Go to elite and check the scoreboard on the games you played. Take all the KDR's and add them up then divide them by the number of players that were in the game. See what you get. I promise it will not always be 1 and it will not always be less than 1.
It has to always be less than 1. I am NOT saying that everytime I have a kill, that I have a matching death. I'm saying that K/D on the whole is a zero sum game.
There can NEVER be a KILL without a DEATH.
There can be a DEATH without a KILL.
There is a huge difference between these statements.
Everytime I get a kill, the person I killed had a death, thereby offsetting the average KD of 1-1 by nothing. The only time the average K-D of 1-1 gets offset is by someone dying without a matching kill (suicide). Therefore, the aveage K-D must be below 1.
Not sure how else to explain it.
Please, find me one game in your history where the average K-D in the match was greater than 1. The only possible way for that to happen is for the data connection between Elite and the game to mess up and not accurately count a death.
You're making a mathematical error. The true average K/D is not the average of all KDs in a match because you're not taking into account the weight behind each K/D. For example, 30 Kills / 11 deaths receives more weight than 3 Kills / 20 Deaths because statistically it accounted for more of the average kills and deaths.
The true average K/D of that match is the # of kills divided by the # of deaths.
Kills: 30 + 28 + 16 + 10 + 3 = 87
Deaths: 11 + 15 + 19 + 22 + 20 = 87
87 / 87 = 1
The true average K/D in Call of Duty would be slightly below 1. This is because, for each kill, there must be a death. However, for each death, there does not have to be a kill (suicide). Therefore, the average K/D will be a little bit below 1.
It depends on how you look at it I guess. When I think of "Average K/D", I'm thinking of a genuine average K/D, gained by taking everyone's kdr, adding it up, and dividing by the number of players. That would show the true average kdr from player to player, which I think is how the OP was trying to figure. In other words, if you compare player to player, what is the average amongst everyone? He's comparing himself to the average of other players, so we can't rely on simply total kills and total deaths across the platform.
In this instance, since you're trying to figure average (and not probability), the weight of something isn't a factor (at least not in the sense I assume you're referring to). We're just taking 5 numbers, and trying to find the average amongst those 5 numbers (in my example anyway). The reason I used an equal number of kills and deaths was to show that even with an even number, the average can still be shown as something unexpected (in my example it was 1.20).
I can't say that with a larger sampling that this will hold up however. Maybe if we averaged out everyone's KDR across an entire platform, the average would come down to below 1.0, but at least on a smaller scale, it's possible for the average to be higher than 1, despite equal kills and deaths.
I was apparently wrong about the OPs intent. I just read his reply, and it seems he is indeed simply looking for a 1-1 average for kills and deaths.
Look I'm not telling you this garbage for my health. So we are talking about the average KDR right? I'm not talking about all the kills in the game over the deaths. By normal distribution (http://en.wikipedia.org/wiki/Normal_Distribution) that would have to be 1. I get that, you get that so lets talk about the average KDR of every player. Again, from player to player there is nothing that gurantees a death after a kill or vice versa. Let's say I play and I never die. My kills will be over zero deaths and any number over zero is undefined (plug it into any calculator). So now throw this undefined number into a set of ratios and divide it by the number of ratios and you get a number that is still undefined, completely dismantling your theory. So going back to what I originally said: we as players who do not have access to the data sets cannot define the average KDR.
Statistically, people think that an average KDR should be around 1.0 given that for every Kill, there's a death. But people are looking at the equation all wrong. They're assuming total kills and total deaths, without regard for who those kills are attributed to. By that I mean, let's take this scenario for 5 people, all in the same match playing a FFA:
Player 1 - 30 Kills - 11 Deaths - 2.72 Ratio
Player 2 - 28 Kills - 15 Deaths - 1.86 Ratio
Player 3 - 16 Kills - 19 Deaths - 0.84 Ratio
Player 4 - 10 Kills - 22 Deaths - 0.45 Ratio
Player 5 - 3 Kills - 20 Deaths - 0.15 Ratio
Now, statistically, there's 1 death per kill. But the average ratio doesn't add up to 1.0. If you get the average of all their K/D Ratios, it comes out to about 1.20. So you can see how the average ratio isn't as simple as just adding up all the kills and deaths in the game. The average K/D for that match turned out to be 1.20 despite an even number of kills and deaths.
You messed up somewhere in your math. If you take the average of all ratios in your example, it comes out to 1.00333333etc. The additional 1/3 of 1% comes from having rounded the individual kdrs to two decimal places. Broken down:
2.72+1.86+0.84+0.45+0.15 = 6.02/6 = 1.00333333
If you wanted to find the median k/d level among a group of players, you would need to create a bell curve using all k/ds and find where it peaks. This would best be done using a much larger sample size than a single lobby, but I suspect it would still end up somewhere right around 1, give or take maybe 0.2.
Where in the hell did you people learn math?
He was right its the avg kdr in his example was 1.2. You divided your set by 6 as opposed to 5. He specifically said 5 people and there are 5 kdrs in the example.
During MW3 I actually recorded about 300 k/d from players in the lobby with me (from several different lobbies, obviously) and I did get a roughly normally distributed bell curve with the average around 1.25. It was slightly skewed to the lower end, it was not 100% normally distributed (though I think my r value was like .98...this was back when I was in school, I have since forgotten a lot about stats). The lowest k/d you can have is 0, and very few pleople had one lower than .5, and the highest you can have is infinite, though very few had higher than 3.
Anyway, the average may be a little lower on this game, but 1.25 seemed like a solid result for an "average" k/d in MW3.
Unless you play constant Team deathmatch, K/D means nothing for example my K/D at the moment is 0.80 but that's only because when I play Domination I always jump onto B when no one else really is meaning I take atleast 10 deaths trying to get 1 capture but when I go into Search and Destroy I usually get atleast a 2.00 K/D.