So I was wondering after doing some empircal statistics in school this year, why not apply that to Call of Duty ....
Asking the Question does a high K/D ratio = a high W/L ratio.. and in my subjective opinion "skill"...
So I would ask the community to fill out the following form (if you feel like it)... and a week from now I'll run the vectors through a statistical program called R, and tell you fine people the results.
Just fill out the submission form... I am looking for at least 250 entries
I do ask to keep the results honest.. and I will collect only those reults which link to your Call of Duty Elite page. To get a link to your Elite page go to Elite - go to Find a player - Type in your online indentiy in "Find a player" for Example mine would be "v Sim CO", and then press enter... you should get to a page which looks much like you homepage but with a different URL link, please use this URL link in completing the form.
So argue amongst yourselves... about the topic...
Sorry I effed up the elite link on my first one.. imo There is no link between A high KD = a high WLR. It really depends on what you would call high, Whats your idea of a high kdr or high wlr? Do you take what gamemode you play into consideration?
Your experiment has a flaw in that it's not taking into account other factors, such as:
1) Consistently playing in a party
2) Game Mode Normally played
3) Mid-match quitting
A high K/D in any event doesn't necessarily correlate into a high W/L, nor does a high W/L correlate to a high K/D. There's just too many underlying factors that your experiment/analysis doesn't factor in to be accurate.
Sure I know the flaws, but the only way to get rid of them would be to do quite a bit of data work... which would just take too much time. sometimes... in analysis you have to make assumptions... and simplifications to account for the real world problems of data analysis. This analysis if I get a large enough sample... would probably tend to the average beta anyways.... mitigating the problems you listed.. eq 1. is y = xB + error Test for correlation of x and error, and correct with AR(1) if a problem exists. Use a t-test for significance etc... Hypothesis is that there won't be a statistical relevance.