Is K/D really that accurate?

Games Other Halo
1

Is it really a good measure of skill? I personally find that it isn’t really an accurate measure of actual skill.

I mean, if someone has 10000 kills and 5000 deaths, sure, they may have earned through normal gameplay, but they also could just be playing Grifball, in which it’s much easier to do good in than in normal playlists. Even my 2000 kills more than deaths are mostly due to Grifball lol. So with this in mind, could it really be considered an accurate measure of skill? Especially in Halo 4, seeing as how it has been made more accessible, thus easier to do good in.

On the topic of skill, has anyone really noticed any difference made by the introduction of CSR? Personally, I really haven’t, not that it bothers me, but I’m just curious if anybody has noticed any significant difference in their matchmaking games with the feature.

2

My K/D is 0.03 points lower than Ninja’s (MLG player for those who don’t know, and probably the most active MLG player on Halo 4.)

What are the odds that I’m anywhere near his skill range?

I’ve got a better K/D than other MLG players I played and still got kicked into the dust.

The moment you pass 1.0 K/D is where it stops being a measurement of skill.

As for CSR, I haven’t noticed any difference. The player pool is too shallow for CSR to actually function properly.

3

If the player has a high k/d and only plays slayer then that’s pretty good. I have noticed a slight increase in the competition in games since csr came in but still a lot of csr 1 and very unskilled players in almost every game I play.

4

k/d plus win effectiveness = skill

5

> My K/D is 0.03 points lower than Ninja’s (MLG player for those who don’t know, and probably the most active MLG player on Halo 4.)
>
> What are the odds that I’m anywhere near his skill range?
>
> I’ve got a better K/D than other MLG players I played and still got kicked into the dust.
>
> The moment you pass 1.0 K/D is where it stops being a measurement of skill.
>
> As for CSR, I haven’t noticed any difference. The player pool is too shallow for CSR to actually function properly.

I agree.

I simply feel that in BTB that if you drop at least 140 pts and have a positive K/D then are you an asset to the team.

I have had good games with double figure assists but with a slightly negative K/D. I still feel that I contributed is softened a ton of enemies for my team mates.

6

> k/d plus win effectiveness = skill

Unless you are comparing two people who play differently and in different conditions.

Someone who plays with a full team all the time will undoubtedly have a higher win percentage than someone who plays as a random. Even if they’ve got the same K/D and but different win percentages, there is no way to determine who the better player is because one person has better conditions than the other.

Now, lets assume they have the same win percentage but different K/D. According to you, the person with the higher K/D is the better player, but what if the lower K/D player play objective and is the designated objective player? So they’re the one who runs the flag and carries the ball. They are undoubtedly going to have a lower K/D because of this, but are they the worse player?

There is no way to tell who the better player is through stats. You can only get an approximation of their skill level.

Someone with an 85 win percentage and a 1.5 K/D will (probably) be better than someone with a 42 win percentage and a 1.1 K/D.

> > My K/D is 0.03 points lower than Ninja’s (MLG player for those who don’t know, and probably the most active MLG player on Halo 4.)
> >
> > What are the odds that I’m anywhere near his skill range?
> >
> > I’ve got a better K/D than other MLG players I played and still got kicked into the dust.
> >
> > The moment you pass 1.0 K/D is where it stops being a measurement of skill.
> >
> > As for CSR, I haven’t noticed any difference. The player pool is too shallow for CSR to actually function properly.
>
> I agree.
>
> I simply feel that in BTB that if you drop at least 140 pts and have a positive K/D then are you an asset to the team.
>
> I have had good games with double figure assists but with a slightly negative K/D. I still feel that I contributed is softened a ton of enemies for my team mates.

This is a good point too. Assists definitely mean a lot, because you still contributed a very good amount to the game. You might not have gotten the kills, but your teammates might not have either had you not been there to help.

I like the points in this game. It actually rewards assists and objective play, which is a good thing.

Someone might have gotten 30 kills, and they were useful to the team. But the guy who got the 4 flag caps was probably a lot more useful because he was a good flag runner.

I know that doesn’t exactly apply to BTB, but it helps.

7

I can look at their service record in game and look at k/d then look at how many wins they have…ive played enough to know how many wins a good player should have along with the amount of kills. At this stage if you dont have 1000 wins by the time you reach 20,000 kills, thats not good enough.

I think a decent player should have at least 500 wins per 10,000 kills

8

It’s only a mild indicator. A negative K/D says a lot more than a positive one. Though if it’s a super high one (2.0+), then there are one of four things that goes through my mind when I see their service record:

1: Stat pads in objective.
2: Plays BTB with a party.
3: Abuses the hell out of the broken things in this game.
4: Is legitimately that good. (Quite rare)

K/D matters, but not as much as some people think. A near-even or negative ratio will tell your more about one’s skill though compared to very/super positive to be honest. If slayer-based playlists showed you everyone’s K/D in just that specific playlist, then it would matter a lot more.

Halo 5 should display K/D for slayer playlists on player cards, and W/L for objective-based/MLG. Individual playlist stats like those being displayed at least somewhere in-game would help a lot, possibly more than ranks would. H5 should use both though.

9

> I think a decent player should have at least 500 wins per 10,000 kills

Eh that’s a decent way to look at it. At least for team based games. I’m sitting on about 10k kills now with a bit over 500 wins so I’ll roll with that.

10

> Is it really a good measure of skill? I personally find that it isn’t really an accurate measure of actual skill.
>
> I mean, if someone has 10000 kills and 5000 deaths, sure, they may have earned through normal gameplay, but they also could just be playing Grifball, in which it’s much easier to do good in than in normal playlists. Even my 2000 kills more than deaths are mostly due to Grifball lol. So with this in mind, could it really be considered an accurate measure of skill? Especially in Halo 4, seeing as how it has been made more accessible, thus easier to do good in.
>
> On the topic of skill, has anyone really noticed any difference made by the introduction of CSR? Personally, I really haven’t, not that it bothers me, but I’m just curious if anybody has noticed any significant difference in their matchmaking games with the feature.

No. People back out everytime they see someone that “looks” like a challenge, so they can increase their K/Ds and W/L against worse players instead.

Either way, it isn’t a very good way to measure skill.

11

> I can look at their service record in game and look at k/d then look at how many wins they have…ive played enough to know how many wins a good player should have along with the amount of kills. At this stage if you dont have 1000 wins by the time you reach 20,000 kills, thats not good enough.
>
> I think a decent player should have at least 500 wins per 10,000 kills

Well, I’m at 9400 kills and only 346 wins, and would consider myself a more than decent player.

It’s hard to base anything on stats with things like JIP in the game, and the fact that some people play with teams and others don’t.

Then there’s things like Internet Connection to bring into a count, along with a whole lot of other things.

You have no idea if the person is sacrificing their stats for commendations (like plasma pistol) or if they play gametypes like flood where a loss is a lot easier to obtain than a win.

I myself play a LOT of FFA (or I did, until FFA got boring as all hell because it’s terrible in this game) and consider myself a good FFA player, but with JIP and the fact that it’s you against 8 other people (and largely luck based when you combine 8 people, instant spawns and sprint) you are much, much more likely to lose than if you played a team based game.

Gotta take stats with a grain of salt. There is a LOT to take in when looking at a persons stats, and many different scenarios that would have a massive affect on it.

12

It can if you look at it the right way. But, you will have players who are good, but look bad since Assists are not considered into the ratio.

It can paint a picture, but not a detailed picture. You will have to look further than that in order to see the whole story, unless that is really all there is to it.

One thing I can say is this: Never use the in-game kill and deaths stats. It takes everything, making the player look better than he should, or worse.

13

In my opinion yes, k/d is relatively an accurate representation of a players ability to do well in game. In my experiences when i see a person with a low k/d they generally play poorly compared to other players. Same goes for players with higher k/d, they tend to play better. K/d is not entirely accurate but i would say from my experiences that 8/10 times a player with a low k/d will play poorly and a player with a higher k/d will play well.

Also i have not noticed anything different since csr has been released, which is quite disappointing.

14

Is K/D the best way to measure skill? No.

Is it somewhat useful in measuring skill? Yes.

Realistically no one is going to have a K/D of 2.5+ by playing against good players in competitive playlists (hence why most pro players have K/D’s between 1.5-1.7). Someone with a K/D of 1.4 but 7 CSR 50’s is going to be better than someone with a K/D of 3.0 but whose highest CSR is a 20 in Action Sack.

K/D isn’t a bad way to measure skill, but factors such as CSR and what playlists a player plays are much more important.

15

> I can look at their service record in game and look at k/d then look at how many wins they have…ive played enough to know how many wins a good player should have along with the amount of kills. At this stage if you dont have 1000 wins by the time you reach 20,000 kills, thats not good enough.
>
> I think a decent player should have at least 500 wins per 10,000 kills

Oooh… so close. I only have 476 wins and 9887 kills. Do I get extra credit since I usually play by myself, never use Camo, PV, Boltshot, or Gauss Hogs and I always rush the objective?

16

I hope K/D isn’t too much of a skill indicator, because then it would show how much of a failure I am, lol. I’ve got a .9 K/D (600 more deaths than kills, approximately). Does this mean I am terrible? I hope not. :stuck_out_tongue:

17

I don’t have a high k/d about 1. 26 but my win ratio on objectives are high.

I don’t mind going negative as long as the team wins

18

> > I can look at their service record in game and look at k/d then look at how many wins they have…ive played enough to know how many wins a good player should have along with the amount of kills. At this stage if you dont have 1000 wins by the time you reach 20,000 kills, thats not good enough.
> >
> > I think a decent player should have at least 500 wins per 10,000 kills
>
> Oooh… so close. I only have 476 wins and 9887 kills. Do I get extra credit since <mark>I usually play by myself, never use Camo, PV, Boltshot, or Gauss Hogs and I always rush the objective?</mark>

This sounds a LOT like me. <.< | >.>

The only difference is that You K/D is .42 higher than mine. :stuck_out_tongue:

19

Neither K/D nor wins are a great measure of skill because of skill matching.

All your K/D says is that, against players with similar TrueSkill values (since that’s who the system will generally match you against), you are able to kill X.XX of them for every time you die. If, however, you were matched against people with higher TrueSkill values, your K/D would suffer greatly.

Win rate is not that great, either, depending on what type of game you play. So if you like multi-team or FFAs, your win rate might suck but you’re still fairly good.

CSR was an attempt to fix that. For the Team CSR scores, it probably does what it’s supposed to. All that matters for the team one is Win/Loss. This is not really an individual measurement . . . simply because the Patriots have a good W/L percentage does not mean that every individual on the team is good, or even contributes. But then again, it’s not meant to. It just shows the strength of the team. So team CSR does give an indication of how well you play the objective in those types of games.

Individual CSR is more problematic. What is really needed is a true statistical measurement of how likely you are to win an engagement against an opponent with a certain TrueSkill value. Ideally, your individual CSR number should correspond to the point at which you will beat an identically ranked opponent 1/2 of the time. Because it just uses score, though, this doesn’t work out in practice (and JIP messes it up, too). But it’s an attempt - if a slightly poor one - to answer the individual skill question. Although, just showing the TrueSkill number would be far more accurate.

However, an objective “overall” skill ranking that combines both individual and team skills is not really possible. You first have to decide how much to weight team elements vs. individual elements, and that is arbitrary.

So, unfortunately, there’s no real way to answer the question unless you rise to MLG level . . . at which point the question is answered by whether you win or lose in the controlled environment presented during the competition.

20

As a baseball fan there is an analogy I like to use for K/D. For years baseball used battin average as the defining statistic for how good a hitter was. Recent research over the past 15 or so years proves this to be false… Obp, OPS,Wrc+, war… All these new metrics better reflect a players skills and value.

I feel KD is the batting average of halo. A good starting point but by no means the whole picture. Assists, performance in objective based games, and general teamwork play a far more important role than can be measured. So when people #humblebrag about their KD I don’t really listen too closely… It’s only part of the picture.

As for CSR, I have noticed an increase in the number of teams, especially in objective, slayer, and dominion. Kinda rough on those of us who don’t always party up, but I do like a challenge.