Topic: Rating system

81 — 90 of 160 Posts ‹‹ previous 1 2 ... 7 8 9 10 11 ... 15 16 next ››

king_of_nowhere Joined: 2014-09-15, 17:35 Posts: 1668 One Elder of Players	Posted at: 2019-08-03, 19:20 teppo wrote: WorldSavior wrote: But I guess that if it will be good if both players have to agree before every match about that? `ranking = 1000.0 * number_of_win_games / (number_of_win_games + number_of_lost_games)` This isn't a meaningful value. No. If player A is better than player B, player A can still easily have a much lower winning percentage than player B. And number of draws is missing also. If we adopt this system, I will admit to ranking only once, and make sure to win that one game. After that, I will shine on the top forever with my perfect score. which just shows that that way of ranking is stupid. there are perfectly tested and true algorithms that has been mentioned and that work, and don't have that problem, and it wouldn't be complicated to use them. hell, i could calculate ELO manually if I had to. Top Quote

einstein13 Joined: 2013-07-28, 23:01 Posts: 1118 One Elder of Players Location: Poland	Posted at: 2019-08-03, 22:16 I have a question(s): Why are we talking about ranking that is supposed to be at most secondary? Arguing about something that is pure statistics? As I understand, most of us agreed that major ranking should be Glicko-2 on officially selected maps. The only "argue" can be about if we count other than 1 vs. 1 games and how often the ranking should be updated. einstein13 calculations & maps packages: http://wuatek.no-ip.org/~rak/widelands/ backup website files: http://kartezjusz.ddns.net/upload/widelands/ Top Quote

teppo Joined: 2012-01-30, 08:42 Posts: 423 Tribe Member	Posted at: 2019-08-04, 05:34 king_of_nowhere wrote: there are perfectly tested and true algorithms that has been mentioned and that work, and don't have that problem, and it wouldn't be complicated to use them. hell, i could calculate ELO manually if I had to. Oh, I forgot about that while reading through the thread.. Thanks for pointing out. I agree with you. einstein13 wrote: Why are we talking about ranking that is supposed to be at most secondary? Arguing about something that is pure statistics? As I understand, most of us agreed that major ranking should be Glicko-2 on officially selected maps. I must have misunderstood an older post in this thread, and thought that the current suggestion is to use a win/lose ratio to rank players. Top Quote

WorldSavior Joined: 2016-10-15, 03:10 Posts: 2094 OS: Linux Version: Recent tournament version One Elder of Players Location: Germany	Posted at: 2019-08-04, 12:13 teppo wrote: king_of_nowhere wrote: there are perfectly tested and true algorithms that has been mentioned and that work, and don't have that problem, and it wouldn't be complicated to use them. hell, i could calculate ELO manually if I had to. Oh, I forgot about that while reading through the thread.. Thanks for pointing out. I agree with you. einstein13 wrote: Why are we talking about ranking that is supposed to be at most secondary? Arguing about something that is pure statistics? As I understand, most of us agreed that major ranking should be Glicko-2 on officially selected maps. I must have misunderstood an older post in this thread, and thought that the current suggestion is to use a win/lose ratio to rank players. I interpreted the discussion the same way Winning/losing/drawing percentages as secondary values are welcome. So 100 may be a better factor than 1000 But maybe there shouldn't be a list where players are sorted by the percentage, the percentages could be in the list where players are sorted by ranking... teppo wrote: WorldSavior wrote: But I guess that if it will be good if both players have to agree before every match about that? `ranking = 1000.0 * number_of_win_games / (number_of_win_games + number_of_lost_games)` This isn't a meaningful value. No. If player A is better than player B, player A can still easily have a much lower winning percentage than player B. And number of draws is missing also. If we adopt this system, I will admit to ranking only once, and make sure to win that one game. After that, I will shine on the top forever with my perfect score. That would only work if I would stop to win every game Wanted to save the world, then I got widetracked Top Quote

king_of_nowhere Joined: 2014-09-15, 17:35 Posts: 1668 One Elder of Players	Posted at: 2019-08-04, 12:49 WorldSavior wrote: I interpreted the discussion the same way Winning/losing/drawing percentages as secondary values are welcome. So 100 may be a better factor than 1000 ;-) actually, a better - and related - parameter would be the average rating of opponents. See, if you have a high rating but low win rate, it means you are consistently playing against other high rated opponents. conversely, if you have low rating and high win rate, you must be going out of your way to find players even weaker than you are. and from the other side, if you have a high rating and your opponent's rating is well below yours, it can be inferred that you win most games. Top Quote

WorldSavior Joined: 2016-10-15, 03:10 Posts: 2094 OS: Linux Version: Recent tournament version One Elder of Players Location: Germany	Posted at: 2019-08-04, 15:08 king_of_nowhere wrote: WorldSavior wrote: I interpreted the discussion the same way Winning/losing/drawing percentages as secondary values are welcome. So 100 may be a better factor than 1000 ;-) actually, a better - and related - parameter would be the average rating of opponents. See, if you have a high rating but low win rate, it means you are consistently playing against other high rated opponents. conversely, if you have low rating and high win rate, you must be going out of your way to find players even weaker than you are. and from the other side, if you have a high rating and your opponent's rating is well below yours, it can be inferred that you win most games. One could implement both Wanted to save the world, then I got widetracked Top Quote

trimard Topic Opener Joined: 2009-03-05, 21:40 Posts: 230 Widelands-Forum-Junkie Location: Paris	Posted at: 2019-08-04, 16:08 One could implement both Yes I think that's the idea here. We'll keep tons of different rating system, nice for test, more relevant for some people. And only keep glicko2 as the official one. edit: I'm advancing slowly but I think I might soon have at least the win/lost ratio ready Edited: 2019-08-04, 16:14 Top Quote

GunChleoc Joined: 2013-10-07, 14:56 Posts: 3324 One Elder of Players Location: RenderedRect	Posted at: 2019-08-09, 17:19 Anyway having a tag 'officially balanced' would be nice. We already have an "official" and a "balanced" tag. Combine them, and presto pronto, you'lll get "officially balanced". Once we start implementing games data submission into Widelands itself, we will also need a privacy policy for that. I have already created a bug report for the "Surrender" button, because this will be an interesting feature in any case https://bugs.launchpad.net/widelands/+bug/1839636 Busy indexing nil values Top Quote

WorldSavior Joined: 2016-10-15, 03:10 Posts: 2094 OS: Linux Version: Recent tournament version One Elder of Players Location: Germany	Posted at: 2019-08-09, 18:44 GunChleoc wrote: Anyway having a tag 'officially balanced' would be nice. We already have an "official" and a "balanced" tag. Combine them, and presto pronto, you'lll get "officially balanced". Exactly I have already created a bug report for the "Surrender" button, because this will be an interesting feature in any case https://bugs.launchpad.net/widelands/+bug/1839636 Good idea. It would be good if one could not trigger it by accident, for example it could need confirmation, and then the confirmation could be "immune" to the ctrl-button which usually skips confirmations. Maybe this would be safe enough, one could also make it even safer, but maybe that's not necessary. Wanted to save the world, then I got widetracked Top Quote

kaputtnik Joined: 2013-02-18, 19:48 Posts: 2439 OS: Archlinux Version: current master One Elder of Players Location: Germany	Posted at: 2019-08-09, 19:18 WorldSavior wrote: GunChleoc wrote: Anyway having a tag 'officially balanced' would be nice. We already have an "official" and a "balanced" tag. Combine them, and presto pronto, you'lll get "officially balanced". Exactly I thought the term was meant to be used for uploaded maps, or a list of maps filled with shipped (installed with the game) and uploaded maps. Maybe i just give the word 'new' (in the last sentence, highlighted here) too many importance. einstein13 wrote: WorldSavior wrote: einstein13 wrote: WorldSavior wrote: Also on very imbalanced maps? It depends on rules that will be applied to the ranking. But if it will be allowed, why not? Because playing on imbalanced maps can be equal to manipulate the ratings. So it shouldn't be allowed. Fair point. So there should be a list of maps that are "officially balanced". But how is it possible to make the list easy to fill with new maps? How should we decide which map is equal enough? Fight simulator for Widelands: https://wide-fighter.netlify.app/ Top Quote

‹‹ previous 1 2 ... 7 8 9 10 11 ... 15 16 next ››