Mountain_Raider

Stats - WR Question

24 posts in this topic

Dumb question that has been driving me nuts...

Is MM screwing with me?

I know the answer is always, yes.  But this seems like they're taking it to an extra level.

My win rate looks like a roller coaster.  A roller coaster with statistically significant and oddly consistent up/down periods.  Meanwhile, all of my other stats show a consistent increase in performance and almost no correlation to the WR numbers.  Is this a result of MM's attempt to balance the teams?  I usually play the same tanks every session.  I almost exclusively play Random battles and generally don't platoon (no one ever wants to beyond one-game, dynamic toons and my clan members seldom play anymore or toon up when they do).  I don't think my play style varies that much.

With all of my other stats being consistent, you'd think that the WR would be consistent, too.  What am I doing wrong?  Is it even me?  And, why do I care so much?

MR_Stats.gif

Share this post


Link to post
Share on other sites

While DPG may be on the rise, DMG is a way too low to impact (and possible skew) WR line. Practically you are passenger dependent on the team. Unless you are playing T2 with 1000 dmg per game :D 

Share this post


Link to post
Share on other sites

For the last 1000 games, my average tier played is 7 with avg DMG of 1500 and WN8 of 1957.

That's bad?  Perhaps to the the Unicums, yes.  But I mean, saying I was merely a 'passenger dependent on the team' seems a bit harsh.  I certainly don't camp in the back and am often the highest damage dealer (and at least in the top 5 to 10 XP wise) at the end of the game regardless of whether I'm top or bottom tier in that game.

Share this post


Link to post
Share on other sites
3 minutes ago, Mountain_Raider said:

For the last 1000 games, my average tier played is 7 with avg DMG of 1500 and WN8 of 1957.

That's bad?

It's not. However if you seem to fall short on winning you simply might be doing one of these too often:

A) dying too early and being unable to clutch it out

B) playing too passive meaning you miss out on the actually important damage

C) derp and more derps

Share this post


Link to post
Share on other sites

I've proffered in previous posts that my decision making process for when to play aggressive vs passive needs improvement.  I'll just have to play more to get gud and make better decisions.

However, all things being equal and assuming my (arguably inferior) play style is consistent, how can the WR cycle up and down in an obvious pattern like that without some external influence?  (The standard deviation for this fluctuation is only 0.05% - which is admittedly minor.)

If the observed cycles were all my doing, that would mean I'm consistently brilliant and impactful for almost exactly every 160 games, after which I go full noob for another 160 games.  I somehow then wise up and become brilliant and influential again and the pattern repeats.

It's also not like I'm having a bad session of games one night.  My average number of games per day is around 12.  So for almost exactly every two week period, my WR is either averaging up or averaging down.... consistently... for the last few thousand games.

There's NO WAY that level of consistency can be attributed solely to me.

Share this post


Link to post
Share on other sites

That just means that you can win 54.something% of the games on avarage. And its consistant with your damage dealt.

Fluctuation is almost non existant. Theres probably 1 in 20 games where you get either lucky or unlucky and respectively win or lose a game that you generaly shouldnt.

Share this post


Link to post
Share on other sites

OK, so I'm going to be the math asshole here.  You do realize the scale of that graph shows you have stayed within 0.07% of the same WR over the last 1600 games.  Essentially at the 15000 game mark you had a 54.15% WR and now you after 16600 games you have a . . . . yep, 54.15% WR.   Also notice that on each of those micro swings your high has been a little higher and your low has also been higher each time.

Relax, its mostly just an illusion created by a wonky graph scale.

Share this post


Link to post
Share on other sites
9 minutes ago, Archaic_One said:

Relax, its mostly just an illusion created by a wonky graph scale.

No, you're not a math asshole.  But if you are, I'm a bit of one, too.

I totally get that the fluctuation in WR is insignificant and if the scale were to be normalized, that the graph line would flatten out to a gently rising slope.

AND...

I'm talking about the pattern of fluctuation itself.

I just looked at your WR stats (and Gr1nch_1's, too) and the same pattern of up/down fluctuations is present.  There's a peak / valley every 300 or so games....  consistently.

As a 'math asshole' or more correctly, a 'statistics asshole', this is significant and points to an external influence at play on the system.

Shit.... I'm sounding like a conspiracy theorist.

Ok.  I'll leave it and stop stressing.

It's just a game.

(..but that pattern)

Share this post


Link to post
Share on other sites

I think we (new players) learn how to farm damage much more quickly than we learn how to win games. I have pretty much learned how to farm damage, and I focused on that for a long time. Now I'm not as focused on farming damage, and it has hurt my ability to farm wn8 significantly (only a problem if I cared). But my focus is on winning now, and my WR graph no longer looks like your roller coaster graph, it is a much smoother upward trend. But it's way harder, since even on a loss you can farm a lot of damage and if that's the only stat you're focusing on, you can tell yourself, "well, I did MY job at least". But I think that's a counterproductive mindset. Just my two pence. Keep going, look for opportunities for relevant damage - 240 damage to a low HP heavy or TD locking down a corridor might be a game-changing contribution, even if that's all the damage you do all game.

Spotting, tracking, and finishing off low HP guns has way more impact than sheer damage output.

Share this post


Link to post
Share on other sites
1 hour ago, Mountain_Raider said:

I just looked at your WR stats (and Gr1nch_1's, too) and the same pattern of up/down fluctuations is present.  There's a peak / valley every 300 or so games....  consistently.

300 games, about the equivalent to doing T28 and 110 stock grinds simultaneously . . . AKA the amount of games WG has calculated it takes us to say fuck it and pay to free-XP through it.

Illuminati confirmed

Share this post


Link to post
Share on other sites

solution: a certain donation to a certain paypal will fix that curve for ya real quick :kappa: 

Share this post


Link to post
Share on other sites
11 minutes ago, Kolni said:

solution: a certain donation to a certain paypal will fix that curve for ya real quick :kappa: 

You realize you have to say @CarbonWard for him to see this :gaytroll:

Share this post


Link to post
Share on other sites

Have you tried to correlate your win rate fluctuations (and potentially DPG and XPG using a finer scale) to a weekly or diurnal cycle? Theoretically, both of these factors will influence the playerbase and therefore the relationship between individual statistics and win rate. The playerbase may not necessarily be worse at any point, but it would not take a huge shift in metagame or population to have the slight effects you are observing. 

I would personally do a frequency domain decomposition of your moving-average (say 1000 game) win rate, checking if any statistically significant periodicity does exist. It is quite possible that random chance is sufficient to explain everything. 

You could also just get fatigued and play worse (or more farmer-like) then usual once a week or so, I distinctly remember having days when I played like a shitlord and other days where I would make Kewei complain that I was stealing his damage. 

Stick with it and it will even out over time - you are well above average already, with nothing to be ashamed of. 

Share this post


Link to post
Share on other sites

Does the "every 300 games or so drop in WR" match up with say event weekends?  Just a thought...we generally have one whether it's a popular on track or whatever every 2-3 weeks or so.   My stats usually see a slight drop during those too.

Share this post


Link to post
Share on other sites
Just printed out mine and took a ruler to it, lo and behold, 400 game fluctuation. I have suspected MM to be a part of this and it's interesting that you are seing the same trend. MM goes +10 RNG for 200 games then -10 RNG for the next
 

win rate fluctuation 400 games.JPG

Share this post


Link to post
Share on other sites

I probably have the most anomalous graph for WN8. 

3uEnFck.png

Share this post


Link to post
Share on other sites

This is what happens when you play extremely infrequently

4bbe1576bb.png

 

Share this post


Link to post
Share on other sites
1 hour ago, Henderson27 said:
Just printed out mine and took a ruler to it, lo and behold, 400 game fluctuation. I have suspected MM to be a part of this and it's interesting that you are seing the same trend. MM goes +10 RNG for 200 games then -10 RNG for the next
 

win rate fluctuation 400 games.JPG

You get the "rollercoaster-effect" because wotlabs creates "save-points" every 100 or so games(You can see them by hovering over the graph).

Tracking the results of every single battle for every single player would be utterly retarded so the site creates these save points and the graph just interpolates between them.

These fluctuations are not only normal but to be expected since you will never perform exactly the same every single session you play, especially at these low battlenumbers you will see large fluctuations even if you are a robot and perform exactly the same every single game since this is how randomness works.

At 8000 battles played you only have to win/lose 8 more battles than average in a 100 games-session to increase/decrease your overall winrate by 0.1%.

 

41 minutes ago, MagicalFlyingFox said:

This is what happens when you play extremely infrequently

4bbe1576bb.png

 

That is how the stats of someone look like who just discovered the art of redline sniping.

Share this post


Link to post
Share on other sites
6 minutes ago, Fabunil said:

That is how the stats of someone look like who just discovered the art of redline sniping.

 

Skillz.JPG

Ooh, do me, do me!

Share this post


Link to post
Share on other sites
1 hour ago, Fabunil said:

 

That is how the stats of someone look like who just discovered the art of redline sniping.

I should get my average spots down :D 

 

Oh, and play light tanks. 

 

 

the big issue is that I'd need to actually play the game for that. 

Share this post


Link to post
Share on other sites
13 hours ago, Archaic_One said:

OK, so I'm going to be the math asshole here.  You do realize the scale of that graph shows you have stayed within 0.07% of the same WR over the last 1600 games.  Essentially at the 15000 game mark you had a 54.15% WR and now you after 16600 games you have a . . . . yep, 54.15% WR.   Also notice that on each of those micro swings your high has been a little higher and your low has also been higher each time.

Relax, its mostly just an illusion created by a wonky graph scale.

His overall w/r yes, but what about his w/r for the session when when the chart shows those peaks and dips? He was running a 45% w/r for 70 battles or whatever for the dips and then, all of the sudden, was winning at 65% for the peaks. A swing of 20 percent seems a bit extreme if all other factors are consistent.

Share this post


Link to post
Share on other sites

Fabunil

cheers, that 100 game interval makes sense. There still seems to be a correlation to games #s vs WR fluctuation that is common in lots of people's graphs, irrespective of how many games per day (so not time based) which is interesting. NB for the last month SEA server to Aus has had broken cables/ high ping so last months data is not good for comparisions

Share this post


Link to post
Share on other sites
3 hours ago, Henderson27 said:

Fabunil

cheers, that 100 game interval makes sense. There still seems to be a correlation to games #s vs WR fluctuation that is common in lots of people's graphs, irrespective of how many games per day (so not time based) which is interesting. NB for the last month SEA server to Aus has had broken cables/ high ping so last months data is not good for comparisions

If you want to page someone you have to write it like this @Fabunil

My WR directly correlates with my sleeping habits. If I am sleep deprived I get around 58% WR, and 62-63% if I slept enough. (This is on average ofc.)

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.