Jump to content


Matchmaker vs. RNG shootout - examining the black box.

MM stats pathetic winrate matchmaker noobteam RNG WR WN7

  • This topic is locked This topic is locked
4 replies to this topic

nm_pytre #1 Posted 21 August 2013 - 11:04 AM

    Corporal

  • Player
  • 1742 battles
  • 179
  • Member since:
    05-29-2013
Let's have a look at an epic battle of Match Maker playing against so called Random Number Generator.

So at approx. 450 battles I quit playing with 8.6 SPG change, y'all know why.
I got back last week to have fun analyzing the black box and run no-ammo SPG riot.

That SPG no-ammo riot lasted until after the 1000 battles.
It did have a minor interruption at 850 battles lasting shortly to verify one claim that was posted in forum.

So the noobinator statistics show it from the 500th battle where Match Maker was already scared with my performance of last 50 battles taking a dip from 52% to 45%.
The matchmaker was scared to the point that I have been consistently assigned better chance to win teams.
The Match Maker was scared of my game experience so badly that it did provide me with a bumped up bonus to be on better team every next game.
Since I used the SPG as collision ninja / capture ninja / slow scout few times static not a single shot fired. Try it and you will see it's playing the impossible.
Yet I got my Win Rate graph consistently rising for the period of 500 - 850 battles.

Now what did happen at the 700 battle?
I grinded to tier 4 and the WN7 statistics has therfore reduced the progressive sealclubbing penalty of many battles/low tier.
This also show the flaw in WN7 - you want better WN7?
Then play just one single branch of vehicles per account.
Do not unlock two vehicles of same tier.
Always buy 100% crew. Never grind it to 300% in what WN7 considers as low tiers.
Simple as that and you will outperform almost anyone else in the WN7 rating.
Similar case for the performance rating.

What happened at the 850 battle?
I got ammo for few battles and started shooting to verify forum post claims.
My RNG bonuses were already scared by my performance and I saw next to no bounces.
So did I see massive amount of damage close to shell maximums with every hit.
You know average SPG chance to hit is dependant on static/moving vehicle with the average 3 seconds shell delivery time.
I did shovel the ground with ammo couple battles just to see massive chance to hit on static vehicles (same or better as pre 8.6).
That's when Match Maker started to be scared of me being too good compered to my long term history performance.
So it did start to reset my MM bonuses and start to penalize me for having improved way too quickly to the point of below the 10% chance to win teams.
I can not report on resetting the RNG bonuses since I got back to no-ammo riot for 870 - 1000 battles.
So I almost hit the bottom while having my MM bonuses reduced again.

A bit after the 1000th battle I quit the no-ammo riot, ninja SPG.
Starting the quest to see when's gonna Match Maker put the bar in the door of my WR.
And it did not last too long to reproduce the very same result. See the 850 WR peek and 1150 WR peek and the same angles of the rise/slide.

Attached Files

  • Attached File   mm_vs_rng.jpg   143.48K


nm_pytre #2 Posted 21 August 2013 - 12:01 PM

    Corporal

  • Player
  • 1742 battles
  • 179
  • Member since:
    05-29-2013
WN7 Additons:
For better WN7 rating also do convert excessive experience on the vehicle to free experience. Never leave any on the low tier vehicle.
For best WN7 results buy premium vehicle of tier not penalized by WN7 rating as low tier. Buy 100% crew, Train the crew bonuses to +300%, switch off the accelerated crew training. Use gold to convert premium vehicle experience to free experience. Research the WN7 penalized vehicles with the free experience. Never fight a battle with the WN7 penalized vehicle.
Buy just one branch of vehicles and never take a battle in tier that does WN7 penalize as low tier.
MM / RNG Additions:
See the completely flat performance after 1150th battle.
And as for gaming experience it can be described as cooperative multiplayer of Match Maker + Random Number Generator + player in battles 850 - 870 and 1000 - 1100.
While after 1150 it's a team deathmatch of Match Maker + RNG against Player.
Sorry I did not took the picture of 1150 battle statistics where the performance, damage, exp were skyrocketing so sharply that it was almost vertical line.
Also making any vehicle I touched ace tanker almost no-hands with just avoid death tactics between 1000 - 1150. When the MM and RNG started to shift sides not at the same time. RNG lasting a bit longer by my side playing against MM. More precisely till the reboot/server maintanance.

Edited by nm_pytre, 21 August 2013 - 04:59 PM.


Unch_04 #3 Posted 21 August 2013 - 12:18 PM

    Sergeant

  • Player
  • 59082 battles
  • 272
  • [FE2O3] FE2O3
  • Member since:
    08-23-2011
yaaawn

nm_pytre #4 Posted 21 August 2013 - 03:56 PM

    Corporal

  • Player
  • 1742 battles
  • 179
  • Member since:
    05-29-2013

View PostUnch_04, on 21 August 2013 - 12:18 PM, said:

yaaawn
Spoiler                     


zaphod_4242 #5 Posted 21 August 2013 - 04:43 PM

    Warrant Officer

  • Player
  • 13113 battles
  • 731
  • [GLRYB] GLRYB
  • Member since:
    06-09-2012
To make this more than just a conspiracy theory, you would have to produce some proof that meets at least basic scientific standards. No, it's not up to us to proove otherwise.

You didn't (conclusively) describe the concept and the methodology of your "study" - your description is very erratic.
You didn't describe the measures you took to ensure that it's not just that kind of statistic you can only trust if you rigged it yourself - keywords would be: multiple test subjects, standardized tests, statistically relevant sample sizes, double blind testing, control groups, peer review.





Also tagged with MM, stats, pathetic, winrate, matchmaker, noobteam, RNG, WR, WN7

1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users