Jump to content

Recommended Posts

Hibachi is wotstatgrabber 64 bit? I get an error when trying to execute it on a 32bit win XP PC in the lab... I have w7 at home, so ill try it there later.

 

Just replied to your PM then saw this, yeah the executable is 64 bit.  I'll need to get a 32 bit environment of Python set up (without stomping on my 64 bit install) to generate a 32 bit exe.  If I have some time tomorrow I'll try to do that, otherwise it will be next week.

 

New 64 bit version using top 5 most played low tier tanks for the low tier battles is up though: https://www.dropbox.com/s/afkooktz7x889i3/WotStatGrabber%200.1%20alpha2.zip

 

I think I'll need to make a separate thread for the script soon, anyone know what section that should go in?

Link to post
Share on other sites

Running at 1710 WN8+norm, down from 1735 w/WN7, with a 43.5 point tier malus. Seems to show that my stats are close to unpadded, which I'm happy with, but it's hard to make any assertions as to whether the formula itself is more accurate; this is probably something I'm going to need 60/30/7 day examples of to see.

Link to post
Share on other sites

Gameplay discussion?

 

Rexxie, once everything is up and running correctly I will conduct extensive testing. Nevertheless, keep in mind WN8 is pretty similar to WN7 aside from def, damg and the normalizer term.

Link to post
Share on other sites

The normalizer is kinda working, but the limits seem to be off... I´ll take another look at the formula. People with 22k games and 150 on their top tier 1-3s should be reciveving no normalizer, but they are getting something like -15 for some reason... 

 

battlesNormBattles and battlesNormTier seem spot on though, so it might be a mistake on my side...

Link to post
Share on other sites

You will end up losing a lot of support for WN if you put it through with the normalizer.  Especially with ClanWars now requiring Tier 1's and 3's.

 

Instead of penalizing for tier 3 and below, why not IGNORE stats for tier 3 and below?  It gets rid of any 'penalties' for CW players who have to play those tanks, players who have certain premiums for crew training (FCM36 PaK40 anyone?), and only really 'hurts' players with a huge percentage of their games at those tiers.

Link to post
Share on other sites

Instead of penalizing for tier 3 and below, why not IGNORE stats for tier 3 and below?

 

For christ's sake that's EXACTLY what he's doing!

 

Has anyone read what Praetor's been saying about the normalizer?

 

It's purpose is to estimate what an individual's overall WN8 would be had they had not clubbed seals all those times in their low tier tanks; perhaps it's a little harsh in its current implementation, but it's also fairly accurate, as his posts about ITDude/Pucara show on the regular WoT forums.

 

You cannot directly remove stats for tier 3 and below due to API limitations.

 

This is the next best thing, and it's only natural that some individuals will be hit harshly by it, since the WN7 that they see is heavily padded by the low tier tanks that they play.

Link to post
Share on other sites

Not much more to add to precambrian's post, although the normalizer DOES seem to be applied to players that it shouldnt, and that might also lower the normalizer for players that SHOULD be normalized, like pucara, and ITDUDE.

 

PP, I will go through with it because the numbers show the normalizer is pretty accurate. Obviously people who have padded their stats with low tier sealbashing will be pissed off, but honestly, I don´t really care. The same thing happened when WN was released with no cap points, and everyone with high average cap points was whining and doing everything possible to have me include it in the formula although the data shows cap points have almost no correlation with winning and/or player skill.

 

For the normalizer, the numbers show it works pretty well, actually it works awesomely for such a simple calculation. It has a +-100 error approximately which is pretty small. Also, the normalizer will be recalibrated if needed, since I want to play on the safe side, and give people the benefit of doubt when possible (either due to small number of total games, or to relatively small number of games on tier 1/3).

Link to post
Share on other sites

For christ's sake that's EXACTLY what he's doing!

 

Has anyone read what Praetor's been saying about the normalizer?

 

It's purpose is to estimate what an individual's overall WN8 would be had they had not clubbed seals all those times in their low tier tanks; perhaps it's a little harsh in its current implementation, but it's also fairly accurate, as his posts about ITDude/Pucara show on the regular WoT forums.

 

You cannot directly remove stats for tier 3 and below due to API limitations.

 

This is the next best thing, and it's only natural that some individuals will be hit harshly by it, since the WN7 that they see is heavily padded by the low tier tanks that they play.

 

Yes, I've read, and it DOES NOT ignore tier 3 and below, it severely penalizes you for every game played at tier 3 and below.

Link to post
Share on other sites

Not much more to add to precambrian's post, although the normalizer DOES seem to be applied to players that it shouldnt, and that might also lower the normalizer for players that SHOULD be normalized, like pucara, and ITDUDE.

 

PP, I will go through with it because the numbers show the normalizer is pretty accurate. Obviously people who have padded their stats with low tier sealbashing will be pissed off, but honestly, I don´t really care. The same thing happened when WN was released with no cap points, and everyone with high average cap points was whining and doing everything possible to have me include it in the formula although the data shows cap points have almost no correlation with winning and/or player skill.

 

For the normalizer, the numbers show it works pretty well, actually it works awesomely for such a simple calculation. It has a +-100 error approximately which is pretty small. Also, the normalizer will be recalibrated if needed, since I want to play on the safe side, and give people the benefit of doubt when possible (either due to small number of total games, or to relatively small number of games on tier 1/3).

 

A measurement stat which gives penalties based on tier is not a good measurement.  

 

There are many who will be dinged heavily by this that DID NOT 'seal club'.  All this looks like - to many people, and this is sentiment i've seen from many players from many top clans - is a personal vendetta and hatred against those who do not play the same as you.

Link to post
Share on other sites

Yes, I've read, and it DOES NOT ignore tier 3 and below, it severely penalizes you for every game played at tier 3 and below.

 

You cannot actually be this stupid. I'll reiterate this fairly obnoxiously for you:

 

IT IS IMPOSSIBLE TO IGNORE STATS FROM ANY GROUP OF TANKS DUE TO LIMITATIONS IN WARGAMING'S API

 

This is the next best thing. Your WN7 is inflated because of the games you played at tier 3 and below; removing those games from consideration reduces it to around the level predicted by the normalizer.

Link to post
Share on other sites

You cannot actually be this stupid. I'll reiterate this fairly obnoxiously for you:

 

 

This is the next best thing. Your WN7 is inflated because of the games you played at tier 3 and below; removing those games from consideration reduces it to around the level predicted by the normalizer.

 

Like I said, WN will lose TONS of support once it is released as version 8+norm.  Many clan wars clans/players will drop it once they realize that they will be penalized for playing clanwars.  As I've said elsewhere, this looks like a personal vendetta against play below tier 8, smells like it, and quacks like it.

Link to post
Share on other sites

Yes, I've been hugely penalized as a clan wars player for playing clan wars, because obviously, the ten or so games in a week, compared to the two hundred other games that I play, makes up an inane percentage of my games and will result in an enormous penalty.

 

Seriously, don't talk about something you know nothing about. Play in serious clan wars before you preach about it.

Link to post
Share on other sites

For the win ratio part of the equation why don't you exclude low tiers from the win ratio? As far as I know the api does include win ratios for each tank and the number of games played for each tank as well. At least xvm can show this info. Using that info you could simply calculate the total win ratio for tiers 4-10 which would make the win ratio part of the equation more precise.

Link to post
Share on other sites

Praetor I've updated the script (also made a thread for it) and added 32bit and 64bit downloads here: 

 

This version lets you specify the number of tanks used for the low tier battles calculation, fixes the bug you noticed with said calculation, and adds the sum of low tiers value you were requesting.  Let me know if anything seems off.

Link to post
Share on other sites

Tango you are right, that would probably make WN slightly more precise. And I say slightly because I must remind you the winrate portion is  used as a proxy here, and platooning can affect WR probably much more than low tier winrates. Also, the incremented winrate for low tiers are included into the normalizer term, so it would be kind of like overlapping factors.

 

It would also increase formula complexity for a small gain, IMHO.

 

Hibachi you are awesome!  :thumbup:

Link to post
Share on other sites

Orrie I have an idea what is wrong. This:

 

-Math.abs(Math.max(2200/battles*(battlesNormTop-battlesNorm*83/(1+Math.pow((0.15-0.000072*battles),10))),0)),

 

Should be this:

 

-Math.abs(Math.max(2200/battles*(battlesNormTop-battlesNorm*83/(1+Math.pow(10,(0.15-0.000072*battles)))),0)),

 

The Math.pow number and exponent were backwards.

 

Also, lets change the normalizer multiplier from 2200 to 1700. This would be the formula to use in the script:

 

-Math.abs(Math.max(1700/battles*(battlesNormTop-battlesNorm*83/(1+Math.pow(10,(0.15-0.000072*battles)))),0)),

Link to post
Share on other sites

This isn´t stat denial? Quoting private_public from official forums...

 

 

You are still fixing it wrong.  You have set an arbitrary and imaginary limit on the stats and said "if you play below these tiers, then you are a bad, bad person."  

 
You then claim that people playing that low are able to boost their WN7.  I call complete hogwash on that.  Playing at low-tier in WN results in lower WN scores.  Making them artificially lower so that you and other 'super-unicums' can feel better about yourselves is at best, false pride.
 
How do I know that playing low tier drops score?  If I even dare touch any of my tanks that are under tier 5 (as of current WN7), then my WN score drops like a rock.  I have to abandon my FCM36 PaK40, T-15, Tetrarch, Micro-Maus, etc. or else my score drops 100 points overnight just by playing 1-2 games in them as is.
 
Now I am not a clanwars guy.  But I have told you (and you keep ignoring me) that Clanwars Clans will drop WN like a hot potato because they will get their scores dinged heavily just for participating in clanwars, as many matches now require tier 1 or 3 tanks.  
 
Get rid of the penalties for low-tiers as it only shows a personal vendetta by you and a handful of others who are making the formula.  

 

Link to post
Share on other sites

Playing low tiers it's easier to farm top guns...if you play 10 matches in your t57 doing well...then throw in a match in a mini maus where you get 10 kills your stats go off the chart. 

 

Whats that old math thing mean, median, mode?  Your 3k damage from the t57 looks meh on paper but then when you throw in a T2 match do the math..10X10=100+2   102/11=9.27 average tier suddenly 3k damage looks way better...

 

Private public doesn't see this because he isn't good enough to beat the brand new players lol...

Link to post
Share on other sites

Sorry, the "low tiers are easy kills" completely falls apart when you don't take into account historically broken tanks and currently broken tanks.

 

Gold derping in the M4 and PzIV was even easier than tier 1 to rack up huge numbers of kills. VK3601 was broken for a very, very long time. Going right back the Bobblehead was seeing tier 3 ffs.

 

Quote from that post:

"Any tier tank can, if played well, be a statpadding tank

 

Nope.

 

You're just playing well, playing well and actual stat padding are completely different things.

 

Another:

"I want to know whether you actually ARE that 1700 WN player I need to worry about, or if you're actually just a green wearing clothes that are slightly too big for you.

 

WTF are per tank stats for then?

 

Honestly it's normalizing people no-one cared about anyway and leaving openly super-padded accounts untouched. No wonder it looks sus to the baddies.

Link to post
Share on other sites

Sorry, the "low tiers are easy kills" completely falls apart when you don't take into account historically broken tanks and currently broken tanks.

 

Gold derping in the M4 and PzIV was even easier than tier 1 to rack up huge numbers of kills. VK3601 was broken for a very, very long time. Going right back the Bobblehead was seeing tier 3 ffs.

 

Quote from that post:

"Any tier tank can, if played well, be a statpadding tank

 

Nope.

 

You're just playing well, playing well and actual stat padding are completely different things.

 

Another:

"I want to know whether you actually ARE that 1700 WN player I need to worry about, or if you're actually just a green wearing clothes that are slightly too big for you.

 

WTF are per tank stats for then?

 

Honestly it's normalizing people no-one cared about anyway and leaving openly super-padded accounts untouched. No wonder it looks sus to the baddies.

 

Congratulations, please continue to pick and choose parts of the post without considering the whole thing.

 

Here's the part that invalidates your argument:

 

Now, due to limitations in Wargaming's API, it is IMPOSSIBRU to sort through the data and just remove low tier games in their entirety; all stats are reported in aggregrate.

 

You can't eliminate individual tanks like the M4/bobblehead/whatever the fuck you feel like because the API doesn't allow for it. The only thing you can eliminate for with any semblance of accuracy is entire tiers, and tier 1-3 are the ones with the brokenly high KPG, etc, and the lowest correlation to actual skill. Eliminating tier 4s isn't as good, since t4 scouts/non derps/B1 aren't exactly pad tanks; this continues into all other tiers.

 

Also per tank stats in game blow since they only show WR and number of battles played, which is easily padded in and of itself. I want WN to reflect actual skill to whatever extent is possible, and the normalizer does that better than WN7 does.

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...