medium volume average player, 3695.

 

4k might be the new threshold for "above average/competent."

 

I think I'm at somewhere around 2900...

And after last night, I wouldn't argue with this statement either >.>

 

Share this post


Link to post
Share on other sites

The battle component count is annoying, but won't make the whole thing completely useless.  Your score is deflated 30% at 5k battles, 10% at 10k battles, and 1% at 20k battles.  The factor is small enough that you'll still be able to distinguish between scrub and skill.

 

The hit ratio is annoying.  You get zero points for a 45% hit rate, up to 3300 points for a 100% hit rate (linear).  So for the padder, no more clearing cover or taking low-probability low-risk shots.  But practically speaking, with a typical hit ration between 65% to 80%, there is only 900 points of variability between scores.

 

The battle count in 8.8 component is just crap.  You lose ~2000 WGR (assuming ~500 avg 8.8 xp) if you stop playing for a few weeks?  That's some trollish shit right there.

Share this post


Link to post
Share on other sites

My biggest gripe with hit rate is how much low tier autocannons can screw it up.  I have ~350 games in a Panzer II.  Around 2.3% of my total games played.  This vehicle has fired ~17.5% of my total shots, however.  That's twice as many shots as all of my heavies combined.  So this one tank is twice as influential as an entire class of tanks. 

 

Maybe part of the post 8.8 pad hard strategy will involve my PZ IC getting behind heavies and dumping a few hundred rounds in. 

Share this post


Link to post
Share on other sites

My biggest gripe with hit rate is how much low tier autocannons can screw it up.  I have ~350 games in a Panzer II.  Around 2.3% of my total games played.  This vehicle has fired ~17.5% of my total shots, however.  That's twice as many shots as all of my heavies combined.  So this one tank is twice as influential as an entire class of tanks. 

 

Maybe part of the post 8.8 pad hard strategy will involve my PZ IC getting behind heavies and dumping a few hundred rounds in. 

 

If you make any conscious effort to pad your WG rating, I will campaign to sanction you from these forums, as well as de-friend you in game. And I will probably cry a little over it as well!

Share this post


Link to post
Share on other sites

My biggest gripe with hit rate is how much low tier autocannons can screw it up.  I have ~350 games in a Panzer II.  Around 2.3% of my total games played.  This vehicle has fired ~17.5% of my total shots, however.  That's twice as many shots as all of my heavies combined.  So this one tank is twice as influential as an entire class of tanks. 

 

Maybe part of the post 8.8 pad hard strategy will involve my PZ IC getting behind heavies and dumping a few hundred rounds in. 

 

You made me laugh.

 

Hit ratio is indeed a bad indicator of skill. Obviously low tier machine guns screw it up. But even if those were excluded it would be a bad indicator. Some players like to take snapshots others only pull the trigger when they have a safe shot lined up. Often it is the smart choice to do the former.

Share this post


Link to post
Share on other sites

But even if those were excluded it would be a bad indicator. Some players like to take snapshots others only pull the trigger when they have a safe shot lined up. Often it is the smart choice to do the former.

 

Don't write it off.  If you lob 5% of your shells into mud huts, your WGR drops 300 points.  If, as a result, your win rate and survival rate increase by 1%, you gain 300 points. 

 

Funny how that math lined up nicely...

Share this post


Link to post
Share on other sites

I think the rating is shit and useless. Let me explain.

A performance rating in wot has couple of purposes

1) determine the level of skill of a player

The rating therefore is supposed to tell how good someone is at tenks. Not how well the player's history and gameplay correlates with some ideological economical idea of how to play wot in the "wg correct way". A rating should do its best to give the most accurate rating for every player regardless of how many battles he has fought for example. Not to punish rerolls just for the sake of punishing rerolls.

What does wg do wrong here? A TON of stuff. First of all they make the rating less accurate on purpose by penalizing people with low amount of games. Even at 10k battles you are looking a negative multiplier. Why? Because for no reason at all wg wants to make the rating inaccurate for players with low number of battles. There is absolutely no reason to do that at all. Instead wg should focus to make the rating more accurate for people with high battle counts to show their current level of skill that is now being averaged and reduced over all battles, most of which do not correlate well with the current level of skill the player has.

2) show that rating to others

Because of 1. the rating is useless. It can not be used as a tool to determine who is dangerous and who is not because it is wildly inaccurate. Instead of the system giving huge penalties to rerolls the system should seek to be so accurate that your reroll account (with similar tanks as your main account) should have identical ratings with your main account. Why? Because essentially an account is one person and that rating is supposed to show how good that person is at tenks.

As a result you can not trust the rating to show who is the dangerous player in the enemy team and who is the competent player in your team

3) the rating emphasizes some ideological wargamering way of playing tenks instead of even trying to determine the skill of the player

They want you to play high number of battles to get higher rating. Why? This does nothing good and only makes your rating go up despite you playing on the same level from your first game until your 20th thousandth game. It gives false sense of improvement to the players. Wargamering doesn't want the players to improve - they want you to play more. With the rating they want to reward players with lots of battles instead of improvement in skill.

4) the rating is a tool for monitoring progess and improvement

No. Going from 5k to 10k battles you will get automatic 50% increase to your rating just to reward non-improvement. If your rating gets bigger it does not mean you have improved at all. That makes this rating very bad tool to monitor your progress because the rating is designed to mask your progress. Looking and comparing the rating to your older ratings tells you absolutely nothing whether you have improved, gotten worse or plateaued.

In the last 5k games I've probably improved my wn for about 400 points by playing better and better. From 1100 to 1500. With this wg rating I would have basically gotten relatively similar size increase to my wg rating automatically without learning anything at all, without improving at all. This is not just insanely stupid mechanic in the rating but it is also detrimental to the game. It makes people believe that they are improving because after each battle their rating is little higher just because of higher total battle count. And because it makes people believe they are improving it makes people less likely to put any effort into actual improvement.

With the info wg has I'm amazed how shitty job they once again managed to pull off.

Share this post


Link to post
Share on other sites

Less then 10k battles: check!

Taking a lot of farfetched shots when not in immediate need: check!

Going for the win even if it means dying a lot: check!

 

I think I need to prepare myself for a lot of noobcalling if these stats will ever been shown ingame :smug:

Share this post


Link to post
Share on other sites

It looks like the formula for the new PR rating is out on ftr. 

 

http://ftr.wot-news.com/2013/09/12/new-wg-personal-rating-analysis/#more-3074

 

1dts.jpg

 

It looks like the most rewarding aspects that will contribute to better ratings are ( in the order of greatest to least impact ) :

  • non-prem account experience per game 
  • Survival rate
  • hit rate
  • Damage

Battle count only have a significant enough effect on this rating if under ~10k games played (any more beyond that and the coefficient flattens out to 1), and win rate has the highest effect on the formula when player's win rate is around 50%

 

I think this is actually pretty good. Its not tank specific though so there is still something to wish for..

 

Thoughts? 

Share this post


Link to post
Share on other sites

I like the idea of using non-premium exp as a heavy factor. I find that it tends to correlate very well with how well I did in a match. It also has the unique property of being able to take into account spotting damage, which is in my opinion is WN's greatest shortcoming. 

 

I was going to write something about how I wish they would normalize experience gained to tier, because that heavily biases it towards tiers 7-9, but the more I thought about it the more I can understand the argument for why you wouldn't do that. After all, I agree with the idea that it is more difficult to play well as a high tier tank where the game is more complex, than as a low to mid tier tank. It inherently prevents seal clubbers from getting high scores.

 

The adjusting for battles played is pretty stupid, but at least its effect has been reduced. Win rate I don't particularly like being factored in, as exp gained does that by its very nature. I've never liked win rate being factored into these kind of things because of all the legitimate things you can control, win rate is the most easily affected through the use of platoons and TC.

 

Hit rate is stupid and shouldn't be included at all. Survival rate I can see the argument for keeping in. I don't care especially about its inclusion, but I don't think it's needed.

 

Share this post


Link to post
Share on other sites

I like the idea of using non-premium exp as a heavy factor. I find that it tends to correlate very well with how well I did in a match. It also has the unique property of being able to take into account spotting damage, which is in my opinion is WN's greatest shortcoming. 

...

 

avg XP is flawed in its current form. Victory and team bonus makes it useless.

 

I just have to compare my T32 and KingTiger stats and avg XP. T32 has much higher avg XP because I played often in company (winning more often -> higer avg XP), but KT has better stats (Kills, Damage + Spots).

Share this post


Link to post
Share on other sites

If anyone hasn't noticed it yet you can calculate your average (non-premium) exp by reverse-engineering this rating.

I'll maybe post my excel sheet later after I have downloaded 8.8 and checked it shows the correct numbers...

edit: It takes some time before the average exp is useful though because wg starts "recording it" in 8.8. So to get that average exp number later you'd need to memorize what is your total number of battles and average exp at that moment when 8.8 is deployed on your server.

It would be useful if stat sites stored the battles played and average exp numbers before 8.8 for every player so in future we could use the average exp value for other ratings maybe.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...