Jump to content


Replay analysis - what does this tell us about the MM?

Lets do science statistics

  • Please log in to reply
343 replies to this topic

Schepel #341 Posted 14 October 2018 - 02:01 PM

    Lieutenant Сolonel

  • Player
  • 59602 battles
  • 3,099
  • Member since:
    05-13-2013

View PostUrQuan, on 14 October 2018 - 10:20 AM, said:

 

Tbh, Baldrick proved one thing; players can rig their own games in their favour harder then any potential WG rigging system conspiracy theorists can think of to keep m down. It's called 'platooning'. I remember giving my replays to baldrickk as well & the result being quite skewed in my favour, because I platoon alot.

So if there's any rigging system that keeps good players down, I have a good advice for the conspiracy theorists: platoon up & you can beat the (supposed) rigging, because platooning has a stronger effect then the possible rigging.

 

In case people think my platoons are rigidly run, nope. I toon up & play my fav tank & do my thing, all the while having faith in my toonmates to roll out in their fun tanks & have fun too. No voicechat either, cause I can't use it.

 

 

 

Actually, no.

 

Platooning with tomatoes rigs the MM against you, since you bring bad players with you. Platooning only works in your favour if you bring good players. Even then, it really matters what kind of tanks you bring.



jabster #342 Posted 14 October 2018 - 02:04 PM

    Field Marshal

  • Beta Tester
  • 12537 battles
  • 23,387
  • [WSAT] WSAT
  • Member since:
    12-30-2010

View PostSchepel, on 14 October 2018 - 01:01 PM, said:

 

Actually, no.

 

Platooning with tomatoes rigs the MM against you, since you bring bad players with you. Platooning only works in your favour if you bring good players. Even then, it really matters what kind of tanks you bring.

 

It would seem somewhat obvious that they were referring to platooning with above average players.

UrQuan #343 Posted 14 October 2018 - 02:30 PM

    Lieutenant General

  • Player
  • 19453 battles
  • 6,182
  • [T-D-U] T-D-U
  • Member since:
    08-19-2011

View PostSchepel, on 14 October 2018 - 02:01 PM, said:

 

Actually, no.

 

Platooning with tomatoes rigs the MM against you, since you bring bad players with you. Platooning only works in your favour if you bring good players. Even then, it really matters what kind of tanks you bring.

 

In which case you still beat WG's rigging system, your platooning has a bigger effect then whatever possible rigging WG has implemented according to the 'MM is rigged crowd' We know this, because accounts that platoon alot are picked up by Baldricks replay analyzer as an anomaly (their teams do a bit better/worse statistically & in line with the quality of their toonmates), as the platooners actually rig the MM in their favour when playing with above average players (or rig it against them when playing with below average ones).

And that's just with you bringing one or two players with you. Not the entire team.

So yes, if you feel WG is rigging your matches to bring you down, a simple way to counteract it is to platoon up with above average players & have fun. Not even a need for strict regulated teamplay. 

 

Note, my platoonmates are in most case the Dingers, with the Forumites second & Barumites third; total random strangers are 4th (Hi general chat folks!). Which does mean I generally toon with above average folks in skill.


Edited by UrQuan, 14 October 2018 - 02:32 PM.


Baldrickk #344 Posted 14 October 2018 - 09:55 PM

    Field Marshal

  • Player
  • 30109 battles
  • 14,292
  • [-TAH-] -TAH-
  • Member since:
    03-03-2013

View PostVarzA, on 13 October 2018 - 04:41 PM, said:

I only just saw this thread, nice analysis, and i wish i could understand most of it tbh (really, my background is not in statistics :P ).

 

One thing though, i don't think your analysis is that good overall because it starts off with the wrong data, it would be better if we had replays from 1 patch alone, a large quantity. Even then, what if someone uploads them on a losses streak.

 

One thing is for sure, whatever way we look at data, turbolosses/turbowins happen (especially yesterday evening, it was insane, during friendly friday), and huge losing streaks do happen.

What is the probability, for those streaks of losses/wins to happen, if MM is truly not rigged ?

 

If those streaks are in the end the action of the MM (set to help/scr*w you over or simply because it feels like it), how can we talk of true random behaviour vs the individual (as a group of players maybe, and over a long period of time).

 

It just doesn't make sense to me that those streaks of wins/losses can happen if the MM is truly random.

How is it the wrong data?

 

This isn't the only data I've run it on,  but it's a pretty good dataset.

Lots of battles,  almost all of them watched until the end so results are recorded,  all from one player. 

 

It's long term trends that we are most interested in. It reduces the impact of RNG, which dominates in the short term. 

 

If you want to play 5-10k battles in one patch,  then you for it. 

 

With a statistically significant sample size,  it shouldn't matter if someone had a loss streak just before running the analysis.

 

I have to admit though that I've had some tinfoil moments myself where I lost most og my battles.

You can imagine how pleased I was to see that my team was on average stronger, and the one where there was a 40%! difference between our teams, with my team stronger, we lost 15:2...

 

Random is just that.  It naturally forms streaks.

When you toss a coin, you don't get:

HTHTHTHTHTHTHTHTHTHTHTHTHTHTHTHT

That is most decidedly not random, is it?

Despite that, it's exactly what people tend to think should happen,  that results should come equally.

Instead,  you're far more likely to get something that looks like:

HHHTHHTTTHTTTTTHHHHHTTHTHHTTTTT

and even that is periodic, so a bad example of random, but note the streaks.

 

21:00 Added after 5 minutes

View PostSchepel, on 14 October 2018 - 02:01 PM, said:

 

Actually, no.

 

Platooning with tomatoes rigs the MM against you, since you bring bad players with you. Platooning only works in your favour if you bring good players. Even then, it really matters what kind of tanks you bring.

 Yeah, but as people said,  platooning with above average players is implied.

 

As for tanks,  the analyser cares not. 

If you and platoon in AMX 40s, you might lose more than if you were in, let's say matildas. 

Your ratings are unchanged though. 






2 user(s) are reading this topic

0 members, 2 guests, 0 anonymous users