Jump to content
Sign in to follow this  
You need to play a total of 20 battles to post in this section.
Snargfargle

Proof that Win Rates in the Matchmaking Monitor Don't Necessarily Predict Wins

22 comments in this topic

Recommended Posts

Members
778 posts
2,755 battles

Had 2 games yesterday in a row where my side then the enemy side didn't sink a single enemy ship... My teams loss was so catastrophic it was fun to watch 😊 (nearly all our tier 9+ BBs were annihilated in the first 4m)

Share this post


Link to post
Share on other sites
10,577
[WOLF3]
Members
26,967 posts
23,767 battles

Yeah, if I had a choice to have a couple 60% WR+ guys on my team instead of some 45% guys, I'd be crazy to not take the Unicums.

 

The 2009 Detroit Lions had a record of 2-14.  Hey, they won two games, they're just as good as the best of the best in the league, right?

Edited by HazeGrayUnderway
  • Cool 4

Share this post


Link to post
Share on other sites
7,819
[PSP]
[PSP]
Members
12,596 posts
2 minutes ago, Pytheas said:

As they say, statistics is hard.

I found them so, even though I eventually taught a course in statistics. I made one "B" in grad school, ruining my 4.0 grade average, it was in a biometery course. And don't get me talking about calculus; I simply hated that course, though I eventually pulled off an "A." I have used calculus exactly zero times since college. However, I still work with stats quite a bit.

  • Boring 1

Share this post


Link to post
Share on other sites
5,063
[ERN]
Modder, Alpha Tester, In AlfaTesters, Beta Testers
10,877 posts
4,881 battles
1 hour ago, HazeGrayUnderway said:

Yeah, if I had a choice to have a couple 60% WR+ guys on my team instead of some 45% guys, I'd be crazy to not take the Unicums.

don't drink and drive

liq_uni1.jpg

Share this post


Link to post
Share on other sites
3,953
[RLGN]
Members
14,250 posts
25,088 battles
5 minutes ago, MajorRenegade said:

don't drink and drive

liq_uni1.jpg

Lord, that color...

https://commons.m.wikimedia.org/wiki/File:Zwack_bottle.jpg#mw-jump-to-license

Edited by Estimated_Prophet

Share this post


Link to post
Share on other sites
783
[WOLFC]
Members
1,611 posts
9,445 battles
2 hours ago, Snargfargle said:

image.png.4368687b4f6deeebc3b0ba0bd6e9d55f.png

image.thumb.png.e10e83e9dd27233e13cd6cf90472d233.png

That’s not how statistics work. :cap_old:

Edited by Nevermore135

Share this post


Link to post
Share on other sites
7,819
[PSP]
[PSP]
Members
12,596 posts
13 minutes ago, Nevermore135 said:

That’s not how statistics works. :cap_old:

Quote

 

Hyperbole may also be used for emphasis or effect.

Nit-picking fallacy --  focusing on trivial details instead of directly addressing the main issue in dispute.

 

 

Share this post


Link to post
Share on other sites
1,583
[-BUI-]
Members
2,021 posts
5,668 battles

Lone random exceptions are just that, exceptions.   Exceptions are not "proof" that something isn't generally true.  

  • Cool 1

Share this post


Link to post
Share on other sites
783
[WOLFC]
Members
1,611 posts
9,445 battles
1 minute ago, Snargfargle said:

 

:Smile_amazed:

What is the “main issue of dispute?” All you’ve done is pick a specific example and use it as evidence of a broad-ranging conclusion of yours without any consideration for the wider context. Of course MM Monitor results will not determine with 100% accuracy the outcome of a match - otherwise why play at all? That is what you have shown. However, that is something completely different from claiming players’ winrates (a reflection, by and large, of their level of skill) cannot be used to predict the likely outcome of a match. 

From Miriam-Webster:

Predict:

to declare or indicate in advance
 
especially foretell on the basis of observation, experience, or scientific reason
(bold emphasis mine)
 
3 minutes ago, Zenn3k said:

Lone random exceptions are just that, exceptions.   Exceptions are not "proof" that something isn't generally true.  

^This.

I don’t use MM Monitor. I think it’s a cancer  and I’m not defending it. I also don’t believe in stat shaming and have never and will never do that to any player in this game.

This thread combined with your recent posting history (including a thread along these lines that was recently shut down by the mods), seems to indicate you are trying to find a reason for what you have, individually and without input from any other person on this forum, decided is subpar performance on your part (or at least what you view as a subpar win rate) versus whatever standard you hold for yourself. And I say this without any malice or ill will towards you. MM averages out over a large number of battles: you are just as likely to get a “good” team as a “bad” team. The common factor in every player’s matches, including my own, is the player in question. In this game, as in life, I am the chief controller of my destiny.

  • Cool 1

Share this post


Link to post
Share on other sites
897
[OIL-1]
Members
1,247 posts
5 minutes ago, Nevermore135 said:

MM averages out over a large number of battles: you are just as likely to get a “good” team as a “bad” team.

I have 4,000 random battles. At what point do I get the good team?

Share this post


Link to post
Share on other sites
215
[-RNG-]
[-RNG-]
Members
247 posts
12,729 battles

Usually in my 10K + battles if you are top tier you have bad RNG, bottom tho opposite. 

I use Match Maker to kinda track my own stats and to know who to focus on the opposite team...in recent trends I see the lowest win rate prediction ends up winning...why? All in the hands of the 50% WG mantra! Beat down the good and give the weak a chance. 

It is taxing, it is boring, and will be the end of this so called "game"  

Share this post


Link to post
Share on other sites
943
[WOLF6]
Members
1,849 posts
5,553 battles
28 minutes ago, Brhinosaurus said:

Trying to see how many threads about matchmaking and win rates you can get locked in one day?

You don't learn do you @Snargfargle?

See the sig? The little permanent comment or image that splatters their stats or philosophy all over everyone's faces like a.....ok I'll stop there. It's below the txt of the post. Anyway yes that's the sort of players you're dealing with. Relax mate and enjoy the game. Hit me up too if you want to Div.

 

 

Edited by Sumseaman

Share this post


Link to post
Share on other sites
783
[WOLFC]
Members
1,611 posts
9,445 battles
34 minutes ago, z9_ said:

I have 4,000 random battles. At what point do I get the good team?

How do you determine a “bad” team versus a “good” team? Even if you assume there are vastly more “bad” players than good players (does that make every team a “bad” team?), MM is still random. Of the 23 other players selected from the queue for a random battle, any one of them has the same  47.8% of being selected for your team, and the same 52.2% chance of being selected for the red team. Now, unless you are claiming MM isn’t random (with the exception of divisions, of course), over a statistically significant sample size the influence of these other players on the ultimate outcome of your matches will have a noticeable effect, but how much is determined by where you as a player sit relative to the playerbase as a whole. If you are an above average player you will tend to influence your team’s chance of success in a positive direction. If you are a below average player, you will tend to influence it in a negative direction.

There’s also an inherent bias when assessing how good or bad your team is vs the red team (myself included in this). You have all the information regarding your own play, and a self aware player will be aware of his or her potato moments. You have less, but still a lot of information regarding your team’s play, due to the minimap and the way vision mechanics work in the game. You can be aware of your teammates potato moments as well. You have even less information on the red team’s play and let’s be honest: when a red  player does something stupid and you capitalize on it, are you more likely to attribute it to your own good play or your opponent’s bad play?

 

Edited by Nevermore135

Share this post


Link to post
Share on other sites
5,148
[O7]
Alpha Tester, Alpha Tester
11,904 posts
8,457 battles
4 hours ago, Snargfargle said:

image.png.4368687b4f6deeebc3b0ba0bd6e9d55f.png

image.thumb.png.e10e83e9dd27233e13cd6cf90472d233.png

Could someone point out what the prediction is please?

  • Cool 1

Share this post


Link to post
Share on other sites
443
[KRAB]
Members
890 posts
7,343 battles

The biggest issue with MM monitor data is sample size. When some players have under 10 games in the ship and most have under 100 their win rate has big error bars on it.

It is good at identifying elite players or bots but the rest is a coin flip.

Share this post


Link to post
Share on other sites
897
[OIL-1]
Members
1,247 posts
20 hours ago, Nevermore135 said:

How do you determine a “bad” team versus a “good” team?

Everyone on my team has a 45% w/r. Everyone on the nme team has a 55% w/r. That's how. Doesn't mean we lose every battle, just shows that the game is unfair.

For everyone that still hasn't taken the red pill, it's currently back in your face.
Look at the team scores at the end of the battle. Now look at the clan names.
Everyone on my team belongs to a brown clan.
The players on the enemy team belong to clans colored purple, green, yellow, etc.
This is just one aspect of the 'many' ways that WG and other companies rig games.
The majority of the population has known about it since the early days of WoT.
But some people still insist that the Earth is flat.

  • Confused 1

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.

×