You might think he is playing against rookies, and he could be. In higher level pubs I still have some gorges do this to me.
Tufts is amazing as a gorge, but he's still more effective as nearly anything else. When I Nerf myself its always going gorge. Or pistol. Spit efficacy falls off drastically with armor upgrades until bile bomb, then some more with wep ups.
I've gotten fantastic kds with gorge and decent kill totals. But I would hav gotten far Whittier kds but double the kills as a skulk.
Gorges just can't kill things fast enough. That's not their strength... You want to see lopsided though?/ try getting a base up in their natural. If they don't push it hard enough, it turns into a free win.
Tufts is exactly one of the gorges I speak of.
Isn't that what the mathematical model is supposed to account for though. Since we can't quantify skill directly, we try to estimate it in most circumstances. People who go gorge probably do so often and this would be reflected in their hive score. If not, then they may lose that round.
I think that something like a good player going gorge is something unaccountable for in hive skill.
Turts rather. Going gorge is when you want to dick around though. They'll almost always lose the round if the teams are balanced.
I knew you meant turts, but typed tufts just in case. In most games I end up playing with turts, which is not many, I see him gorge all the time.
Again though, Isn't that what the mathematical model is supposed to account for though. Since we can't quantify skill directly, we try to estimate it in most circumstances. People who go gorge probably do so often and this would be reflected in their hive score. If not, then they may lose that round.
You might think he is playing against rookies, and he could be. In higher level pubs I still have some gorges do this to me.
Interisting video.
He aims always exact the center of the target, prefect tracking. WOW.
That is besides the point. That player doesn't even play ns2 anymore. Turts is the real deal super gorge though. I am sad to say he has at least once chased me across the map to kill me and survived. Silly comm didn't med.
Isn't it a bit silly to try to balance skill purely around win/loss when so many people F4 or disconnect when losing with no repercussions against their ranking?
Isn't it a bit silly to try to balance skill purely around win/loss when so many people F4 or disconnect when losing with no repercussions against their ranking?
Afaik, your score goes up or down partially based on your input on the game. One of the factors of your input is time spent in game. If you were there only 1/2 the round you got half the effect.
Isn't it a bit silly to try to balance skill purely around win/loss when so many people F4 or disconnect when losing with no repercussions against their ranking?
The hive skill is changed based on the time you spent on a team. If you F4 or disconnect it will still update your hive skill. As you team is even more likely to loose when you do this, your hive skill will decrease.
Isn't it a bit silly to try to balance skill purely around win/loss when so many people F4 or disconnect when losing with no repercussions against their ranking?
The hive skill is changed based on the time you spent on a team. If you F4 or disconnect it will still update your hive skill. As you team is even more likely to loose when you do this, your hive skill will decrease.
Yep, and the beginning of the game counts exponentially more than the end. So if you abandon a game and your team loses, your skill will still go down by a similar amount even if it takes a long time to finish. If you join a game late, you won't be punished or rewarded much for whatever happens in the game.
On gorging, IMHO gorges win games. Most of my wins are from knowing how to place a deep tunnel and when to bile rush the power or whichever forward base the marines are paying the least attention to. Naturally if someone is gorging just to have fun with it rather than because it's what their team needs, they aren't as likely to win as the model predicts. Having your bases constantly under harassment so you can never make forward progress is similar to the feeling of fighting a marine who is so dead on accurate you just can't get close to him, so I think this is working as intended.
Yep, and the beginning of the game counts exponentially more than the end.
Why? I had a match today in which I lost 28 pts in a 2 minute romp. Was a complete stack, skills werent even remotely close. Was just a 28 point tank. In other games where skills are balanced, I will win like 5 games in a row and gain like 20 pts. If skills are "balanced" like 1200 vs 1200, then shouldn't I be gaining more than a couple points at a time if I win those matches?
Yep, and the beginning of the game counts exponentially more than the end. So if you abandon a game and your team loses, your skill will still go down by a similar amount even if it takes a long time to finish. If you join a game late, you won't be punished or rewarded much for whatever happens in the game.
The first part shouldn't work like this as @2cough mentions. Joining late doesn't reward or punish you much which is good, however the same should happen if you abandon early. Sure it makes it harder for the team that was left in the rain, but you never know who might fill the spot. Also teams and skill-balance could theoretically change a lot during a long game.
Let's imagine this situation: Someone leaves 1min into the game, because his team consists only of low skilled players and the enemy team is stacked. However the game drags on for whatever reason. Many of the low skilled players leave the game and get replaced by very high-skilled ones. Though too late and they loose... What happens to the skill of the guy that left early? @moultano.
Honestly I don't know what would be ideal and easy to calculate for that player. You can't just do nothing with his skill points, otherwise people would just leave a game when you they see they're loosing. It should somehow consider playtime and track the teamskill during playtime. And then... ?!?
@moultano
Some thoughts on the current implementation:
* I've noticed that FET moves players to make the teams having the exact same skill rating. If you allow for a small difference in skill between teams more players should be able to stay on their chosen team.
This seems reasonable, what would you choose as an acceptable bias? What should the probability of victory be before we say it's good enough?
@moultano I think around 2-4% skill difference between the teams should be ok.
Yep, and the beginning of the game counts exponentially more than the end.
Why? I had a match today in which I lost 28 pts in a 2 minute romp. Was a complete stack, skills werent even remotely close. Was just a 28 point tank. In other games where skills are balanced, I will win like 5 games in a row and gain like 20 pts. If skills are "balanced" like 1200 vs 1200, then shouldn't I be gaining more than a couple points at a time if I win those matches?
You should. The only variable there is the number of players in the match. You should gain or lose 100 * 8 * 0.5 / num_players points from a match of even skill. Is it possible that the composition of the teams changed a lot over the round?
I still think this method is somewhat unfair and unrepresentative of the individual player. It also allows for Hive score mining. Something more encompassing of the player score during the match, in conjunction with the minutes he played would far better represent his contribution and overall skill.
however the 2 min round in question has already been cleared out by other games. And I don't remember it changing a lot, was too short, really. Maybe everybody f4'd on that round? I can see that screwing up the avgs . There was 2 today, skill disparity less than 200 pts, gained 8 on one, 6 on the next. (1st 2 games of march 4th). Also, not sure why it thinks it's march 4th already...
I think I'd like to see player's scores/avg. scores incorporated somehow. I think that shows a player's efficiency in a round more than just w/l. Shows how productive one is on a team, win or lose, which I think is more indicative of skill.
Just a thought. Still think it's far better than previous iterations.
I still think this method is somewhat unfair and unrepresentative of the individual player. It also allows for Hive score mining. Something more encompassing of the player score during the match, in conjunction with the minutes he played would far better represent his contribution and overall skill.
I think I'd like to see player's scores/avg. scores incorporated somehow. I think that shows a player's efficiency in a round more than just w/l. Shows how productive one is on a team, win or lose, which I think is more indicative of skill.
The entire point of the skill rating is to be an unbiased estimator of your tendency to win a match, which is the only goal anyone should strive for in the game. A player's score and k/d do not matter, as these only (biasedly) indicate their efficiency, which is already taken into account in the skill rating, as efficiency is a determinant of the probability of winning a given match. Cutting off the vestigial indicators of performance and relying purely on the w/l statistic eliminates the possibility of attributes irrelevant to winning a match having an effect on the player's skill rating. The w/l statistic incorporates every kill, death, good call, bad call and stroke of luck and even things like the player's tendency to have their internet connection severed during a round. Measures such as score only catch a few of these determinants of victory, and they can be skewed intentionally by the player.
It's statistics. Given a large enough sample of rounds played, the law of large numbers dictates that a player's score should converge towards such a value that when pitted against an opponent (a team) with an equal rating, they win the match 50% of the time, on average. Score has no place here.
There are some problems with the system, nonetheless, but none of them are related to the measure that determines a player's skill rating. One of these problems is the fact that the skill rating of a player's 'opponent' is not that of any single player, but the average of a team composed of multiple individual skill ratings. Is the average alone, however, a suitable estimator of a team's chance of winning a game? There are some limitations within the mechanics of the game themselves that limit the usefulness of individuals of exceptional skill. There are only so many bullets in their magazine and they can only move so fast, no matter their skill rating. This is not really a problem of how to measure a player's skill rating but rather of how to use those ratings to build teams whose skill statistic is actually representative of their innate tendency to win a round.
The entire point of the skill rating is to be an unbiased estimator of your tendency to win a match, which is the only goal anyone should strive for in the game...
Score has no place here.
Touché.
Just thinking of the many, many times I see teams forced even and one team has only one person expected to carry whole team of ppl out-skilled by other team, despite team skills being about even. IMO when these are clearly going to be losses, this person shouldn't be punished as harshly, just wondering if score (gorges can top scoreboards w/ 0 kills) could be incorporated in someway to either lessen the punishment of players who contributed far more in what's ultimately a loss, or to divvy out the pts gained to the winning team's players respective to how much each player contributed via score per minutes in round in a victory. But I definitely see how that's not really a true portrayal of one's skill or ability to win.
Guess it's not needed, and would probably be over-complicated, etc. Comes back to:
It all boils down to creating a measurement based on individual skill ratings that actually, unbiasedly portrays the probability of a team winning a match. Average is the best we've got so far, might be the best we're every going to get. After all, if there is no tendency for team compositions to drift towards having different standard deviations (i.e. one team having a few good players and lots of rookies and the other comprising a balanced set of all-right players), and I see no reason to assume there is, then the average skill rating will, on average, be a good measure of the probability of a result of the match. Incorporating measurements like score or k/d in any way would only serve to spoil the unbiasedness of the rating.
1 / When i decide to play one team i reconfigure the keybindings. So when the FET swap me. "Dude... it's already boring to change the bindings. ho F**k it, F4"...
2 / I never know who i'm playing with. Except one or two that i recognize from time to time. And i like these ones. Not because they can kill as much as me. But when i call for them, they come (what? wait, seriously? -> YES seriously). I prefer 10 times players that play properly doing useful stuff instead of "Pistoleros de la muerte" that just warm up or just able to shoot. I like it as a player on the field as much as a commander. It's so annoying to call ten times for a simple thing.
I don't care about loosing as long as the team did what had to be done at the right moment. Then if it fails, it fails, that's it. Calling hours for nothing is just annoying and render any good action you do useless.
3 / This is not working anyways.
One player that is playing competitive will probably have more lost games than a strong pubber. The "comps" get more 2-2 or 3-1 because the opponent isn't sleeping. So their two respective scores will be clearly different while they are probably of equal force. Or even inverted (big score average, lower score high skill). So the FET has to be the oracle of the day trying to make a choice based on numbers that can be clearly inaccurate and based on the probable behavior of these player (gorge vs fade). No need to say it's impossible.
4/ AFKs.... Daaaaam i hate when that happens. The AFK are often assigned while they got to the toilets. So if a 2000 is having a pee, you're still supposed to win, right ? ... Not sure.
That why i classified this as a lost cause.
The balanced games will come out naturally as soon as players (new or veteran) get the proper behavior and have the proper tools to do it.
* - Teachers / helpers. Ensuring all players get the rules and get the point about tactical stuff and critical structures. I'm not sure everybody does. A "Teacher/trainee pickup systems" that can get in contact players whenever they decide to provide help / teaching. With a good visibility in the main menu like (size 48) "You want help to get better at this game : push button here!!!".
* Language. One of the most durable barrier in this game. It renders sometimes microphones useless.
The menu could have been of great help. We still don't have a menu and sound to say "come over here, i'm on the PG...". We got taunts and stuff but nothing really tactical like "hey make a diversion".
* - Status of players. Nothing can signal you're AFK of something of the kind.
The balanced games will come out naturally as soon as players (new or veteran) get the proper behavior
The society without murders will come naturally as soon as people get the proper behavior.
Why don't you simply say "never"?
There's always an idiot in the landscape. One can say that NS2 doesn't provide the tool for ensuring better behavior. The gaming standards just cannot apply to NS2. It's not COD...
What tool do you have in mind against instinctive stacking behavior? I thought you are against automatic balancing.... Or are you just criticizing what kind of math is used for it?
What tool do you have in mind against instinctive stacking behavior? I thought you are against automatic balancing.... Or are you just criticizing what kind of math is used for it?
Re-read. You'll find.
And there are manys post about this specific topic to "strengthen the community". We had other tools/things, that people barely remember now.
I won't. Thats why I am asking. You oscilate between "force even is good, just polish the math behind it" and "force even is evil", which is it then? You state and restate how people should and should not behave at length. How do you actually make them behave that way. You say something vague like "make tools". Never found anything specific at that point from you in this thread. That's why I am making the sociological joke at your expense above. You can't just tell people to behave some way. It's the opposite. People simply do behave some way. They do stack (from observation). They will stack as long as it is physicaly possible in the game.
@moultano: Maybe there should be a penalty in the team skill calculation, which is higher, the higher the skill spread is within a team. I think we can all agree that a 2000-2000-1000-1000 (total 6000) skill team is better than a 4000-1000-500-500 (total 6000), and may even be better than a 4000-1000-1000-1000 (total 7000). Not sure how to fine tune such a penalty tho, I suck at math
I won't. Thats why I am asking. You oscilate between "force even is good, just polish the math behind it" and "force even is evil", which is it then? You state and restate how people should and should not behave at length. How do you actually make them behave that way. You say something vague like "make tools". Never found anything specific at that point from you in this thread. That's why I am making the sociological joke at your expense above. You can't just tell people to behave some way. It's the opposite. People simply do behave some way. They do stack (from observation). They will stack as long as it is physicaly possible in the game.
I can't point to you a complete bibliography. There is a good load of threads about this very topic on this forum. With good ideas (not mine) in there. Good reading.
I'm not in between. FET doesn't work !
Yesterday there was a FET vote that passed. It ended up Me (+2000) and a pile of 4 rookies on my team. The other team was at an average of 1.5k in which most players where personally at this very level. I stated "we loose" before the very beginning of the game. I was right, to my despair. You can't make the difference with 4 rookies on the team. We asked already that rookies should be equally dispatched. I mean some of them where still setting up the keyboard shortcuts...
How come an anti-stacking algorithm is piling up rookies on the same team; ultimately making a stack on the other side ??? I like jokes... when it's funny.
As @CmdrKeen pointed out; the distribution is wrong.
But also one thing. This algorithm relies on statistics. One must admit they aren't accurate. Because on the difference of "100% pubbers" vs Competitive and also because it is simply not the right parameters to choose.
Being a superstar Pistolero isn't at all what NS is. It's all about TeamWork. Actually there is no stats on TeamWork actions (except Assist in the scoreboard). I understand it can be tricky to implement but would be damn better than the number of won games. And obviously many things are already there.
*Biting RTs (damage on structures - which exists in the engine (but apparently not submitted to hive stats)
*Assists (not a surprise)
*Time being with the teammates (close range) vs Time as solo. And the fun thing is you can be solo and still earn points by killing RTs.
Take a set of this kind of things and you have a clearly better snapshot of what a player is doing and what skill you can give to him/her. Again being good at aiming is ok, but it's not the only thing.
PS: What i call "tools" are implementation in the game for a better life like described before. Devs tried with the match system. A missed shot to say the least.
But also one thing. This algorithm relies on statistics. One must admit they aren't accurate.
Statistics are a thousand times more accurate than personal anecdotes. You're pitting tens of thousands of games worth of data against a one-time personal story of one individual player. You are arguing against mathematics. And when you argue against mathematics, you lose.
Being a superstar Pistolero isn't at all what NS is. It's all about TeamWork. Actually there is no stats on TeamWork actions (except Assist in the scoreboard). I understand it can be tricky to implement but would be damn better than the number of won games. And obviously many things are already there.
*Biting RTs (damage on structures - which exists in the engine (but apparently not submitted to hive stats)
*Assists (not a surprise)
*Time being with the teammates (close range) vs Time as solo. And the fun thing is you can be solo and still earn points by killing RTs.
You are missing the entire point. Every single thing you mentioned is already captured by the player's win-lose ratio, those and every single other determinant of game victory. And it's unbiased, because every determinant affects the win-to-lose ratio by the amount it actually affects the underlying probability of winning and not by the amount some 'expert' says it does (if we went by your route, SOMEONE would have to decide WHICH factors are taken into account and HOW MUCH weight is given to those factors, and there is so much room for error there). By changing the algorithm to forget about the player's victories and concentrate on the measurements you suggest, you're not taking more things into account, you're making the model take less things into account. See numerous posts above (mine included) and the original proposition by moultano in the original thread.
The problem does not lie in the statistic used to calculate a player's skill rating. The problem lies in how to use these individual ratings to a) create equal teams and b) calculate a team rating actually representative of the combined competency of individual players. The problem, again, is NOT in the way the individual player's skill is calculated. Under the assumption that the only thing to strive for in a game round is to win, which is an assumption I don't think anyone disagrees with, the only sensible statistic to use in the measurement of player skill is the win-lose ratio.
These problems are difficult, if not impossible, to address with only discussion. I proposed earlier that the standard deviation of individual skill ratings within a team might have an effect on the outcome that's not captured by the team's average skill rating. I'm not suggesting any actions, however, since all I have is anecdotes, opinions and educated guesses.
What we need is DATA. Data on whether, given equal average skills, the team with LESS standard deviation in skill ratings consistently has a larger probability of winning that match or not. If we find that this is the case, standard deviations should probably be taken into account when creating balanced teams. How? Hell, if I know.
I hope that data is available. The whole skill system relies on data, so I hope this was thought out.
Comments
Isn't that what the mathematical model is supposed to account for though. Since we can't quantify skill directly, we try to estimate it in most circumstances. People who go gorge probably do so often and this would be reflected in their hive score. If not, then they may lose that round.
I think that something like a good player going gorge is something unaccountable for in hive skill.
I knew you meant turts, but typed tufts just in case. In most games I end up playing with turts, which is not many, I see him gorge all the time.
Again though, Isn't that what the mathematical model is supposed to account for though. Since we can't quantify skill directly, we try to estimate it in most circumstances. People who go gorge probably do so often and this would be reflected in their hive score. If not, then they may lose that round.
Interisting video.
He aims always exact the center of the target, prefect tracking. WOW.
Afaik, your score goes up or down partially based on your input on the game. One of the factors of your input is time spent in game. If you were there only 1/2 the round you got half the effect.
What would you suggest to make it better?
The hive skill is changed based on the time you spent on a team. If you F4 or disconnect it will still update your hive skill. As you team is even more likely to loose when you do this, your hive skill will decrease.
Yep, and the beginning of the game counts exponentially more than the end. So if you abandon a game and your team loses, your skill will still go down by a similar amount even if it takes a long time to finish. If you join a game late, you won't be punished or rewarded much for whatever happens in the game.
On gorging, IMHO gorges win games. Most of my wins are from knowing how to place a deep tunnel and when to bile rush the power or whichever forward base the marines are paying the least attention to. Naturally if someone is gorging just to have fun with it rather than because it's what their team needs, they aren't as likely to win as the model predicts. Having your bases constantly under harassment so you can never make forward progress is similar to the feeling of fighting a marine who is so dead on accurate you just can't get close to him, so I think this is working as intended.
Why? I had a match today in which I lost 28 pts in a 2 minute romp. Was a complete stack, skills werent even remotely close. Was just a 28 point tank. In other games where skills are balanced, I will win like 5 games in a row and gain like 20 pts. If skills are "balanced" like 1200 vs 1200, then shouldn't I be gaining more than a couple points at a time if I win those matches?
Let's imagine this situation: Someone leaves 1min into the game, because his team consists only of low skilled players and the enemy team is stacked. However the game drags on for whatever reason. Many of the low skilled players leave the game and get replaced by very high-skilled ones. Though too late and they loose... What happens to the skill of the guy that left early? @moultano.
Honestly I don't know what would be ideal and easy to calculate for that player. You can't just do nothing with his skill points, otherwise people would just leave a game when you they see they're loosing. It should somehow consider playtime and track the teamskill during playtime. And then... ?!?
@moultano I think around 2-4% skill difference between the teams should be ok.
You should. The only variable there is the number of players in the match. You should gain or lose 100 * 8 * 0.5 / num_players points from a match of even skill. Is it possible that the composition of the teams changed a lot over the round?
Can you link to your hive page?
however the 2 min round in question has already been cleared out by other games. And I don't remember it changing a lot, was too short, really. Maybe everybody f4'd on that round? I can see that screwing up the avgs . There was 2 today, skill disparity less than 200 pts, gained 8 on one, 6 on the next. (1st 2 games of march 4th). Also, not sure why it thinks it's march 4th already...
I think I'd like to see player's scores/avg. scores incorporated somehow. I think that shows a player's efficiency in a round more than just w/l. Shows how productive one is on a team, win or lose, which I think is more indicative of skill.
Just a thought. Still think it's far better than previous iterations.
The entire point of the skill rating is to be an unbiased estimator of your tendency to win a match, which is the only goal anyone should strive for in the game. A player's score and k/d do not matter, as these only (biasedly) indicate their efficiency, which is already taken into account in the skill rating, as efficiency is a determinant of the probability of winning a given match. Cutting off the vestigial indicators of performance and relying purely on the w/l statistic eliminates the possibility of attributes irrelevant to winning a match having an effect on the player's skill rating. The w/l statistic incorporates every kill, death, good call, bad call and stroke of luck and even things like the player's tendency to have their internet connection severed during a round. Measures such as score only catch a few of these determinants of victory, and they can be skewed intentionally by the player.
It's statistics. Given a large enough sample of rounds played, the law of large numbers dictates that a player's score should converge towards such a value that when pitted against an opponent (a team) with an equal rating, they win the match 50% of the time, on average. Score has no place here.
There are some problems with the system, nonetheless, but none of them are related to the measure that determines a player's skill rating. One of these problems is the fact that the skill rating of a player's 'opponent' is not that of any single player, but the average of a team composed of multiple individual skill ratings. Is the average alone, however, a suitable estimator of a team's chance of winning a game? There are some limitations within the mechanics of the game themselves that limit the usefulness of individuals of exceptional skill. There are only so many bullets in their magazine and they can only move so fast, no matter their skill rating. This is not really a problem of how to measure a player's skill rating but rather of how to use those ratings to build teams whose skill statistic is actually representative of their innate tendency to win a round.
Just thinking of the many, many times I see teams forced even and one team has only one person expected to carry whole team of ppl out-skilled by other team, despite team skills being about even. IMO when these are clearly going to be losses, this person shouldn't be punished as harshly, just wondering if score (gorges can top scoreboards w/ 0 kills) could be incorporated in someway to either lessen the punishment of players who contributed far more in what's ultimately a loss, or to divvy out the pts gained to the winning team's players respective to how much each player contributed via score per minutes in round in a victory. But I definitely see how that's not really a true portrayal of one's skill or ability to win.
Guess it's not needed, and would probably be over-complicated, etc. Comes back to:
If we had multiple-choice voting, we could replace current FET with
-FET VOTE-
- F1 - Soft balance (try to avoid switching players)
- F2 - Hard balance (switch players to optimise fairness)
- F3 - Don't skill balance
Why i F4 (in fact everybody does)?
1 / When i decide to play one team i reconfigure the keybindings. So when the FET swap me. "Dude... it's already boring to change the bindings. ho F**k it, F4"...
2 / I never know who i'm playing with. Except one or two that i recognize from time to time. And i like these ones. Not because they can kill as much as me. But when i call for them, they come (what? wait, seriously? -> YES seriously). I prefer 10 times players that play properly doing useful stuff instead of "Pistoleros de la muerte" that just warm up or just able to shoot. I like it as a player on the field as much as a commander. It's so annoying to call ten times for a simple thing.
I don't care about loosing as long as the team did what had to be done at the right moment. Then if it fails, it fails, that's it. Calling hours for nothing is just annoying and render any good action you do useless.
3 / This is not working anyways.
One player that is playing competitive will probably have more lost games than a strong pubber. The "comps" get more 2-2 or 3-1 because the opponent isn't sleeping. So their two respective scores will be clearly different while they are probably of equal force. Or even inverted (big score average, lower score high skill). So the FET has to be the oracle of the day trying to make a choice based on numbers that can be clearly inaccurate and based on the probable behavior of these player (gorge vs fade). No need to say it's impossible.
4/ AFKs.... Daaaaam i hate when that happens. The AFK are often assigned while they got to the toilets. So if a 2000 is having a pee, you're still supposed to win, right ? ... Not sure.
That why i classified this as a lost cause.
The balanced games will come out naturally as soon as players (new or veteran) get the proper behavior and have the proper tools to do it.
* - Teachers / helpers. Ensuring all players get the rules and get the point about tactical stuff and critical structures. I'm not sure everybody does. A "Teacher/trainee pickup systems" that can get in contact players whenever they decide to provide help / teaching. With a good visibility in the main menu like (size 48) "You want help to get better at this game : push button here!!!".
* Language. One of the most durable barrier in this game. It renders sometimes microphones useless.
The menu could have been of great help. We still don't have a menu and sound to say "come over here, i'm on the PG...". We got taunts and stuff but nothing really tactical like "hey make a diversion".
* - Status of players. Nothing can signal you're AFK of something of the kind.
Why don't you simply say "never"?
There's always an idiot in the landscape. One can say that NS2 doesn't provide the tool for ensuring better behavior. The gaming standards just cannot apply to NS2. It's not COD...
Re-read. You'll find.
And there are manys post about this specific topic to "strengthen the community". We had other tools/things, that people barely remember now.
I can't point to you a complete bibliography. There is a good load of threads about this very topic on this forum. With good ideas (not mine) in there. Good reading.
I'm not in between. FET doesn't work !
Yesterday there was a FET vote that passed. It ended up Me (+2000) and a pile of 4 rookies on my team. The other team was at an average of 1.5k in which most players where personally at this very level. I stated "we loose" before the very beginning of the game. I was right, to my despair. You can't make the difference with 4 rookies on the team. We asked already that rookies should be equally dispatched. I mean some of them where still setting up the keyboard shortcuts...
How come an anti-stacking algorithm is piling up rookies on the same team; ultimately making a stack on the other side ??? I like jokes... when it's funny.
As @CmdrKeen pointed out; the distribution is wrong.
But also one thing. This algorithm relies on statistics. One must admit they aren't accurate. Because on the difference of "100% pubbers" vs Competitive and also because it is simply not the right parameters to choose.
Being a superstar Pistolero isn't at all what NS is. It's all about TeamWork. Actually there is no stats on TeamWork actions (except Assist in the scoreboard). I understand it can be tricky to implement but would be damn better than the number of won games. And obviously many things are already there.
*Biting RTs (damage on structures - which exists in the engine (but apparently not submitted to hive stats)
*Assists (not a surprise)
*Time being with the teammates (close range) vs Time as solo. And the fun thing is you can be solo and still earn points by killing RTs.
Take a set of this kind of things and you have a clearly better snapshot of what a player is doing and what skill you can give to him/her. Again being good at aiming is ok, but it's not the only thing.
PS: What i call "tools" are implementation in the game for a better life like described before. Devs tried with the match system. A missed shot to say the least.
Statistics are a thousand times more accurate than personal anecdotes. You're pitting tens of thousands of games worth of data against a one-time personal story of one individual player. You are arguing against mathematics. And when you argue against mathematics, you lose.
You are missing the entire point. Every single thing you mentioned is already captured by the player's win-lose ratio, those and every single other determinant of game victory. And it's unbiased, because every determinant affects the win-to-lose ratio by the amount it actually affects the underlying probability of winning and not by the amount some 'expert' says it does (if we went by your route, SOMEONE would have to decide WHICH factors are taken into account and HOW MUCH weight is given to those factors, and there is so much room for error there). By changing the algorithm to forget about the player's victories and concentrate on the measurements you suggest, you're not taking more things into account, you're making the model take less things into account. See numerous posts above (mine included) and the original proposition by moultano in the original thread.
The problem does not lie in the statistic used to calculate a player's skill rating. The problem lies in how to use these individual ratings to a) create equal teams and b) calculate a team rating actually representative of the combined competency of individual players. The problem, again, is NOT in the way the individual player's skill is calculated. Under the assumption that the only thing to strive for in a game round is to win, which is an assumption I don't think anyone disagrees with, the only sensible statistic to use in the measurement of player skill is the win-lose ratio.
These problems are difficult, if not impossible, to address with only discussion. I proposed earlier that the standard deviation of individual skill ratings within a team might have an effect on the outcome that's not captured by the team's average skill rating. I'm not suggesting any actions, however, since all I have is anecdotes, opinions and educated guesses.
What we need is DATA. Data on whether, given equal average skills, the team with LESS standard deviation in skill ratings consistently has a larger probability of winning that match or not. If we find that this is the case, standard deviations should probably be taken into account when creating balanced teams. How? Hell, if I know.
I hope that data is available. The whole skill system relies on data, so I hope this was thought out.