Have to say that KPM is the closest individual stat to skill. Kdr is usually meaningless in games like etqw where it’s useful to run in, kill 2-3, die and spawn a second later. Pretty much what happens on refinery last stage. A high kdr is usually reached by not taking damage and quite surely a player with 27/24 k/d is better than a player who has killed 5 times and died only once per match.
Getting high amount of kills means you’re killing packs of enemies and that usually happens near the spawn, near the objective or at a bottleneck between the two. Can’t see how one could get a lot of kills by sniping on a hill.
ET-style battlesense has the same problem as kdr, it takes too much away from tactics involving dying, but i like how it requires taking a hit while killing to gain a higher multiplier.
Community Question: Measuring Player Skill
Killing players further away from their spawn means more time wasted on their part and thus a higher value kill. That’s why kpm is a perverse indicator.
Killing someone within 19 seconds is still close to the spawn and has him up and running in no time. In this context the respawn timer needs to be considered a random factor.
Only in your head toka. You’re talking about the difference between a very valuable action and an extremely valuable action. Its always good to kill.
The only situation in which KPM becomes so high as to be meaningless is complete shut-out spawn camping, which indicates hugely unbalanced teams and is an edge situation.
And it doesn’t bother you at all that the difference between a valuable and an extremely valuable action is completely unaccounted for in your indicator?
And let’s not forget the huge way classes, map, numbers of player on a server and being the defender or attacker skews this number.
Going for KPM is fun and I sometimes also do so, but it shouldn’t be used by an objective-oriented game itself to stimulate the people at frag whoring or rambo-medicing. Ones of most famous objective/CTF games (QuakeWorld TeamFortress, W:ET, etc.) ended their lives in a total ramboing mess if you will, and additional stimulus in form of KPM will only contribute to the early decline due to gameplay degradation. This **** is really infectious. Once this starts on a couple of servers, there’s no way to stop it. Next thing you see is people getting kicked from servers because they started playing objective… Been there. IMHO the only thing advancing this parameter will bring is more people caring less about objective. Even XP stat is a lesser evil because at least people try to construct objectives or build/destroy deployables, etc. Especially because you don’t get too much XP for keeling.
People who chose to specialise killing a lot and excel at it should be recognised. But so should people who chose to excel at healing or doing any other useful task in the game.
An indicator for skill is pretty much a infinitely complex system that would take into account all these smaller stats. But wouldn’t it be better to keep things simple? We don’t really need 100% precision. An objective based shooter has 2 parts, the combat and the tactics, but which one weighs more? If you can’t even kill an enemy, there’s little you can do with tactics, unless the enemy is totally disoriented and not even around the objective.
KPM is just a stat that a more complex system could be based on. Add some weights to kills, for example a bonus multiplier for killing the objective class, another one for the distance to the objective, etc. But lets not forget that every time we add such a rule, more exceptions arise. Is killing 5 players near their spawn really less significant if they were eventually heading for the objective? Is killing a lemming engineer near the objective useful if the objective bar is empty and the defending team has full control of the area? Even attempting to incorporate objective play into a skill stat brings more problems than it solves. The biggest problem is surely judging skill over a course of multiple matches. ELO only works for predetermined teams or clans.
Back to combat skill, i’d rather see battlesense slightly altered to add kills and deaths. Usually the game checks every 45 sec if the player is dead and rewards points if not, but i would add a forced evaluation after a kill or a death. This way a player who kills a lot while taking damage gains the highest amount of battlesense while a player dying while dealing damage is also awarded points. I doubt the regular time interval needs changing as the average kpm for a player is around 1.3 in etqw.
One thing is measuring skill, another is providing stimulus for playing for the objective. The two don’t really mix. This actually makes me wonder why it’s important to measure it? Maybe a behind the scenes calculation used to rank players without actually giving them the score would work.
if i had to make some automated system to find out who was the most important player in some match of etqw 12v12 i would choose it this way: it would be the player with most kills in the defending team, or the player with most xp in the attacking team (one of those 2 is the best, depending on the team who actually won).
but there are a few situations where this does not work so well … so i would includ this other thing:
a player can also be chosen from the losing team in any of these (uncommon, i.e., unbalanced teams) situations:
he was attacking and lost, but he got the most kills
he was defending and lost, but he got the most xp
PS: remember, 12v12 matches, not 5v5 etqwpro games (about those i have no guesses).
Think this is why should perhaps just offer some important basic stats at end of the round like we had in ET and ET:QW and let players make what they like out of them and throw in that idea about the random ‘outstanding achievements’ for that round. Those on the server at the time know who’s been xp farming and those who’ve actually been contributing to the game. Then on top of this also allow the community to gather other stats from the server to make their own leaderboards however they wish to base it on, think that’s what splatterladder does no?
What players prefer to do and what is the best contribution to a victory isn’t the same thing.
An indicator for skill is pretty much a infinitely complex system that would take into account all these smaller stats
All these smaller stats are only valuable if they don’t result in tangible feats within the game. A high accuracy is useless if it doesn’t lead to your opponents dying when you need them to. Staying alive or building battlesense or a k/d ratio is completely meaningless if that doesn’t lead to valuable contributions that in turn lead to winning the match.
That’s why xp/min is superior to anything you can think of. All the tiny important stats are already manifested within the xp score. The only premises are that the xp distribution is accurately (and not used as some dangling carrot to make players do things they ordinarily wouldn’t) awarding the things that matter and that the victory is holding a lot of weight in the final score.
People here have been thoroughly jaded by a few loopholes and the emphasis on flat xp score in ETQW and the way Brink completely butchered any meaning out of that score by skewing it to reward certain behaviour, regardless of whether it’s valuable or not. Nobody wants that again.
However, to say that players can make up themselves what’s important is just foolish. You only create lots of people chasing their own little agenda regardless of what’s best for the team. Not everything a player values is important in the end. The only way to incorporate a multitude of valid play-styles is to have them weighed in a single score. That’s why xp/min is the only way you can put a fragger, a supporter, a saboteur, a sniper, an objective lemming and a pilot next to each other. It’s the only system that assesses these people merits without ambiguity.
in ETQW the main problem about XP/min is the limited number of vehicles. if everybody could choose any kind of vehicle to play, this xp/min would be much more precise. but since most maps have one flyer per team, this creates an unbalanced situation from the start.
ratio win/loss is good, but not perfect. people will stack the easier side, and also will leave if they are loosing. but it is where you can see the good players, those who are willing to sacrifice their individual score to achieve victory. meaning they will choose the class the team is lacking, and will do those little things that dont get them kills or xp, but result in victory (like gibbing when defending, or destroying mines when attacking, or many other things i cant remember right now).
kills is not good measure for skill. and yet, try to win some map defending when your team is doing nothing or afk. no way you can do it without getting the most kills award in the end.
the w:et battlesense style is not good either. it would turn etqw into a counterstrike game style. with everybody camping instead of working to complete objective.
PS: either way (kills or xp) ins’t good because people will abuse it. they will attack when they are suposed to defend and vice-versa. so this leave us with the hardest thing to abuse: victories.
but without some sort of autobalancing, out of the server admins control, it is also useless to see who won and who lost.
Don’t worry, we’ve been over this already… XP/min is a terrible, broken, flawed, easily manipulated idea. He knows this, he’s just stubborn is all.
Yeah, 12 flyers for each side would definitely fix your xp-whore problem.
I also got an idea, 24 tanks/cyclops on the map so we don’t need the stupid infantry weapons anymore 
Nowai! You forget that they would also need an unlimited supply of deployables to farm. Even then, it doesn’t give a great definition for skill.
All these flat statistics for rating skill are all pretty meaningless if they don’t take into account your opponents’ skill rating.
It’s a fair point, one flyer allows one person -not necessarily the most skilled- in the team to be a very effective asset and thus garner more experience. It’s not just a problem for skill indication, the entire notion that one player gets so much power by simply being the first to grab it is problematic regardless of the scores he gains from it. A game like BF3 has even a bigger discrepancy while a game like Brink has none of this.
It’s worth looking into toning things like these down and allow multiple weaker vehicles over individual powerful vehicles. Another solution would be to push powerful vehicles more into support roles and/or make them more dependant on their team. Powerful land vehicles being more dependant on repairs in order to be more effective and powerful flyers taking up roles as supply drops, spawning players form drop-pods or parachutes and having their offensive potency tied to laser-painters on the ground.
ratio win/loss is good
Only if it has a high enough frequency. In 1v1 this required frequency is already quite high due to the random opponents but in team games, where external factors increase by several magnitudes, this frequency needs to be even higher for it to have any statistical significance.