@SupaDupaNoodle - There is a search function on this forum, please use that rather than necro threads.
The short answer to your question is; your K/D is not a factor in your skill assessment. You can have a 100/0 K/D, but if your team loses the game it doesn't matter.
I didn't necro any thread. When I was thinking about this problem, I checked the board today and someone else had already made a comment recently, so I added my own question.
I have been on this forum far longer than you. I know that there is a search function. If what you linked is the desired result of the "search option," it is wholly inadequate. Some line of code in the thread, and then some algorithmic gibberish on a linked googledoc doesn't explain how the skill system works. Someone should write an actual concise, English explanation.
Relating to my question specifically, whether the other team wins or loses, my score rating at the end of the round is not congruent AT ALL with the others who performed equal to or worse than me. Quite frankly, the system doesn't work. Someone needs to fix it.
@santaclaws, he really did not necro it. That was several posts back and yesterday. This thread is still really active and has been showing some decent discussion recently. I think this discussion is still valid, recent, and not a necro.
Recently, I've seen it balancing like this:
-Alien : 2000 + 3 rookie and 2~3 between 500 1500
-Marine all between 500 1500
Marine team have 3 more players than the alien team.... whut ?! (no F4 involved & second best laugh about it)
If one have a high score doesn't mean he's got more units on the field. or more hands...
It's a lost cause for me. Better work on a "buddy" system / interface to get people trained with potent players. So NS2 can put up with some real numbers.
The uneven amount of players might be related to a current bug of that vote which is related to the hive timing-out.
Recently, I've seen it balancing like this:
-Alien : 2000 + 3 rookie and 2~3 between 500 1500
-Marine all between 500 1500
Marine team have 3 more players than the alien team.... whut ?! (no F4 involved & second best laugh about it)
If one have a high score doesn't mean he's got more units on the field. or more hands...
It's a lost cause for me. Better work on a "buddy" system / interface to get people trained with potent players. So NS2 can put up with some real numbers.
The uneven amount of players might be related to a current bug of that vote which is related to the hive timing-out.
That might explain the force even vote causing one team to have twice the skill rating compared to the other. It was funny til I was shuffled on the bad team. Made for really poor games (and heaps of in joke about the even vote being broken (again)). Random ready room had better success.
Social Experiments, compliments of Nordic, the minister of silly.
It started out as "Skill system is bad" because I found something broken about it. The discussion cascaded into 10 pages of mostly negativity. Somewhere along the line I have begun to think the hive skill system does a decent job at balance so I changed the title to "Skill System is Good" to see if the discussion would get more positive. It did not, but the quality of the comments improved. I highly doubt this was related to the title change so I don't think the title has a real effect. To find out I did something ridiculous, hence the reference to 1984 in the title. That seemed to end the discussion but it was noticed (Thank you Squishpoke). So now it is "To Skill or Not to Skill" because it is less ridiculous but still a bit odd. I have my metaphorical popcorn ready.
I have played outside my usual servers this week, where I find that the skill system is working pretty well. When trying to vote for balance there are those few people claiming it does absolutely nothing. Are these people stuck in old ways? Or are they right?
If one of the player has a big number compared to the rest, it fails. It usually ends up having a stack of rookies and one uber player one the same side. When everybody has more or less the same numbers, it works better. But in this case a standart random is just the same.
The only case where it is useful is when you have, say, one even group score of 500 and one of 1000. Of course all 1000 should be evenly distributed. Then the 1000 of both teams should play the same kind of roles. Aggressve vs passive.
i dont really think it works at all. if you always switch teams to the strongest team you will eventually hit over 2000 skill regardless of how useful u actually are. If you constantly fight the stack your skill level wil suck. Imo assists and kill death ratio are more telling of skill than win loss ratio
i dont really think it works at all. if you always switch teams to the strongest team you will eventually hit over 2000 skill regardless of how useful u actually are. If you constantly fight the stack your skill level wil suck. Imo assists and kill death ratio are more telling of skill than win loss ratio
No. If you always play with the strongest team. It's not said that you win all the time. Sometimes the stronger team looses and when that happens, you loose way more skill points than you got over large number of stacked games.
KDR based skill system we basically had before and I wouldn't wan't to go back to it. Atleast this system works for regulars.
i dont really think it works at all. if you always switch teams to the strongest team you will eventually hit over 2000 skill regardless of how useful u actually are. If you constantly fight the stack your skill level wil suck. Imo assists and kill death ratio are more telling of skill than win loss ratio
No. If you always play with the strongest team. It's not said that you win all the time. Sometimes the stronger team looses and when that happens, you loose way more skill points than you got over large number of stacked games.
KDR based skill system we basically had before and I wouldn't wan't to go back to it. Atleast this system works for regulars.
i didnt say you win all the time and i didnt say strongest team is based off of skill level. I said if you play with the strongest team you will eventually have 2000 skill. strongest usually implies who is in each team and if you play regularly you will know who is good and who isnt. but when its a skill rating of 700 vs 1600 you know who is likely to win. But if you know the players and are familar with them you have a far better idea who will win compared to the skill rating
the skill system is only helpful with extremes of skill difference. But when a match is even skill wise and yet 1 team gets wiped without even securing a single harvester within 10 mins you know something is wrong
It still will take the stacker very long to get to 2000 and then will loose it very rapidly once he plays outside of his safety zone. I'm not telling the current system is flawless, but if we had a KDR based one, the guy would just go farm some rookies or only play Marines. Gorges and Commanders become "useless" too.
the skill system is only helpful with extremes of skill difference. But when a match is even skill wise and yet 1 team gets wiped without even securing a single harvester within 10 mins you know something is wrong
Teamplay needs to work too and if they can't secure a harvester they just trolled or something.
Social experiments are testing. It appears I need a simplistic emotive title to draw people into the thread. Maybe a positive connotation of that title does produce better discussion. I would hypothesize it is because they are arguing against why it is not good instead of a simple agreement of why it is bad.
Having just played a few games using force even, it's still as useful as a condom machine in the Vatican.
FE is still better than no FE, and better than random.
Remember, it can't guarantee balanced games -- nothing can, in fact. It can't predict strategies of a comm, or a gorge getting that ninja tunnel at the right time, or the rambo soloing a hive etc. It's only ever an indicator, but it's indicator (by now) of thousands of games and how when a player is on a team, the frequency of that team winning. Everyone is ranked that way, such that when teams are allocated it's allocating people in way to create as best is possible an equal chance of winning -- but even in a perfectly equal game, there's still a winner and still a loser. And that's not assuming ninja gorges, uber comm strats, sudden breakouts of teamwork on a pub etc.
Frankly, I think it's doing an awesome job. The way to measure it is the frequency of bad, horrible games. And hands down, since the skill system, these have dropped like a stone. Many more games are fun and balanced since the @moultano skill system changes to FE.
Oh, and don't complain if you play games where people F4 after being allocated to a team in FE. Completely defeats the purpose if people switch or exit to RR. Like a child whining that he lost his ice-cream after he turned the cone up-side down. If you want better games, stick to your FE allocation. No, you won't always win. You also won't always lose.
Does the skill system take into account the standard deviation of skill ratings within a team? One or two great players cannot carry a team of rookies against a team of steadily distributed all-right players, even if the average skill ratings were the same. This is especially prevalent when the great players are on the marine side; even though they could absolutely decimate any opposition that comes their way, it can easily be countered by attacking positions where those players aren't and slowly stealing the macro game from under their feet. As an alien, this isn't that much of a problem since a good alien player can have a much wider map presence than a marine and more easily take care of their teammates' mistakes.
I do F4, sorry to say that but FE isn't working great. Im usually joining first and most of times i get 5 complete rookies to carry. When the other teams have middle skilled players. I don't think it should work like that.
I do F4, sorry to say that but FE isn't working great. Im usually joining first and most of times i get 5 complete rookies to carry. When the other teams have middle skilled players. I don't think it should work like that.
Probably because rookies start at 1000 skill points, as far as I know. (Which is really dumb)
I do F4, sorry to say that but FE isn't working great. Im usually joining first and most of times i get 5 complete rookies to carry. When the other teams have middle skilled players. I don't think it should work like that.
Probably because rookies start at 1000 skill points, as far as I know. (Which is really dumb)
Well that was true when the system was introduced (idk why that was done) but today rookies start with 0. Which lead to this issue at pubs of having really large skill gaps between certain players making it kinda hard to balance teams.
I do F4, sorry to say that but FE isn't working great. Im usually joining first and most of times i get 5 complete rookies to carry. When the other teams have middle skilled players. I don't think it should work like that.
Protip RR if you see a FET vote passing. Having players already on the teams seems to kind of mess with how it tries to distribute players, especially if there's a couple of the highest skilled players already on the same team.
My issue at the moment is that since FET was changed to try and avoid switching players, if players want to stomp (or just dont care about having an even game, or stubbornly want to only play one team with their friends, for example), they just have to join before the vote passes. When it re-assigned people regardless of what their choice was, this was less of an issue outside of obtuse post-FET F4ing.
@Therius makes a good point; if it were possible to try and distribute the skills evenly based on their deviation from the average score, it should give much better results.
I do F4, sorry to say that but FE isn't working great. Im usually joining first and most of times i get 5 complete rookies to carry. When the other teams have middle skilled players. I don't think it should work like that.
Oh Carnage, you sure the rookies were not carrying you haha j/k
dePARAJoin Date: 2011-04-29Member: 96321Members, Squad Five Blue
I think the system is working ok and its the best we have atm.
It has some issues for sure.
Example:
You have 2 Players with over 2000 skill points each team and the other players are average pub players around 1000 and one is forced to go com.
While the field player has easy going, the com is cant do much. This is more a problem on alienside cause you cant active support your players.
The "skill-vote" fails most of the time when people manually switching or leaving and weaker players fill the open slot.
So the human factor destroys the system.
And this going to happen with every try to implement another system.
Does the skill system take into account the standard deviation of skill ratings within a team? One or two great players cannot carry a team of rookies against a team of steadily distributed all-right players, even if the average skill ratings were the same. This is especially prevalent when the great players are on the marine side; even though they could absolutely decimate any opposition that comes their way, it can easily be countered by attacking positions where those players aren't and slowly stealing the macro game from under their feet. As an alien, this isn't that much of a problem since a good alien player can have a much wider map presence than a marine and more easily take care of their teammates' mistakes.
lolz, thats how we (the mediocre) team beat the other team with 1-2 Hawkeyes. Last week on the Thirsty Onos I think, the team actively reported where the dead shots are (xxx is in hub etc...), and we avoided that area except for a few that went there to keep the top players busy, while the rest of team wrecked the map where the top players aren't. By the time we tech to a few oni(iiiiis) no matter how well you shot, it didn't matter.
I do F4, sorry to say that but FE isn't working great. Im usually joining first and most of times i get 5 complete rookies to carry. When the other teams have middle skilled players. I don't think it should work like that.
Probably because rookies start at 1000 skill points, as far as I know. (Which is really dumb)
Well that was true when the system was introduced (idk why that was done) but today rookies start with 0. Which lead to this issue at pubs of having really large skill gaps between certain players making it kinda hard to balance teams.
Given the asymptotic nature of the skill rating (ie. you tend toward a particular level and approach it closer over time). it might be best to start rookies on say 500-700 (lower than normal, but not 0). I am about 950 after 800 hrs, and play pub only. Given that is a n=1 it doesn't give you a good idea about where rookies should sit. May be you can do an average of the current active player base, and subtract 200 (this is arbitrary) from the active players atm as the rookie start.
I do F4, sorry to say that but FE isn't working great. Im usually joining first and most of times i get 5 complete rookies to carry. When the other teams have middle skilled players. I don't think it should work like that.
What's your skill value? Can you link to your hive page?
Edit: Assuming this is you: http://hive.naturalselection2.com/profile/439402 You are still winning 3/4ths of your games, so whatever the skill system is attempting to do to handicap you, it clearly isn't enough
Does the skill system take into account the standard deviation of skill ratings within a team? One or two great players cannot carry a team of rookies against a team of steadily distributed all-right players, even if the average skill ratings were the same. This is especially prevalent when the great players are on the marine side; even though they could absolutely decimate any opposition that comes their way, it can easily be countered by attacking positions where those players aren't and slowly stealing the macro game from under their feet. As an alien, this isn't that much of a problem since a good alien player can have a much wider map presence than a marine and more easily take care of their teammates' mistakes.
This seems to be a pretty common complaint, but I'm not sure what the right fix is. The root problem seems to be that there's one player on the server who is just way way better than everyone else. What should balanced teams with that player look like? If there isn't a counterbalancing player who is equivalently skilled on the other team, then it's really hard to compose teams that will make for a satisfying game. Expecting the amazing player to carry a poor team seems reasonable to me, but maybe there's an alternative?
Lets say your server is as follows.
skill 3000 : 1 player
skill 1000: 8 players
skill 0 : 5 players
Comments
I didn't necro any thread. When I was thinking about this problem, I checked the board today and someone else had already made a comment recently, so I added my own question.
I have been on this forum far longer than you. I know that there is a search function. If what you linked is the desired result of the "search option," it is wholly inadequate. Some line of code in the thread, and then some algorithmic gibberish on a linked googledoc doesn't explain how the skill system works. Someone should write an actual concise, English explanation.
Relating to my question specifically, whether the other team wins or loses, my score rating at the end of the round is not congruent AT ALL with the others who performed equal to or worse than me. Quite frankly, the system doesn't work. Someone needs to fix it.
The uneven amount of players might be related to a current bug of that vote which is related to the hive timing-out.
Since the discussion has gone on so long I keep changing the title to see if it effects the discussion.
That might explain the force even vote causing one team to have twice the skill rating compared to the other. It was funny til I was shuffled on the bad team. Made for really poor games (and heaps of in joke about the even vote being broken (again)). Random ready room had better success.
It started out as "Skill system is bad" because I found something broken about it. The discussion cascaded into 10 pages of mostly negativity. Somewhere along the line I have begun to think the hive skill system does a decent job at balance so I changed the title to "Skill System is Good" to see if the discussion would get more positive. It did not, but the quality of the comments improved. I highly doubt this was related to the title change so I don't think the title has a real effect. To find out I did something ridiculous, hence the reference to 1984 in the title. That seemed to end the discussion but it was noticed (Thank you Squishpoke). So now it is "To Skill or Not to Skill" because it is less ridiculous but still a bit odd. I have my metaphorical popcorn ready.
The only case where it is useful is when you have, say, one even group score of 500 and one of 1000. Of course all 1000 should be evenly distributed. Then the 1000 of both teams should play the same kind of roles. Aggressve vs passive.
KDR based skill system we basically had before and I wouldn't wan't to go back to it. Atleast this system works for regulars.
i didnt say you win all the time and i didnt say strongest team is based off of skill level. I said if you play with the strongest team you will eventually have 2000 skill. strongest usually implies who is in each team and if you play regularly you will know who is good and who isnt. but when its a skill rating of 700 vs 1600 you know who is likely to win. But if you know the players and are familar with them you have a far better idea who will win compared to the skill rating
the skill system is only helpful with extremes of skill difference. But when a match is even skill wise and yet 1 team gets wiped without even securing a single harvester within 10 mins you know something is wrong
Teamplay needs to work too and if they can't secure a harvester they just trolled or something.
Remember, it can't guarantee balanced games -- nothing can, in fact. It can't predict strategies of a comm, or a gorge getting that ninja tunnel at the right time, or the rambo soloing a hive etc. It's only ever an indicator, but it's indicator (by now) of thousands of games and how when a player is on a team, the frequency of that team winning. Everyone is ranked that way, such that when teams are allocated it's allocating people in way to create as best is possible an equal chance of winning -- but even in a perfectly equal game, there's still a winner and still a loser. And that's not assuming ninja gorges, uber comm strats, sudden breakouts of teamwork on a pub etc.
Frankly, I think it's doing an awesome job. The way to measure it is the frequency of bad, horrible games. And hands down, since the skill system, these have dropped like a stone. Many more games are fun and balanced since the @moultano skill system changes to FE.
Oh, and don't complain if you play games where people F4 after being allocated to a team in FE. Completely defeats the purpose if people switch or exit to RR. Like a child whining that he lost his ice-cream after he turned the cone up-side down. If you want better games, stick to your FE allocation. No, you won't always win. You also won't always lose.
Well that was true when the system was introduced (idk why that was done) but today rookies start with 0. Which lead to this issue at pubs of having really large skill gaps between certain players making it kinda hard to balance teams.
My issue at the moment is that since FET was changed to try and avoid switching players, if players want to stomp (or just dont care about having an even game, or stubbornly want to only play one team with their friends, for example), they just have to join before the vote passes. When it re-assigned people regardless of what their choice was, this was less of an issue outside of obtuse post-FET F4ing.
@Therius makes a good point; if it were possible to try and distribute the skills evenly based on their deviation from the average score, it should give much better results.
Oh Carnage, you sure the rookies were not carrying you haha j/k
It has some issues for sure.
Example:
You have 2 Players with over 2000 skill points each team and the other players are average pub players around 1000 and one is forced to go com.
While the field player has easy going, the com is cant do much. This is more a problem on alienside cause you cant active support your players.
The "skill-vote" fails most of the time when people manually switching or leaving and weaker players fill the open slot.
So the human factor destroys the system.
And this going to happen with every try to implement another system.
lolz, thats how we (the mediocre) team beat the other team with 1-2 Hawkeyes. Last week on the Thirsty Onos I think, the team actively reported where the dead shots are (xxx is in hub etc...), and we avoided that area except for a few that went there to keep the top players busy, while the rest of team wrecked the map where the top players aren't. By the time we tech to a few oni(iiiiis) no matter how well you shot, it didn't matter.
Given the asymptotic nature of the skill rating (ie. you tend toward a particular level and approach it closer over time). it might be best to start rookies on say 500-700 (lower than normal, but not 0). I am about 950 after 800 hrs, and play pub only. Given that is a n=1 it doesn't give you a good idea about where rookies should sit. May be you can do an average of the current active player base, and subtract 200 (this is arbitrary) from the active players atm as the rookie start.
What's your skill value? Can you link to your hive page?
Edit: Assuming this is you: http://hive.naturalselection2.com/profile/439402 You are still winning 3/4ths of your games, so whatever the skill system is attempting to do to handicap you, it clearly isn't enough
This seems to be a pretty common complaint, but I'm not sure what the right fix is. The root problem seems to be that there's one player on the server who is just way way better than everyone else. What should balanced teams with that player look like? If there isn't a counterbalancing player who is equivalently skilled on the other team, then it's really hard to compose teams that will make for a satisfying game. Expecting the amazing player to carry a poor team seems reasonable to me, but maybe there's an alternative?
Lets say your server is as follows.
skill 3000 : 1 player
skill 1000: 8 players
skill 0 : 5 players
How would you allocate these players?