ITT: People totally polarised one way or the other, and somehow can't see how they are all correct.
You have people who understand that that Win/Loss is the only reliable metric, and those people are correct.
You have people who are frustrated that their "skill level" can go down if they lose a game while they individually did really well. Those people are correct to be annoyed because its not a skill level at all, its a number that contains your chances of being on the winning team (in the long run). It purposely integrates unfair things like being on the team full of rookies, because on average for every time you are put on the weak team, you'll be put on a strong team so it should cancel out.
Some people see errors in the system and think its because the system is totally unsuitable, when most of it is known problems that will be fixed. For example the strongest one is that Marines and Aliens need different skill ratings. I proved this with the example of shellmo, who at 75% play time on marines had a skill level of 300. She then played nothing but Aliens for several hours and now its up at 1400. Which brings me to the final point.
Skill numbers aren't easy to understand for the typical person, is a 2000 skill level player twice as good as a 1000 skill player? How many more times does one win over the other. In shellmo's example it was entirely down to the fact that Marines only win 46% of games. Her skill level changed from 300 to 1400. Thats how knife edge it can seem. Improving your W/L from 0.99 to 1.01 makes your skill level go up by nearly 5 times.
1. Add seperate ratings for Marines and Aliens.
2. Hide the number itself and just show ranking, or rename it so people don't get confused.
3. Potentially stop using the average of (what seems to be) a logarithmic value to show total team strength.
Then we can all get back to the threads about how the game is dying instead!
w/l isnt enough because those who play even teams are likely to lose 50% of the games they play and yet completely destroy balance in a typical public game.
I don't understand. If you have 50% chance to win or lose, is that not, by definition, a balanced game?
yea when the teams are evenly matched regardless of the skill system but because you actively balance them with people you know.. but in most pub games they are subjected to players with bloated skill levels and greens with 0 skill combined with high level vets and comp players with modest skill levels who dont usually play on those servers and this basically throws the whole skill system out the window and since the majority of people dont know each other well either it just becomes a mess
the system currently is very exploitable. If you constantly fight the stack on a public server then overall your skill level will suck regardless of how good you are. If you constantly stack it will be the opposite and there are a lot of people who do both.
i have seen a team with 700skill completely demolish a team with 1500 and when i say demolish i mean the other team struggled to hold more than 2 harvesters for the 8 min duration the match lasted.
the system currently is very exploitable. If you constantly fight the stack on a public server then overall your skill level will suck regardless of how good you are. If you constantly stack it will be the opposite and there are a lot of people who do both..
This is not true, and if you cannot see why, then you lack understanding of what the skill system does. It has been explained dozens of times in other threads.
w/l isnt enough because those who play even teams are likely to lose 50% of the games they play and yet completely destroy balance in a typical public game.
I don't understand. If you have 50% chance to win or lose, is that not, by definition, a balanced game?
Balance is not necessarily that simple since you must look at individual components rather than a simple win/lose ratio. But since we are talking about the skill system, I won't continue on that.
Look, the game is balanced for 2-round matches, simply because it's aysmmetrical.
All balance then lies in maps. A 100% win-rate alien map is still balanced in 2-round matches because both teams win once, 50/50
edit: All other balancing for 1-round matches is what makes this game boring and filled with band-aids
Being asymmetrical is not an excuse to be imbalanced for 1 round matches. A 2 round map with 2 alien wins can either be perfectly balanced with one team being better or the map/game being alien sided. You can't just put a blanket statement on all matches with 1-1 round outcomes being perfectly balanced, because usually they aren't.
Being balanced 50/50 means that each team had an equal chance, not taking skill into account, to win one round regardless of what faction they are on.
Being balanced 50/50 means that each team had an equal chance, not taking skill into account, to win one round regardless of what faction they are on.
This blanket statement can only apply to balance around a single map. Which is why people love summit so much
This doesn't make sense. My statement that you quoted was not a generalised statement, it was a definition of what 50/50 balance in this game is. The reason people like summit so much is not because its balanced for 2 rounds, but because its balanced for 1 round. When you balance so that the game is 50/50 based on 1 round, it is balanced 50/50 for 2 rounds inherently. When you balance for 2 rounds, you can still be unbalanced.
If you want to call it a skill system, you need to actually use measures of skill to rate each player. This should include accuracy, kd ratio, structure damage (as a ratio of actual structures killed to prevent crag farming), enemy player damage, and time spent constructing/welding/healing. Lo and behold, all these things are measured already. Too bad their actual purpose is ignored. The weighting for each of these components needs to be carefully decided. Alien and marine skills should also be separated.
At the moment, a well-timed winning gorge rush by a low average skill team (skewed down by a few rookies) against a high average skill team leads to a load of proficient players losing a ton of skill points and undeserving players getting a boost of points. This is of course fun and that last second bile rush should always be part of the game and should be allowed to lead to victory against an arrogant commander who doesn't drop a second CC, but such an eventuality should not determine one's "skill" at the game. A bile rush is in fact a very low-skill tactic and a victory by this method should not be rewarded with a boost of skill points.
A proper measurement of one's skill can then lead to balanced teams being created. The current system also rewards winning team joiners by measuring, as it does, nothing more than likelihood of being on the winning team. Likelihood of being on the winning team in NS2 clearly has nothing to do with skill, especially on the public servers where most of the matches are played.
Something else that needs to be done is regular resetting of everyone's skill rating so that those who have longer game time aren't disadvantaged, so that improving yourself through hard work and practice are rewarded, and so that new players (if they will exist) can have a better chance at being correctly allocated to a team. At the moment, the "skill system" just acts as a log of whether or not you were on the winning team for your history of playing NS2, which is something for the most part out of your hands in such a team-based, asymmetrical game.
I really do wonder what was going through the "skill" system designers' heads when they activated such a flawed and useless system.
Time for a complete overhaul of the skill scoring system.
At the moment, a well-timed winning gorge rush by a low average skill team (skewed down by a few rookies) against a high average skill team leads to a load of proficient players losing a ton of skill points and undeserving players getting a boost of points. This is of course fun and that last second bile rush should always be part of the game and should be allowed to lead to victory against an arrogant commander who doesn't drop a second CC, but such an eventuality should not determine one's "skill" at the game. A bile rush is in fact a very low-skill tactic and a victory by this method should not be rewarded with a boost of skill points.
Comments
You have people who understand that that Win/Loss is the only reliable metric, and those people are correct.
You have people who are frustrated that their "skill level" can go down if they lose a game while they individually did really well. Those people are correct to be annoyed because its not a skill level at all, its a number that contains your chances of being on the winning team (in the long run). It purposely integrates unfair things like being on the team full of rookies, because on average for every time you are put on the weak team, you'll be put on a strong team so it should cancel out.
Some people see errors in the system and think its because the system is totally unsuitable, when most of it is known problems that will be fixed. For example the strongest one is that Marines and Aliens need different skill ratings. I proved this with the example of shellmo, who at 75% play time on marines had a skill level of 300. She then played nothing but Aliens for several hours and now its up at 1400. Which brings me to the final point.
Skill numbers aren't easy to understand for the typical person, is a 2000 skill level player twice as good as a 1000 skill player? How many more times does one win over the other. In shellmo's example it was entirely down to the fact that Marines only win 46% of games. Her skill level changed from 300 to 1400. Thats how knife edge it can seem. Improving your W/L from 0.99 to 1.01 makes your skill level go up by nearly 5 times.
1. Add seperate ratings for Marines and Aliens.
2. Hide the number itself and just show ranking, or rename it so people don't get confused.
3. Potentially stop using the average of (what seems to be) a logarithmic value to show total team strength.
Then we can all get back to the threads about how the game is dying instead!
yea when the teams are evenly matched regardless of the skill system but because you actively balance them with people you know.. but in most pub games they are subjected to players with bloated skill levels and greens with 0 skill combined with high level vets and comp players with modest skill levels who dont usually play on those servers and this basically throws the whole skill system out the window and since the majority of people dont know each other well either it just becomes a mess
the system currently is very exploitable. If you constantly fight the stack on a public server then overall your skill level will suck regardless of how good you are. If you constantly stack it will be the opposite and there are a lot of people who do both.
i have seen a team with 700skill completely demolish a team with 1500 and when i say demolish i mean the other team struggled to hold more than 2 harvesters for the 8 min duration the match lasted.
This is not true, and if you cannot see why, then you lack understanding of what the skill system does. It has been explained dozens of times in other threads.
Balance is not necessarily that simple since you must look at individual components rather than a simple win/lose ratio. But since we are talking about the skill system, I won't continue on that.
All balance then lies in maps. A 100% win-rate alien map is still balanced in 2-round matches because both teams win once, 50/50
edit: All other balancing for 1-round matches is what makes this game boring and filled with band-aids
Being asymmetrical is not an excuse to be imbalanced for 1 round matches. A 2 round map with 2 alien wins can either be perfectly balanced with one team being better or the map/game being alien sided. You can't just put a blanket statement on all matches with 1-1 round outcomes being perfectly balanced, because usually they aren't.
Being balanced 50/50 means that each team had an equal chance, not taking skill into account, to win one round regardless of what faction they are on.
This blanket statement can only apply to balance around a single map. Which is why people love summit so much
This doesn't make sense. My statement that you quoted was not a generalised statement, it was a definition of what 50/50 balance in this game is. The reason people like summit so much is not because its balanced for 2 rounds, but because its balanced for 1 round. When you balance so that the game is 50/50 based on 1 round, it is balanced 50/50 for 2 rounds inherently. When you balance for 2 rounds, you can still be unbalanced.
At the moment, a well-timed winning gorge rush by a low average skill team (skewed down by a few rookies) against a high average skill team leads to a load of proficient players losing a ton of skill points and undeserving players getting a boost of points. This is of course fun and that last second bile rush should always be part of the game and should be allowed to lead to victory against an arrogant commander who doesn't drop a second CC, but such an eventuality should not determine one's "skill" at the game. A bile rush is in fact a very low-skill tactic and a victory by this method should not be rewarded with a boost of skill points.
A proper measurement of one's skill can then lead to balanced teams being created. The current system also rewards winning team joiners by measuring, as it does, nothing more than likelihood of being on the winning team. Likelihood of being on the winning team in NS2 clearly has nothing to do with skill, especially on the public servers where most of the matches are played.
Something else that needs to be done is regular resetting of everyone's skill rating so that those who have longer game time aren't disadvantaged, so that improving yourself through hard work and practice are rewarded, and so that new players (if they will exist) can have a better chance at being correctly allocated to a team. At the moment, the "skill system" just acts as a log of whether or not you were on the winning team for your history of playing NS2, which is something for the most part out of your hands in such a team-based, asymmetrical game.
I really do wonder what was going through the "skill" system designers' heads when they activated such a flawed and useless system.
Time for a complete overhaul of the skill scoring system.
Googled that, can't find anything in a wiki. Sounds familiar though.
https://en.wikipedia.org/wiki/Elo_rating_system
https://en.wikipedia.org/wiki/Logistic_regression