Re: k/d ratio

Well, since you want to cop an attitude, I most certainly will give you something more to ignore by responding to.

What purpose does it serve to tell someone how to do something the wrong way? I'm going to assume that even you are bright enough to know that telling someone how to do something the wrong way does more harm than in any way helping with the issue at hand. Furthermore, the OP did not ask "how to average" and, given the nature of his post, I think it safe to assume that he already knew how to determine and "average" - if that had been what he was looking for. However, the nature of the original post clearly indicated what he was actually trying to determine was the mode k/d ratio.

At least the OP had the decency to recognize he did not know enough about statistics in general and did not try act like he did.

I believe I had stated something to the nature that I would not even bother showing you how "to deliver an example of kdr that could potentially average to a value less than 1" only after you demonstrated you either knew basic statistical math beyond second grade "(1+1) / 2 = 1" averaging ... or you at least demonstrated a willingness to listen to how statistics actually works. If I did not, then I'm telling you now that is why I did not bother - because ultimately you would not listen to anything beyond "the world has always been flat."

Math is certainly not my area of concentration, but I'm not exactly known for taking "easy" paths to knowledge. That's why I know about statistics. But where I am well trained is in the written word. That said, omission can be just as loud as admission, sometimes even louder. In your case, that's what you're now trying to pull off and that's why you "never said blah, blah, blah." If any of what I'm saying comes across as being a hard arse, all I can tell you is that you certainly would not have made it through the instructors I've had because, frankly, I'm being about as gentle as I can be.

That said, I get exactly why you wrote what you wrote. You made the same mistake a lot of people make when considering the OPs questin: "what is an average k/d ratio." That's why on day one in statistics the instructor explains two things: 1) statistics is more about reading comprehension than it is math and 2) "average" =/= average. There is a popular connotation of "average" but that connotation is not the mathematical definition of "average."

That's why I clarified mode, median ... and MEAN in my original post to this thread. So that we would not have to listen to someone providing the inaccurate information of "Well, you take the kdr of all the players, add them up, and divide them by the number of players and there you have the average of averages."

Now let's deal with the crux of your argument -

"in the example your article gives it would be the equivalent of stating "the average unemployment ratios." But the article was intelligent enough to state the average unemployment level. Does it make sense to calculate the average of the ratio, NO. but this forum chose to look at the "average kill to death ratio" not the kill to death ratio of an average player."

You're actually agreeing with me and trying to spin around as if this is what you were saying all along. It is NOT what you originally stated nor what you originally implied. I quote you again ...

"Hate to say it but your math is wrong. It is almost impossible to have a KD average out to less than 1. you are making the assumption that if every kill = a death then the average must be one because there are the same amount of kills as deaths, and this is not true."

So, actually, yeah, you DID claim and argue that the "average kdr" was greater than 1. In fact, you went on to argue that an average kdr of less than 1 was impossible. See below.

"If you still choose to hold onto the belief the average is less than one i would like to see an actual example where the average is calculated out to be less than 1 (try to use realistic #s) "

I NEVER used the statement that suggested using the phrase "average kill to death ratio." Quite the opposite, see quote of myself below:

"When asking what the "average" KDR is, you are better off asking either what is the "mode" KDR or what is a "typical" KDR. You'll get a far more accurate answer."

Now, you can continue to have a chip on your shoulder because you got "shown up" or you can say, "Hey, I was wrong. No problem. I learned something. Let's move on." To wit I will, most assuredly, give you accolades in this discussion and future discussions. Just as others have done for me.

All of that said, I will demonstrate to you in two ways proof of my claim. First, the following results from a TDM match from BO2 (I would show results from a Ghosts match, but such data is no longer available thanks to the untimely demise of Elite). The following information is from one of three matches followed. In the first match, the winning team had 75 kills and 62 deaths, the losing team had 61 kills and 76 deaths. Obviously a player on each team suffered a "suicide." In the second match, the winning team had 75 kills and 30 deaths and the losing team had 30 kills and 75 deaths. Nothing "abnormal" there.

In the third match (these were randomly selected matches, to the whatever degree that they could be "randomly" selected) the winning team scored 75 kills and suffered 52 deaths. The losing team scored 52 kills and suffered 78 deaths. Here are the kills and deaths for each team, GTs eliminated for obvious forum rules reasons ...

Winning Team

24 - 5

21 - 9

9 - 10

7 - 9

7 - 6

7 - 7

0 - 4

0 - 2

Total = 75 - 52

Losing Team

18 - 10

9 - 13

10 - 18

7 - 15

8 - 9

0 - 1

0 - 12

Total = 52 - 78

The reason there are more than six players on each team should be obvious: some players left the match early; other players entered the match late (I have the times those events took place, as well).

Thus, if you look at the total population within this match, there were 127 kills and 130 deaths.

127 / 130 = 0.9769~ kills to deaths ratio

The second way of looking at this is to look at players outside of a particular match. I'm not going to use the players from these three matches because, frankly, doing so neither proves I am correct in my assertion nor does it prove I am incorrect. Instead, I'm going to look at players from the 10^th, 25^th, 50^th, 75^th, and 90^th percentiles. This is NOT an accurate reflection of the actual percentile in which each player lies. For ease of selection, it is assumed that the page number of the global career leader board (Black Ops 2) relatively reflects the percentile in which players, at large, fall in rank according to total score. Hence, Page 1 would be the highest percentile of players, whatever that percentile would be. To determine the page number for such percentiles, I simply looked at the total number of pages and did the math accordingly (ex: 50^th percentile =  1,700,984 / 2 = page # 850,492. Duplicating these exact results may not be possible because of continual change in total player population.

All disclaimers made, here are the results …

List of KDRs.

Histogram of KDRs.

While a sampling of 25 certainly has a fairly sizeable margin of error possible, even being off by a lot is still likely to yield a right skewed bell curve (noticeable in the histogram link).  The reason is that, if you will notice, below the 75^th percentile both KDR and total kills begin to drop substantially.

The above results are the reason myself and many other veteran forumites and Call of Duty players claim that the MOST COMMON kdr is going to likely be between 0.75 and 0.84 to 1. That claim is probably being generous. It is fully possible the “typical” kdr (commonly referred to as the “average” kdr) is BELOW 0.85 to 1.

Thus, I've shown that even in TDM matches where the winning team scores 75 kills, the KDR for THAT match CAN be below 1.0. I've shown that a random sample of players can have a combined KDR below 1.0. And I've demonstrated the most relevant statistic to the conversation is that the MOST COMMON (i.e., "mode") KDR is below 1.0.

Regardless of the VALID method of determining the “average” player’s kdr, the typical or MODE (correct terminology) … is … below … 1.0.

edit - reduced spacing between paragraphs; bolded quotes, corrected a percentile error in the text (not in the math).