There is some really bad statistical math being bantered about in this thread. Really bad. I've read only two posts that can be considered remotely accurate.
professorgenki above has about as close to accurate explanation as you're going to get.
First off, the MEDIAN kdr is the most important statistic you can look at if you want to know what the kdr of the "typical" player performance is. The problem in this entire conversation is people throwing around commonly used words that, when it comes to statistics, have VERY SPECIFIC meanings that often do not coincide with their general use definitions.
You can all do yourself a favor by stop thinking in terms of "average player" and, instead, think of the "typical player."
There are multiple ways to define "average." For example, you can have a mean, a median, and a mode. All three are measures of center, or "averages." Sort of. The "mode" is NOT an average but will certainly tell you a lot about the "typical." The mode kdr of COD players, for example, I can assure is going to be below a 1.0 kdr. In fact, I'd bet that it is less than 0.80 to 1.0.
The "median" kdr is going to be below 1.0, as well.
What you guys are throwing around as a "weighted" kdr is not exactly accurate, either. Believe it or not, a low kdr is going to pull the overall kdr of a set of players lower moreso than a high kdr player is going to pull up the overall kdr. You can use common experience and common sense to understand this.
Think of your grades in school. Suppose you run a "B" average all semester long. Will scoring 50% on a final exam hurt you more than scoring a 100% help you? Of course an F on a final is going to "hurt" you more than an A is going to "help" you. The same principle applies when looking at the kdr's of a set of players.
Regardless, you need a "large" sample to get a more accurate reading of typical kdr than a single lobby of players. Ideally you need to do exactly what professorgenki above describes, though I would argue that if you looked at 3 or 4 n = 1000 samples, you can do further statistical math to determine the reliability of the statistical information you uncover.
If you want to take a short cut to figure it all out, use samples of at least 30 (that is, n = 30) and gather up six or seven samples ( "n" ) where n = the number of players in your sample.
You can't ask "What is the average kdr of COD players?" as an overall number because, yes, duh, the answer is going to be slightly less than 1.0* (as stated by those with the common sense to see this) due to suicides. As professorgenki noted, however, the asterisk isn't even really necessary because, statistically, the confidence level alone would negate the suicides.
The total number of kills scored by the combined community of COD players DIVIDED BY the total number of deaths suffered by the combined community of COD players = slightly less than 1.0 (due to suicides, accidental or on purpose, as well as team kills in hard core modes).
But that DOES NOT tell you the typical KDR of individual members of the COD community. Personally, I don't know what the typical kdr is, but I would say that it is well below 1.0, probably in the 0.75 to 0.80 area.
The "better" kdr's DO NOT carry enough weight to lift the typical kdr above 1.0. You only need to do two or three sample sets to see that reality. Still, like I said, to remotely get anything that could remotely be considered an accurate representation of kdr's or other data, you need to do 3 or 4 or more sets of 1000 sample sizes (because the community consists of MILLIONS OF PLAYERS). You MUST have large sample sizes in order to get a result with a strong confidence level. You could, also, just do one set of 4000 players or so, then do the confidence level math - and still you would not have an "accurate" number; you'd only have a number that you could a confidence level of 95% or so in reporting.
And all 4000 players would have to be selected randomly. Choosing to look at only players you have played against IS NOT a random sample. Sorry.
Too much to read? Ha. You should see the actual math that would have to be involved in figuring it all out.
Just trust those that have familiarity with statistics: the typical kdr is probably in the 0.80 area. Maybe less. And by typical, that includes probably 80% of the community as someone else already noted.
The number of players with above 1.0 kdr is VERY SMALL in relation to the overall size of the community.
And that's why Activision does not share the information.
Explanation was long but needed, the issue isnt simple!
i said by common-sense, but also by participant-observation (from playing CoD and for usually check some stats of some players), that the "overall KD" isnt 1.0. That may seem the most "accurate" answer but really? Do you really think the average KD is 1? Nope, I almost bet that is below that, and so I made my bet on the 0.75/0.8.
Because 1.0 isnt the average: the 1.0 "overall" is a number that very few people can actually achieve (considering the ammount of thousands that play this game). If you check lobbies youll see a huge number of people of the community is around 0.6-0.8 KD. That doesnt mean, obviously, that all this people cant have games of 1.0 Kd, or 2.0 KD, or even 3.0 KD. But doing that some games is attainable easily. Doing that in "a regular basis" is by far more difficult! If you think better even to have a 1.0 KDR you need to go 1.0 on most of your matches BUT, sib«nce we all have games where we go 0.5, you need a big ammount of games with 1.5 KD so you stay on the 1.0 KD and not on the 0.9 or 0.8 KD!
One thing i bet: 1.0 isnt the average!
That's essentially what myself and two others said. Mathematically, it is impossible for the overall average kdr to be 1.0. Suicides and team kills preclude 1.0 from being a possibilty.
I'm with you, Shadow_Alq ... typical player's kdr is going to be 0.80 or less.
Glad you noted that it takes a lot of matches with a kdr > 1.0 just to have a 1.0 kdr.
It's also worth noting that attempting to compile your own sample of KDR's (from lobbies) would be useless, aside from the fact that it's convenience sampling over random sampling.
Algorithims are used to place players in lobbies of comparable opponents. Some higher, some lower, always attempting a balance towards the center player of the lobby (though outliers exist). Samples used would hold the average you find to be more or less centered on your own statistics.
Obviously the 'average' COD player's KDR (and I think we all know that the assumption of average is in laymen conversation is the mean) is slightly below 1.0 (as others have noted due to the nature of non-opponent induced deaths).
If anyone is concerned if their KDR is normal or average it's fairly irrelevant. The pairing system is always attempting to evenly match you. If you improve your gameplay then the matching (as sloppy as it is) will continue to assess your data and pair you appropriately.
What would be interesting is getting access to the data and finding the standard deviation, and developing what KDR's fall into what percentiles.
The original question was what is considered the average k/d ratio for all black ops players. So were looking the average of ratios, 1.0 as you said is not the average of the ratios it is simply the ratio itself. 1.0 is the K/D ratio of the entire game, but it is not related to the average of many such k/d. The actual average is actually impossible to predict, it must be examined experimentally. If I had to take a wild guess, I would say it's between 0.6 to 0.8, but could be as low as 0.3. These numbers come from the fact that very few people who bought the game play multiplayer regularly enough to get good. So they play once in a while maybe getting 7 / 16, but because they constitue such a large number of the sampled people the average will be quite low.
As an attempt you can try to look at the leaderboards to see your ranking position, mine is around 311k with sps of 204 and k/d of 0.83, assuming k/d is slightly correlated to sps (which would make sense) that would mean that out of 3m people 311k are above or at 0.83 and 2.7m are below that k/d, you can see what the average comes out to.
it takes both guys trying to supress and guys running the caps . in my clan we alternate as 3 teammates cap flags and 3 hold down main routes to b flags. once we have that locked down we start spawn trapping . most of us only get 4 caps a game the spawn caps and b caps each round because once we have it locked down there is nothing else to capture ... triple capping forces random spawing which means were most likely to lose a flag behind us