Welcome to the Forum Archive!

Years of conversation fill a ton of digital pages, and we've kept all of it accessible to browse or copy over. Whether you're looking for reveal articles for older champions, or the first time that Rammus rolled into an "OK" thread, or anything in between, you can find it here. When you're finished, check out the boards to join in the latest League of Legends discussions.

GO TO BOARDS


My theory on winrate...and other mathy stuff

12
Comment below rating threshold, click here to show it.

Moby the White

Senior Member

11-02-2012

so in the short time i have lets break it down simply...

EDIT PLEASE NOTE: Correlation does not imply causation. Any assumptions that I make come with a certain degree of error whether it be alpha or beta error. Also not the chance of positive influence is equal to the chance of negative influence within a controlled environment. Meaning that your chances of going positive vs negative are the same and depend upon your skill and your personal winrate while being only possibly increased by the number of times you play that champion.

People whom tend to sheep are more likely to play a champion with a higher win-rate than a champion with lower win-rate. Also people whom follow lolking like its the bible also apply to this stipulation.

Now a champions win-rate is determined by the number of times that champion wins divided by wins+losses

We can say that E(sigma) (Wins+ Losses)= total games played

so the probability of winning is directly affected by the number of games that champion is played. Ergo we can assume that the more you pick a champion the more they are played obviously. And also from that we can see that the more you pick a champion the more likely you are to increase the chances of affecting the winrate.

A negative correlation would show that the more you pick your champ the less you lose with the champ whereas a positive correlation would suggest that the more you pick your champ the more you win. Of course these are both with seperate varibles on X and Y...

Continuing on... so if a champ, let's call him H, becomes popular and more people pick him then the chances of directly creating a positive influence upon his winrate is highly more likely. This makes sense due to the fact that we now have more opportunities to let champ H win. Now assuming players win with him a respectable amount of time at a higher popularity his current winrate will be validated more quickly due to a surplus of sample size. However suggest H wasn't as popular and still won the same respectable amount of times. In this instance H does not have a chance to fulfill a large enough sample size for our tests due to not being played as frequently and we cannot validate his current win-rate.

From that we can assume that a champions popularity drastically alters his ability to win due to the fact that he is played more and creates a larger sample size than underplayed champions. We now can suggest that wins are not just directly an effect of skill but more or less a combination of N's (sample size) and skill combined. However skill can also be directly traced to a correlation with popularity and picked probability. We can suggest here that skill with a champion has a correlation to popularity aka "sheeping"

I highly believe that winrates of higher played champions in comparison to lower played champions is an unfair basis of champion ability in game. I feel that due to skill being directly affected by popularity in instances that lower played champions skew towards the negative due to not receiving proper play percentiles with players whom are experienced with the champion.

tldr: mathy math allegations.


Comment below rating threshold, click here to show it.

aKa Bad BoY

Senior Member

11-02-2012

i heard math, i see a lot of reading

i quit


Comment below rating threshold, click here to show it.

Moby the White

Senior Member

11-02-2012

Quote:
Moby the White:
so in the short time i have lets break it down simply...



*and then after saying this writes a full page thread :S*


Comment below rating threshold, click here to show it.

Moby the White

Senior Member

11-02-2012

Quote:
BadBoY9x:
i heard math, i see a lot of reading

i quit


sorry :S

Don't QUIT

its extremely ...yeah its boring if you aren't interested :S


Comment below rating threshold, click here to show it.

ViktorKush

Senior Member

11-02-2012

"People whom tend to sheep"

Who*


Comment below rating threshold, click here to show it.

Moby the White

Senior Member

11-02-2012

Quote:
FatBear5090:
"People whom tend to sheep"

Who*


whom...
girls play too

and since i do not know the direct gender of whom i am speaking whom is appropriate.
although the ruling on this kinda says that if its he/she then its who and if its him/her its whom

but im referring to an entity People :S


Comment below rating threshold, click here to show it.

lmaonade200

Senior Member

11-02-2012

Quote:
Moby the White:

A negative correlation would show that the more you pick your champ the less you lose with the champ whereas a positive correlation would suggest that the more you pick your champ the more you win.


you certainly took the long way of saying the exact same thing

Quote:
Moby the White:
whom...
girls play too

and since i do not know the direct gender of whom i am speaking whom is appropriate.


Who*

it has nothing to do with gender but rather if you are referring to the subject or the object of the sentence, "people" is the subject of the sentence, and therefore should be referred to by "who"


Comment below rating threshold, click here to show it.

Digitality

Senior Member

11-02-2012

More people picking a champ should actually push them towards a more accurate win rate indicative of their match up strength. Not specifically in either direction.

Granted it's not a perfect science with multiple uncontrollable variables given the rest of your team.

The idea that more games played by itself correlates to a higher win ratio is fallacious. Based on the assumption that more games played alone indicates increases in personal win rates.


Comment below rating threshold, click here to show it.

Moby the White

Senior Member

11-02-2012

Quote:
lmaonade200:
you certainly the long way of saying the exact same thing


Kind of but if you saw it on a chart you would understand :S

Quote:
lmaonade200:

Who*

it has nothing to do with gender but rather if you are referring to the subject or the object of the sentence, "people" is the subject of the sentence, and therefore should be referred to by "who"


yea...upon recent google search i noticed this...I was only stating something some english teacher told us ... :S


Comment below rating threshold, click here to show it.

Moby the White

Senior Member

11-02-2012

Quote:
Digitality:
More people picking a champ should actually push them towards a more accurate win rate indicative of their match up strength. Not specifically in either direction.

Granted it's not a perfect science with multiple uncontrollable variables given the rest of your team.

The idea that more games played by itself correlates to a higher win ratio is fallacious. Based on the assumption that more games played alone indicates increases in personal win rates.


Remember correlation is not causation...
so first we are basing this on a system where the skill level of the player is not yet introduced...more tba


12