Page 7 of 7 FirstFirst ... 567
Results 61 to 65 of 65

Thread: Which ranking system for Hex ?

  1. #61
    HEX Employee Chark's Avatar
    Join Date
    Aug 2010
    Location
    Lake Forest, CA
    Posts
    1,436
    Quote Originally Posted by frychikn View Post
    try telling this to the people at riot. they will put a bunch of bullshit in front of you(or nothing) and still claim they are right.
    People at Riot are pretty smart. The scale with which they work is completely different then anything I've encountered before I worked there. Their system has to be a black box to avoid exploitation as much as possible, so it doesn't get the benefit of scrutiny of players to improve it. Overall their ladder is a very difficult problem to solve because it has different goals at different levels and because of the massive scale of it. But theoretically they have near unlimited resources, so I suspect they can hire some pretty smart people to improve it (just like for example they did with the player behavior team [the team that made the tribunal system]).

    Quote Originally Posted by Jugg View Post
    Because they ARE right. There is literally no other way to get comparative player rankings, other than basing it purely on wins, losses, and who you won/lost against. Yes it sucks for league due to the combination of long games and its team based nature... but it is literally the only option you have without distorting the game itself. Any other method results in people gaming the system rather than just trying to win their games.

    BTW, they don't use elo, they use Glicker2.
    I think Microsoft's True Skill (tm) would suit their game type better than Elo or Glicko2 (because Elo and Glicko are based on heads-up comparisons between two players, while TS can look at your individual performance within a team or a free-for-all dynamic), but it would require a licensing deal of some sort with Microsoft, since TS isn't for commercial use.
    Last edited by Chark; 03-08-2016 at 05:49 PM.

  2. #62
    Back on the subject of ranking for Hex, for me no free-to-play organized play is the main reason I'm reluctant to spend more than the $200 I have on the game. The fact that not only do I have to pay for my collection, but to play PvP as well is a pretty big turn-off to a game I really want to like.

  3. #63
    Quote Originally Posted by Chark View Post
    I think Microsoft's True Skill (tm) would suit their game type better than Elo or Glicko2 (because Elo and Glicko are based on heads-up comparisons between two players, while TS can look at your individual performance within a team or a free-for-all dynamic), but it would require a licensing deal of some sort with Microsoft, since TS isn't for commercial use.
    They really aren't very different in effect. As I'm sure you know, when Elo/Glicko (and their derivatives) are used in team games, combined ratings of each team are combined/averaged for calculation of the resulting rating changes. Trueskill is great for FFA, but from what I understand, doesn't really offer any huge advantages over the others for team games.

    By "no huge advantages" I'm referring to the fact that when people bitch and moan about "Elo Hell", and rating systems in team games... it is always coming from the position that they want specific in game stats to affect their rating post game. In league it is always "I went 10/0 but still lost due to noob teammates", thinking that their 10/0 implies they should be increasing in rating. To this end, Trueskill and Elo/Glicko are no different, as they only care about who wins and who loses, and who those wins and losses are against. None of them care about KDA, or headshot percentages, or anything.

    This is my understanding anyway... possibly my info is out of date?

  4. #64
    Elo Hell also referred to a bug in earlier seasons of League of Legends where, due to the split 2 Elo system, you could get stuck in bronze rank while fighting people in higher rank. Your ELO gain was based off difference in rank, but your rank game was based (at least during promotion series between ranks) on pure win rate. This would lead to the lowest rank players sometimes getting stuck fighting the highest ranked opponents for the rest of the seasons.

  5. #65
    HEX Employee Chark's Avatar
    Join Date
    Aug 2010
    Location
    Lake Forest, CA
    Posts
    1,436
    Quote Originally Posted by Jugg View Post
    They really aren't very different in effect. As I'm sure you know, when Elo/Glicko (and their derivatives) are used in team games, combined ratings of each team are combined/averaged for calculation of the resulting rating changes. Trueskill is great for FFA, but from what I understand, doesn't really offer any huge advantages over the others for team games.

    By "no huge advantages" I'm referring to the fact that when people bitch and moan about "Elo Hell", and rating systems in team games... it is always coming from the position that they want specific in game stats to affect their rating post game. In league it is always "I went 10/0 but still lost due to noob teammates", thinking that their 10/0 implies they should be increasing in rating. To this end, Trueskill and Elo/Glicko are no different, as they only care about who wins and who loses, and who those wins and losses are against. None of them care about KDA, or headshot percentages, or anything.

    This is my understanding anyway... possibly my info is out of date?
    You're right, it doesn't help team ratings as much as I thought it did (just FFA). It's been a while since I read True Skill implementation. Mostly because I assume I can never use it in a professional setting (because it's proprietary).

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •