{{
`${userDetail.user.displayName}${userDetail.user.cachedBungieGlobalDisplayNameCode ? `#${standardizeBungieGlobalCode(userDetail.user.cachedBungieGlobalDisplayNameCode)}` : ""}`}}
{{getUserClans()}}
Edited by ScientificBEST:
1/16/2016 1:49:56 AM
12
[b]Wrong! Wrong! Wrong! :)[/b]
You're not considering the fact that a K/D ratio above 1.0 influences the average more than a K/D below 1.0.
For illustration, let's say there is a 2-player match, 1v1. Player A kills Player B 10 times and Player B kills Player A 5 times.
[b]Player A finishes with a K/D of 2.0, while Player B has a K/D of 0.5. Now let's take the average:[/b]
[b](2.0 + 0.5) / 2 = 1.25[/b]
[b]As you can see, a high K/D tends to skew the average more than a low K/D.[/b] Why? Because the K/D score is based around 1.0, and is expressed as a fraction. The upper bound of a fraction is infinity, while the lower bound is 0. Naturally the numerator (kills) is going to be more influential then the denominator (deaths).
[b]OP is incorrect when he says the overall average is 1.0. Higher K/Ds skew the average more than low K/Ds, because the K/D is expressed as a fraction. Kills influence the K/D value more than deaths. Therefore, the actual average K/D of the Destiny community is probably closer to 1.15 if I had to guess.[/b]
[Edit]
I'm not positive, but I think if you take the inverse of K/D, which is D/K (deaths over kills), and take the average of those two numbers, it comes out to 1.0 exactly. So:
(K/D - D/K) / 2 = 1.0
[Edit 2:]
Turns out my math is wrong in the first edit. You get 1.0 if you [i]multiply[/i] K/D and D/K together. But that is just a mathematical truism so it's not super helpful.
The point is that the "average K/D of all players" is not derived from the global tally of kills/deaths. It is derived from the individual K/Ds of each player. So given what I've said above, you can't say that an average K/D is 1.0. That's decidedly below average.
English
-
Dont do it, these posts attract full blown potatoes which just dont understand it " but if a team gets 80 kill that means there was 80 deaths That means the K/D was 1 dumb dumb head".. i cant even anymore.
-
Lol, yeah, I feel ya. This usually flies over a lot of heads before someone gets it.
-
Edited by bilbo t baginzz: 1/16/2016 4:08:59 AMIf theres 80 kills in a game there will be 80 death (not including suicides) which is an obvious thing and people dont realise that this cant be used as a reference to someones "skill" or average K/D.. We need to take the average individual K/Ds of each player to work out if they are "above average" or not. I find it really hard to find a game which i have recently played where the average K/D is 1.0 or just below.. amost all the itme its around 1.05 or above Also looks like OP now understands as well.
-
Edited by ScientificBEST: 1/16/2016 8:18:04 PMYeah, honestly you probably explained it there better than me. Or it's probably less confusing lol. But yeah, I mentioned something about it on Reddit. Basically I went back through a bunch of recent games of mine and took the average of each match, which was generally around 1.12 or 1.13. Another thing to think about is how that number might be useful in other ways as it varies from match to match. Basically the higher the K/D average, the bigger the gap between the top and bottom of the scoreboard. So a match with an average K/D around 3.00 was probably a one-sided murderfest, whereas a 1.10 would suggest a fairly even match of skill. So, I bet if you took the global average K/D, and tracked it week by week, you could get a pretty good idea of how well the matchmaking system is working, or perhaps how a recent weapon rebalance has affected PvP outcomes. You would have to leave one of those things alone to disentangle the data, but that's probably something Bungie looks at. For example if it changes from 1.2256 to 1.2175, then you know either (A) a small change in the matchmaking algorithm has improved the quality of skill-based matchmaking; or (B) when they nerfed a subclass or buffed a certain weapon class, it actually improved overall PvP balance. That seems like a small change, but those changes really do matter. Anyway, I'd say the observed average of about 1.12 is a good thing. You can go 5.0 every now and then, and 15.0 on rare occasions, but for the most part a good competitive player will hover around 1.8-3.0. And for the rest of us plebeians, a 1.13 will have to do. ;)
-
Yeah me too. I'm pretty sure it's actually impossible unless everyone goes exactly 1.0 or there are just a TON of suicides. The math simply won't allow it otherwise.
-
I feel as though you should make a post correcting just about everyone out there.. ive done so a few times but i find it difficult to explain or put into words yet you seem very switched on with it all lol. [quote]As you can see, a high K/D tends to skew the average more than a low K/D. Why? Because the K/D score is based around 1.0, and is expressed as a fraction. The upper bound of a fraction is infinity, while the lower bound is 0. Naturally the numerator (kills) is going to be more influential then the denominator (deaths).[/quote] This part is great.. if someone gets 50 kills on someone and the other person gets 50 deaths 1 persons K/D is 50 and the others is still 0 not - 50 which explains perfectly how the positive K/D outweighs the negative ones.
-
Oh boy, if I had a post like this on everything, I could probably swim in my own hate mail. But thank you. Takes one to know one. :)
-
I made a post about it... Hate mail inbound!!! haha
-
This. Your edit is what I've always said. .5 KD basically represents a negative number (for simplicity sake). So you have to inverse to get proper KD. True KD, discounting suicides, is exactly 1. But the average KDR (ratio) is not a true representation for the exact reason you said, it weights the positives more. The true average KD across all players is 1. The GG averages and bungie's averages just pull the straight ratios without the proper conversions.
-
Edited by ScientificBEST: 1/15/2016 10:57:34 AMExactly! Finally someone gets it! Haha. I was also thinking about how the game will record the same K/D if you have just 1 death as it would if you had no deaths at all. Obviously you can't divide by 0, so it is necessary to do that for practical reasons. But it does kinda mess with the data, I assume, much like suicides do.
-
I've had games with just 1 death and it records it the same as 0 deaths. A 10-1 game records the same as a 10-0 game, 10.0 k/d.
-
Yes, so that alone would probably tend to counteract the influence of suicides on the overall average K/D.
