Bad ratings for imbalanced games

2

Comments

  • The_Welsh_WizardThe_Welsh_Wizard Join Date: 2013-09-10 Member: 188101Members, Reinforced - Supporter
    I appreciate all the effort you Put into that Nordic, but I can't agree. The System is too simple and therefore does a Bad Job at measuring the true skill of a player. It is the reason 50% of games are a stomp after a forced shuffle. And when you compare Hive points of players you know it Shows even more that it is very inaccurate.

    And also it would be nice if you could go into some Kind of Group with a buddy. I am tired of switching servers because you always get shuffled into different teams.
  • VetinariVetinari Join Date: 2013-07-23 Member: 186325Members, Squad Five Blue, Reinforced - Shadow, WC 2013 - Silver
    edited June 2016
    I appreciate all the effort you Put into that Nordic, but I can't agree. The System is too simple and therefore does a Bad Job at measuring the true skill of a player. It is the reason 50% of games are a stomp after a forced shuffle. And when you compare Hive points of players you know it Shows even more that it is very inaccurate.

    But... Nordic literally just demonstrated that the system is alright. With numbers. All you ever come up with are subjective statements without any backup. I find it hard to take that seriously.
    And also it would be nice if you could go into some Kind of Group with a buddy. I am tired of switching servers because you always get shuffled into different teams.

    This, on the other hand, is true. Although personally I think this should take lower priority than fixing the other issues hive has already.

    edit: quote fix
  • RevanCoranaRevanCorana Join Date: 2015-08-14 Member: 207125Members
    edited June 2016
    The hive system is mathematically correct but has a huge delay which leads to imprecise shuffles.

    This can be fixed easily by accounting for individual players' hive playtime acting as a multiplier of hive skill.

    Simple example: A veteran with 1500elo would be rated 1500x2=3000 effective rating while a rookie with 1500elo would be rated 1500x0.5=750 effective rating.

    Done
  • Me9aMe9a Join Date: 2008-03-27 Member: 63981Members
    Player Rentation:

    Migalksi has a important Information about Player Rentation in his post:
    http://forums.unknownworlds.com/discussion/comment/2296205/#Comment_2296205


    That he cant Team up with his Mates
    and all other Modern games have a Party/Team System.
    (thats why he (and probably a lot of of other People) prefer to play other Games)
    THIS needs to be added to NS2.
    (maximum party size should probably be 4-6 Players)





    Suggestion:

    There also should be 1 Hiveskill for each Race
    1 for Aliens
    1 for Marines

    As soon as Marines and Aliens have Commanders just Reshuffle so the Game has a Higher chance in being Hiveskill-equal
    (without takeing the Commanders Hiveskill into account because a high skill Player as Commander imbalances the Game).
    (BUT dont separate the "Partys/Teams" while doing that.)


    Maybe add an Hiveskillranking for Commanders too as a Badge with number in the Scoreboard.
    (tbh idk but commanders have to be somehow rewarded for commanding).


    Optional:
    If these Teams then are very badly balanced after all shuffles suggest a split up with a popup message to the Party/Group.

    or get started on a matchmaking #makehive2happen.
  • NordicNordic Long term camping in Kodiak Join Date: 2012-05-13 Member: 151995Members, NS2 Playtester, NS2 Map Tester, Reinforced - Supporter, Reinforced - Silver, Reinforced - Shadow
    edited June 2016
    Now the conversation goes in a circle.
    Hive is the reason 50% of games are a stomp after a forced shuffle.
    Nope.
    Nordic wrote: »
    The biggest reason that makes balance hard to get in ns2 is our small playerbase combined with a high skill ceiling. There are not enough players for low skill and high skill players to have servers of their own. We all have to play together. When you have the highest skilled players playing with the lowest skilled players, it does not make for a good experience. The only way to fix this is to make the game more popular, which is a chicken and egg problem.
    Nordic wrote: »
    And also it would be nice if you could go into some Kind of Group with a buddy. I am tired of switching servers because you always get shuffled into different teams.
    It would be nice if we could play with friends, but again that is a problem of the small playerbase. If the game had a lot players, there would be more people at every skill bracket, and you could play at your skill bracket with your friends in some sort of hypothetical matchmaking.
    The hive system is mathematically correct but has a huge delay which leads to imprecise shuffles.
    Nope. Hive is not perfect as I have described in this thread. It is not so much that hive is imprecise, but that we have a small playerbase that ruins shuffles. No one can find near skilled games. There is practically no such thing as near skilled games in ns2.
    Nordic wrote: »
    Hive is not perfect, as I mentioned. It can be a lot better. For example:
    • Hive has some rookies start at 0 hive skill and others start at 1000 hive skill. This is known, and has a proposed fix.
    • Hive takes too long to find a reasonable estimate of a players skill. There is also a proposed fix.
    • Hive has a single skill value for both marines and aliens, even though a player can be wildly more effective with one faction over the other. The proposed fix that one is self evident.
    This can be fixed easily by accounting for individual players' hive playtime acting as a multiplier of hive skill.
    Simple example: A veteran with 1500elo would be rated 1500x2=3000 effective rating while a rookie with 1500elo would be rated 1500x0.5=750 effective rating.
    You didn't read the proposed fix I wrote up. I will link it a fourth time here. It literally adjusts a players skill as a function of their playtime.
    Me9a wrote: »
    Player Rentation:
    Migalksi has a important Information about Player Rentation in his post:
    http://forums.unknownworlds.com/discussion/comment/2296205/#Comment_2296205

    That he cant Team up with his Mates
    and all other Modern games have a Party/Team System.
    (thats why he (and probably a lot of of other People) prefer to play other Games)
    THIS needs to be added to NS2.
    (maximum party size should probably be 4-6 Players)

    If these Teams then are very badly balanced after all shuffles suggest a split up with a popup message to the Party/Group.
    It would be great if we could have a party system. The best we can get right now is a gather. I can't imagine a party system working all that great with our current server browser method of finding games. As far as I am concerned it would need some sort of matchmaking. The problem with all that is we don't have a large enough playerbase to support matchmaking. With the current playerbase we could never find players of near skill to create the mythical balanced game.
    You do make a suggestion on how to make parties work with our current NS2. I am not saying it is impossible. There may be a decent way of doing it with our current server browser method of finding games. Balance will be harder to find, but what does it matter if we never find it as we are now.
    Me9a wrote: »
    Suggestion:
    There also should be 1 Hiveskill for each Race
    1 for Aliens
    1 for Marines
    Nordic wrote: »
    [*] Hive has a single skill value for both marines and aliens, even though a player can be wildly more effective with one faction over the other. The proposed fix that one is self evident.
    [/list]
    Maybe add an Hiveskillranking for Commanders too as a Badge with number in the Scoreboard.
    We agree on that. One of the supposed fixes of hive 2.0 is that it would have a separate skill value for marines, aliens, commanders, and khammanders.
    I don't like the idea of showing a commanders skill on the scoreboard as an extension of that I think everyones skill value should be completely hidden.
  • YojimboYojimbo England Join Date: 2009-03-19 Member: 66806Members, NS2 Playtester, NS2 Map Tester, Reinforced - Supporter, Reinforced - Silver, Reinforced - Shadow
    Therius wrote: »
    I love these conversations. One guy posts a massive post composed of statistical data and sound conclusions while the other goes "nope, can't agree because my own intuition along with this one single number I just pulled out of my arse contradicts your research".

    Such is the basis of all internet arguments :smiley:
  • VetinariVetinari Join Date: 2013-07-23 Member: 186325Members, Squad Five Blue, Reinforced - Shadow, WC 2013 - Silver
    Therius wrote: »
    I love these conversations. One guy posts a massive post composed of statistical data and sound conclusions while the other goes "nope, can't agree because my own intuition along with this one single number I just pulled out of my arse contradicts your research".

    The worst thing about it is that it repeats over and over again.
  • NordicNordic Long term camping in Kodiak Join Date: 2012-05-13 Member: 151995Members, NS2 Playtester, NS2 Map Tester, Reinforced - Supporter, Reinforced - Silver, Reinforced - Shadow
    Therius wrote: »
    I love these conversations. One guy posts a massive post composed of statistical data and sound conclusions while the other goes "nope, can't agree because my own intuition along with this one single number I just pulled out of my arse contradicts your research".

    The worst thing about it is that it repeats over and over again.

    I sometimes feel sad about my chosen use of time when I look back and I see I have been apart of nearly everyone of these conversations since hive began. :frowning:
  • WakeWake Join Date: 2003-03-05 Member: 14351Members, Constellation
    Yojimbo wrote: »
    Therius wrote: »
    I love these conversations. One guy posts a massive post composed of statistical data and sound conclusions while the other goes "nope, can't agree because my own intuition along with this one single number I just pulled out of my arse contradicts your research".

    Such is the basis of all internet arguments :smiley:

    Remove internet from your prosposition, it'superfluous :)
  • WakeWake Join Date: 2003-03-05 Member: 14351Members, Constellation
    Therius wrote: »
    I love these conversations. One guy posts a massive post composed of statistical data and sound conclusions while the other goes "nope, can't agree because my own intuition along with this one single number I just pulled out of my arse contradicts your research".

    The worst thing about it is that it repeats over and over again.

    It's the Groundhog Day syndrome.
    That's what hell is made of.
  • SantaClawsSantaClaws Denmark Join Date: 2012-07-31 Member: 154491Members, Reinforced - Shadow
    I don't think there's anything wrong with basing the skill rating on winrates, aka. ELO.

    However, I have very big doubts on the effectiveness of global ranking in general, like what has been attempted with Hive.

    Notice that Starcraft, ie., uses 'regions'. A diamond league player from the Korean server, could likely beat a Masterleague player from the American server.

    NS2 suffers from a similar effect. NS2 thrives in somewhat-isolated communities. Some of these communities, attract certain skill levels, some better and worse than others.

    As long as individuals in these communities play among themselves, Hive should, over time, estimate the individual players real skill relative to the other individuals in that community.

    However, the problem arises the moment one of these individuals decides to intermingle with another community. His hive ranking will no longer be accurate relative to the new community he just invaded.

    This is why you can have high level players, with really low Hive rank. Because they insist on, as they should, playing with other high level players. They are ranked, relative to other high ranking players. But the moment they join a random pub, this relative rank no longer applies, and you get imbalanced games.

    One solution, unlikely as it is, would be to confine everyone in isolated communities, so that big skill gapped players can't intermingle.

    Another solution, more likely perhaps, would be to localise the rankings. So when you join Wooza you have one rank, and when you join Hellfire you have another rank. This way, outsiders can't compromise the accuracy of the relative rankings, except for the fact that they have to be re-evaluated each time per community. That, I think, would make for a much better system. The global ranking thing - I sincerely doubt would ever work.
  • VetinariVetinari Join Date: 2013-07-23 Member: 186325Members, Squad Five Blue, Reinforced - Shadow, WC 2013 - Silver
    Nordic wrote: »
    Therius wrote: »
    I love these conversations. One guy posts a massive post composed of statistical data and sound conclusions while the other goes "nope, can't agree because my own intuition along with this one single number I just pulled out of my arse contradicts your research".

    The worst thing about it is that it repeats over and over again.

    I sometimes feel sad about my chosen use of time when I look back and I see I have been apart of nearly everyone of these conversations since hive began. :frowning:

    One day you will look back at your life with your grandkids and say: "Look at all that quality content I produced between 2012 and 20XX. Look at how many awesomes they gave me."
    And it will totally be not worth it.
  • WakeWake Join Date: 2003-03-05 Member: 14351Members, Constellation
    Nordic wrote: »
    I sometimes feel sad about my chosen use of time when I look back and I see I have been apart of nearly everyone of these conversations since hive began. :frowning:

    You walked the walk instead of talking the talk and you regret it ???? Someone snip that, quick, PM that guy !
  • RevanCoranaRevanCorana Join Date: 2015-08-14 Member: 207125Members
    Nordic wrote: »
    The hive system is mathematically correct but has a huge delay which leads to imprecise shuffles.
    Nope. Hive is not perfect as I have described in this thread. It is not so much that hive is imprecise, but that we have a small playerbase that ruins shuffles. No one can find near skilled games. There is practically no such thing as near skilled games in ns2.
    Nordic wrote: »
    Hive is not perfect, as I mentioned. It can be a lot better. For example:
    • Hive has some rookies start at 0 hive skill and others start at 1000 hive skill. This is known, and has a proposed fix.
    • Hive takes too long to find a reasonable estimate of a players skill. There is also a proposed fix.
    • Hive has a single skill value for both marines and aliens, even though a player can be wildly more effective with one faction over the other. The proposed fix that one is self evident.
    As I said. Skill rating based on winrate works mathematically as a concept but it is slow in reality (ie it requires many hours of game to be accurate which means it is not accurate for most players).

    Nordic wrote: »
    This can be fixed easily by accounting for individual players' hive playtime acting as a multiplier of hive skill.
    Simple example: A veteran with 1500elo would be rated 1500x2=3000 effective rating while a rookie with 1500elo would be rated 1500x0.5=750 effective rating.
    You didn't read the proposed fix I wrote up. I will link it a fourth time here. It literally adjusts a players skill as a function of their playtime.
    This thing is about adressing the unbalance created by old rookies having 1000 elo and new rookies with 0 elo (which is not even that big of a deal to start with).
    What I was talking about was the more general idea that ELO score alone does not represent actual effectiveness because it doesnt account for game knowledge.
  • NordicNordic Long term camping in Kodiak Join Date: 2012-05-13 Member: 151995Members, NS2 Playtester, NS2 Map Tester, Reinforced - Supporter, Reinforced - Silver, Reinforced - Shadow
    edited June 2016
    Nordic wrote: »
    The hive system is mathematically correct but has a huge delay which leads to imprecise shuffles.
    Nope. Hive is not perfect as I have described in this thread. It is not so much that hive is imprecise, but that we have a small playerbase that ruins shuffles. No one can find near skilled games. There is practically no such thing as near skilled games in ns2.
    Nordic wrote: »
    Hive is not perfect, as I mentioned. It can be a lot better. For example:
    • Hive has some rookies start at 0 hive skill and others start at 1000 hive skill. This is known, and has a proposed fix.
    • Hive takes too long to find a reasonable estimate of a players skill. There is also a proposed fix.
    • Hive has a single skill value for both marines and aliens, even though a player can be wildly more effective with one faction over the other. The proposed fix that one is self evident.

    As I said. Skill rating based on winrate works mathematically as a concept but it is slow in reality (ie it requires many hours of game to be accurate which means it is not accurate for most players).
    Read what you quoted.
    [*] Hive takes too long to find a reasonable estimate of a players skill. There is also a proposed fix.

    A solution was proposed. Is it a flaw? Yes. Is it a flaw of the whole system? No.
    Nordic wrote: »
    This can be fixed easily by accounting for individual players' hive playtime acting as a multiplier of hive skill.
    Simple example: A veteran with 1500elo would be rated 1500x2=3000 effective rating while a rookie with 1500elo would be rated 1500x0.5=750 effective rating.
    You didn't read the proposed fix I wrote up. I will link it a fourth time here. It literally adjusts a players skill as a function of their playtime.

    This thing is about adressing the unbalance created by old rookies having 1000 elo and new rookies with 0 elo (which is not even that big of a deal to start with).
    What I was talking about was the more general idea that ELO score alone does not represent actual effectiveness because it doesnt account for game knowledge.
    Does more game knowledge increase the chance of someone winning? If the answer is yes, then it is indirectly measured by hive.
  • RevanCoranaRevanCorana Join Date: 2015-08-14 Member: 207125Members
    Does more game knowledge increase the chance of someone winning? If the answer is yes, then it is indirectly measured by hive.

    Basically you're saying that if you match team of vets (2k hour) vs team of newbs (100 hours) and every single player has 1000 elo each team have equal chance of wining.

    laughable


  • NordicNordic Long term camping in Kodiak Join Date: 2012-05-13 Member: 151995Members, NS2 Playtester, NS2 Map Tester, Reinforced - Supporter, Reinforced - Silver, Reinforced - Shadow
    edited June 2016
    Does more game knowledge increase the chance of someone winning? If the answer is yes, then it is indirectly measured by hive.
    Basically you're saying that if you match team of vets (2k hour) vs team of newbs (100 hours) and every single player has 1000 elo each team have equal chance of wining.
    laughable

    If you could hypothetically find a team of veterans (2k hours) and a team of newbs (100 hours) and they all had 1000 skill, this is what would happen. The veterans, being veterans, would win more. A LOT more. The veterans hive skill would go up, while the newbs hive skill down. Game knowledge was just measured.

    There is nothing wrong with hive in this scenario. Hive is just learning. All skill systems must learn. Even the supposed industry best skill system, called TrueSkill, takes an average of 90 games to find a statistically significant skill value. We both agree that hive learns a too slow right now, but there is a proposed fix for that. Is it a flaw of the whole hive skill system? No.
  • RevanCoranaRevanCorana Join Date: 2015-08-14 Member: 207125Members
    Wow, you almost understood, that was really close. :/

    Good luck trying to shuffle even teams with a system blind to playtime. lol isn't it so trivial?

    Whoever thought this was a good idea is an idiot.
  • VetinariVetinari Join Date: 2013-07-23 Member: 186325Members, Squad Five Blue, Reinforced - Shadow, WC 2013 - Silver
    Wow, you almost understood, that was really close. :/

    Good luck trying to shuffle even teams with a system blind to playtime. lol isn't it so trivial?

    Whoever thought this was a good idea is an idiot.

    I've actually gone up and re-read your posts, and I've come to the conclusion that you have no understanding of what the math does here or how statistics work.
    What you're proposing is simply not how it works.

    Besides, in this example:
    Does more game knowledge increase the chance of someone winning? If the answer is yes, then it is indirectly measured by hive.

    Basically you're saying that if you match team of vets (2k hour) vs team of newbs (100 hours) and every single player has 1000 elo each team have equal chance of wining.

    laughable


    You basically say "If the hive score is incorrect, shuffle is useless".
    Well... yeah. Good job figuring that out.
    The thing is that after that bad match created by these arbitrary numbers is that the scores will adjust, to better reflect the situation. That's, like, the whole point of the system.
    Wow, you almost understood, that was really close. :/

    Good luck trying to shuffle even teams with a system blind to playtime. lol isn't it so trivial?

    Whoever thought this was a good idea is an idiot.

    It doesn't need to factor in playtime. The longer you play, the closer your value will be to the "correct" value in relation to everyone else. How exactly would you shuffle differently in the meantime? I would like to point out that your proposed solution of a multiplicator would most definitely not work. Take SoulRider, for example. He has played over 2000 hours, but is, by his own accord, not very good at the game. His skill ranking would ridiculously over inflate.

    Besides, the only thing you would achieve with your theoretical system is that it adjusts for 1000 hour rookies. There's a better, one time fix to this. I would like to quote Nordic.
    Nordic wrote: »
    • Hive has some rookies start at 0 hive skill and others start at 1000 hive skill. This is known, and has a proposed fix.
    I suggest this time you actually click on that link and read it. It's not a lot to read.

    I think we have a small disagreement about who is the idiot here, I'm afraid.
  • MaxAmusMaxAmus UK Join Date: 2003-12-26 Member: 24779Members, Constellation, NS2 Playtester, Reinforced - Shadow
    ahh man, i kinda feel like we have been here before? Different Topic - Same Context?
    jfedu.jpg
  • WakeWake Join Date: 2003-03-05 Member: 14351Members, Constellation
    I think we have a small disagreement about who is the idiot here, I'm afraid.

    That's what Reggie Watts ask at the end of this song ...
  • HandschuhHandschuh Join Date: 2005-03-08 Member: 44338Members, NS2 Playtester, NS2 Community Developer
    I'll see lots of people with 1000+ hours who still can't use a map... playtime doesnt necessarily equal more skill or knowledge.. with that assumption you decrease the use of hiveskill :)
    This experience should lead to better odd and thus more winnings.. if not you havent learned to reflect ylon your errors.

    Its quite simple actually
  • RevanCoranaRevanCorana Join Date: 2015-08-14 Member: 207125Members
    Just pointing out the obvious, winrate alone doesnt produce even teams now feel free to go back put your head in the sand.
  • MaxAmusMaxAmus UK Join Date: 2003-12-26 Member: 24779Members, Constellation, NS2 Playtester, Reinforced - Shadow
    Just pointing out the obvious, winrate alone doesnt produce even teams now feel free to go back put your head in the sand.
    article-2525602-1A2B2A3600000578-553_634x408.jpg

    AM i doing it right?!
  • AeglosAeglos Join Date: 2010-04-06 Member: 71189Members
    Its amusing how people still can't come to terms that the best measure isn't necessarily a perfect/good measure.
  • NordicNordic Long term camping in Kodiak Join Date: 2012-05-13 Member: 151995Members, NS2 Playtester, NS2 Map Tester, Reinforced - Supporter, Reinforced - Silver, Reinforced - Shadow
    Aeglos wrote: »
    Its amusing how people still can't come to terms that the best measure isn't necessarily a perfect/good measure.
    This 1000 more. So true.
    Just pointing out the obvious, winrate alone doesnt produce even teams now feel free to go back put your head in the sand.
    The conversation circles some more. Do I need to go make a condensed version of the massive post I wrote before so you will actually read it?
  • VetinariVetinari Join Date: 2013-07-23 Member: 186325Members, Squad Five Blue, Reinforced - Shadow, WC 2013 - Silver
    Just pointing out the obvious, winrate alone doesnt produce even teams now feel free to go back put your head in the sand.

    Please explain how.
  • RevanCoranaRevanCorana Join Date: 2015-08-14 Member: 207125Members
  • VetinariVetinari Join Date: 2013-07-23 Member: 186325Members, Squad Five Blue, Reinforced - Shadow, WC 2013 - Silver
    I'm serious. Please explain it to me. I do not find it obvious.
Sign In or Register to comment.