How would I go about analyzing this data? Green is the winners and red is the losers. The green stats are usually always higher than the red ones. BUT, how do I go about seeing how much a team needs to have to be considered the "winner"?
I tried calculating the mean between the two and subtracting the difference but I'm not sure if this is the way to go.
Pic related is a very very small sample.
>>293155
What is "Maps" and what is the win condition or the game type?
>>293169
It's the amount of total games played for that team. K/D is the kill/death average for that team.
It's statistics for professional vidya.
what i believe the anon on >>293169 meant is: if the kdr can be lower for the winners than for the losers than there are other factors that can come into play to determine the winners of the match besides the number of kills (like rescuing hostages, or defusing a bomb, or time running out, etc.)
without those extra metrics, the only thing you can conclude about the data you posted is what you already know
>The green stats are usually always higher than the red ones
>>293155
so are you looking to model the binary response of win vs lose, and take in as factors how many maps and the K/D? As in given a certain K/D over a certain number of maps, will a team win
>>293194
if you're wanting some basic insight to the differences, there is no statistical evidence that losing teams have a different KD or Maps value to winning teams. This can be shown using a welch two sample t-test. I did this in R if you want the results lemme know
>>293155
here's the results of fitting a logistic regression. which predicts the probablity of winning given Maps and K/D.
The first eight are the winning teams, the second eight are the losing teams.
>>293201
so for instance, if team 1 vs'd team 9, you would predict that team 1 would win because the probability is higher.
>>293180
I got you mate, here's the extra stats. I didn't include it in the OP because I thought it was a lil confusing.
>Maps = Total games played
>K/D = kill/death ratio
>HS% = headshot ratio
>KPR = Kill Per Round
>APR = (Kill) Assists Per Round
>DPR = Deaths Per Round
>Hours = average recent hours for team
>>293206
How would it work with these stats? You got me a little confused.
anyone?
bumpin
>>293211
>>293238
>>293551
hey forgot about this thread. I couldnt be bothered typing all that data out properly for analysis. you should consider downloading the program R and learning how to fit models if you're interested in this kind of thing. I'm in my final year of studying statistics at uni and R, MATLAB and Python have been my best friends over the last 4 years :)