Let's talk K/D for a moment...

It is my understanding that TrueSkill is used to ensure that players will play against those of their own skill level. For example, on a 1-50 scale, if you have a 40 and a 20, the 40 would be matched with other 40s and the 20 will be matched with 20s.

If the 40 and the 20 played against each other, the 40’s K/D would reflect something better than that of the 20’s. Perhaps the 40’s K/D would be 2.5 and the 20’s 0.4 (for example).

Now take the 40 and play him against a 40. His K/D should be about 1.0.

Since TrueSkill is the context of the game play, the 40 should not play against the 20, and thus should have a K/D reflecting his play against another 40. This is true for every level of skill. So each person’s K/D should be around a 1.0 when TrueSkill works correctly.

TrueSkill doesn’t always work well, we know that. But the fact that it works at all skews the results toward 1.0.

It makes no sense to say that someone’s skill is based upon their K/D, since that number is skewed by TrueSkill in the first place.

Can anyone refute what I just said and explain why K/D reflects a person’s skill?

> Can anyone refute what I just said and explain why K/D reflects a person’s skill?

Nope, you hit the nail on the head. It stands, though, that Halo 4’s matchmaking will put people of uneven skill levels on a team to make the average skill levels of each team somewhat equal, like Halo 3’s social matchmaking system. This allows parties to play together without heavily skewing the matchmaking process.

KD reflects broadly reflects skill.
A person with a .5 KD or less can be labeled inept.
A person with a .8 or higher KD can be labeled Competent.

That about sums it up. K/D isn’t a hard, static indicator of skill. Saying, “I have a 2.5 K/D” really doesn’t mean anything, because it all depends on who you played against to get that stat.

If it were static and accurate, you could theoretically put a team together with players with a positive K/D and always win. In order to get positive, someone has to go negative.

My fine friend, kill to death ratio is so important in Halo 4 because TruSkill barely works! At least it does not for me. Perhaps I do not play the playlists with adequate populations to always match me up with competant players, or perhaps it’s something else…but if it does not work, then you’re matched with people at complete random instances. This said, if it’s completely random, a high ratio goes to show that you succeed many times no matter who the adversary is.

Provided TruSkill did work well, a 1.0 K/D would be…respectable.

> …
> Provided TruSkill did work well, a 1.0 K/D would be…respectable.

No, that was my point. If TrueSkill worked as it should, the K/D would be nearly entirely meaningless, because it would only reflect your skill within a narrow range of players with the same skill.

If you had player A with 1.0, and player B with 1.0, and they played each other 10 times and A won every game, your only conclusion would be that A was a higher skill and that they both achieved their K/D of equal value by playing against players of their own skill level up until they matched against each other.

K/D is pretty much meaningless, except that TrueSkill doesn’t work well, or you are in a party with people of another skill, etc. It is because we don’t know the history behind any one K/D value, it is meaningless - it is without context, and thus has no real meaning in any discussion.

> > …
> > Provided TruSkill did work well, a 1.0 K/D would be…respectable.
>
> No, that was my point. If TrueSkill worked as it should, the K/D would be nearly entirely meaningless, because it would only reflect your skill within a narrow range of players with the same skill.
>
> If you had player A with 1.0, and player B with 1.0, and they played each other 10 times and A won every game, your only conclusion would be that A was a higher skill and that they both achieved their K/D of equal value by playing against players of their own skill level up until they matched against each other.
>
> K/D is pretty much meaningless, except that TrueSkill doesn’t work well, or you are in a party with people of another skill, etc. It is because we don’t know the history behind any one K/D value, it is meaningless - it is without context, and thus has no real meaning in any discussion.

Again, it also has to do with the matchmaking system matching people based on team average, rather than individual trueskill.

to me k/d in halo dosent really matter. i try to keep my kd around 1.00 so i can still play with average players, but not get completely squashed by the really good players with a high kd. Also a high kd dosent mean anything anyway because it isn’t hard to camp a base on dominion with a rocket launcher with the ammo perk on, or to camp in btb. I’ve tried and i went 16-2

> Now take the 40 and play him against a 40. His K/D should be about 1.0.

Assuming of course that they both will forever be 40s i.e. they deserve to be rank 40. A 40 on his way to 41 can potentially wreck a 40. 50s were 40s were 30s were 20s were 10s.

Slaying is how you keep enemies off an objective and how you take it for yourself. You can only do anything when you are not on a spawn timer.

Slaying and staying alive are fundamental to winning. You don’t do either and you can’t win, on the other hand doing both really well doesn’t mean you are winning. What ends up being the case is that a really good player and a really selfish player are indistinguishable. W/L is also an imperfect metric, as is every single metric. The whole picture will tell you a lot more about a player’s ability.

Could you explain, then , how I have a 2.3, and I am really not good at all?

K/D to me means nothing. When I’m playing with a full team (which is rare but has been happening more often) we basically NEVER lose and constantly go positive. Even against 2.0 K/Ds or better. The game honestly doesn’t even try to match randoms against randoms and teams against teams. Most of the time when me and a full team play against 2.0 K/D’s they end up quitting early. Can you really count the K/D of someone who will only stay in the game against terrible players? K/D isn’t a true way to measure skill in any Halo game but in Halo 4 and Halo Reach K/D is specifically less of a skill indicator. W/L is also a pointless stat considering (like said before) the game doesn’t attempt to match evenly what so ever.

Who judges players on their K/D? It is expected to drop anytime. For example, a guy was sneaking his way to our flag but I sniped him with a Binary Rifle out of the blue, which is a random and unfair death. He ended up getting the second-highest score of the match with a flag capture for his team.

You’re right, but the system does a pretty bad job occasionally. A guy could be very skilled, earned a few random deaths like the one mentioned, and is pitted against opponents that are easily conquered. This lowers the OPPONENT’S K/D with the victor earning his/her rightful place. The repercussions cause unhappy whiners to emerge.

You’re right. K/D is the most random fact a player can base their skill on.

> Could you explain, then , how I have a 2.3, and I am really not good at all?

This is why I say K/D is a reflection of how many points you GIVE to the enemy, not how good you are at slaying. It is a measurement of conservative play.

I can control my K/D by concentrating on hiding and surprising the enemy that comes along my path. Or I can try to gain kills and let my K/D slip.

It is also a reflection of being patient enough to wait for the Scorpion on Exile BTB while being conservative and careful not to get into a fire fight. I have had 30/3 from games like that.

K/D means nothing. It can easily be manipulated by how you play, meaning you don’t care about your team at all and you only stay alive and kill cheaply. Your win loss ration shows how good of a player you are. That will always be my opinion towards that topic.

You have a point there. Trueskill is flawed to the point that you aren’t always matched with similarly skilled players; however, K/D can’t just be thrown out the window. It’s not even close to being a good measure of skill, but it is good for being a rough indicator. For example, if you’re positive, you usually help your team, and if you’re negative, you’re most likely a detriment. FFA is different.

> Could you explain, then , how I have a 2.3, and I am really not good at all?

Your service record:

515 victories
511 defeats

> > Could you explain, then , how I have a 2.3, and I am really not good at all?
>
> Your service record:
>
> 515 victories
> 511 defeats

Well honestly win/loss only matters if you play with a team, unless of course we’re talking about FFA playlists. For example as a random playing team slayer you’re only responsible for 1/4 of your teams overall score. Also as generslly every other game is an already lost JIP match win/loss ratio really doesn’t give a very accurate portrayal of a randoms skill.

On the other hand when I’m playing 4v4 slayer and I see a player with over 500 games under his/hers belt and a positive K/D ratio, it’s more than likely that said person will go positive in the next match.

> > > Could you explain, then , how I have a 2.3, and I am really not good at all?
> >
> > Your service record:
> >
> > 515 victories
> > 511 defeats
>
> Well honestly win/loss only matters if you play with a team, unless of course we’re talking about FFA playlists. For example as a random playing team slayer you’re only responsible for 1/4 of your teams overall score. Also as generslly every other game is an already lost JIP match win/loss ratio really doesn’t give a very accurate portrayal of a randoms skill.
>
> On the other hand when I’m playing 4v4 slayer and I see a player with over 500 games under his/hers belt and a positive K/D ratio, it’s more than likely that said person will go positive in the next match.

For a guy who who claims to be so good because his kd is high, you would expect a better win:loss ratio than 1:1, AT LEAST after 1,000 games.

> > > > Could you explain, then , how I have a 2.3, and I am really not good at all?
> > >
> > > Your service record:
> > >
> > > 515 victories
> > > 511 defeats
> >
> > Well honestly win/loss only matters if you play with a team, unless of course we’re talking about FFA playlists. For example as a random playing team slayer you’re only responsible for 1/4 of your teams overall score. Also as generslly every other game is an already lost JIP match win/loss ratio really doesn’t give a very accurate portrayal of a randoms skill.
> >
> > On the other hand when I’m playing 4v4 slayer and I see a player with over 500 games under his/hers belt and a positive K/D ratio, it’s more than likely that said person will go positive in the next match.
>
> For a guy who who claims to be so good because his kd is high, you would expect a better win:loss ratio than 1:1, AT LEAST after 1,000 games.

This is just me trying to cover my own -Yoink- but I find searching solo somewhat decreases your chances of winning.

Try as I might, I cannot capture every base in dominion alone.

As others have said k/d should theoretically be 1.0 if TrueSkill is working properly. That presupposed a large enough population for the mechanics behind TrueSkill to make proper pairings AND that k/d isn’t artificially inflated.

As a stat to measure a player’s skill in H4 k/d is meaningless. This is mainly due to the fact that k/d’s have been artifically inflated due to things like large amount of power weps on maps and in players hands. Also we have to look at TrueSkill pairings and if those pairings are correct. If I had to guess, I’d say TS is only making proper pairings between 25%-30% of the time.

Plus, when the highest population playlist is BTB… well then throw k/d out the window as a stat to measure player skill at all.