EPTAligulac.com changelog and feedback thread - Page 8
| Forum Index > SC2 General |
|
TheBB
Switzerland5133 Posts
| ||
|
graNite
Germany4434 Posts
Is there a way to compare two random players' charts? | ||
|
fezvez
France3021 Posts
I was browsing Remark's page, and I noticed that even though he is supposedly american, when I try to filter the USA players, he doesn't appear (not even under a different name, his rating is 935 and no one has 935) The new all time chart looks super cool! However, players need to have slightly different colors to differentiate them. Also, it is not perfect, when looking at the top 5 protoss, trust appears as the last protoss bonjwa. Clearly, something is wrong | ||
Shellshock
United States97276 Posts
On June 22 2013 10:11 fezvez wrote: Hello, I was browsing Remark's page, and I noticed that even though he is supposedly american, when I try to filter the USA players, he doesn't appear (not even under a different name, his rating is 935 and no one has 935) The new all time chart looks super cool! However, players need to have slightly different colors to differentiate them. Also, it is not perfect, when looking at the top 5 protoss, trust appears as the last protoss bonjwa. Clearly, something is wrong The reason Remark doesnt show up on the America page is that inactive players are removed from the rankings until the play again. He hasn't played a match in a while (about 3 months I believe) so he has been marked as inactive. If you look at FXO NA : http://aligulac.com/teams/6-FXOpen-e-Sports-North-America/ you can see him listed on the inactive roster since it's been so long since he's played a game. | ||
|
jalstar
United States8198 Posts
The dual tournament format is unfair to players and offers advantages and disadvantages based on initial pairings. Here are 3 simulated Flash/Jaedong/Bisu/Stork groups, the first with Flash vs Jaedong first, the second with Flash vs Bisu first, the third with Flash vs Stork first. Top 2 1st 2nd 3rd 4th Top 2 1st 2nd 3rd 4th Top 2 1st 2nd 3rd 4th The chances to advance differ by 10% OR MORE depending on initial pairings. While this format is more entertaining than round robin to me and many others, it's definitely a completely unfair format and no better than flipping a coin to break ties. | ||
|
rename
Estonia329 Posts
On June 22 2013 13:16 jalstar wrote: I'd like to thank Aligulac for offering evidence for a suspicion I've had since I started watching MSL with Lost Saga and Avalon: The dual tournament format is unfair to players and offers advantages and disadvantages based on initial pairings. Here are 3 simulated Flash/Jaedong/Bisu/Stork groups, the first with Flash vs Jaedong first, the second with Flash vs Bisu first, the third with Flash vs Stork first. Top 2 1st 2nd 3rd 4th Top 2 1st 2nd 3rd 4th Top 2 1st 2nd 3rd 4th The chances to advance differ by 10% OR MORE depending on initial pairings. While this format is more entertaining than round robin to me and many others, it's definitely a completely unfair format and no better than flipping a coin to break ties. You are not serious. We could also say that instead of having 8 groups we should just let every player of 32 play each other to guarantee maximum fairness. Because the group seedings aint no better compared to flipping coins to break ties duh huh. | ||
|
TheBB
Switzerland5133 Posts
On June 22 2013 13:16 jalstar wrote: The chances to advance differ by 10% OR MORE depending on initial pairings. While this format is more entertaining than round robin to me and many others, it's definitely a completely unfair format and no better than flipping a coin to break ties. Yeah, you have a point. I'm also strongly in favour of round robin over this dual tournament format. It's not completely correct to say it's not fair because those initial matchups may depend on seeding from previous rounds, but it's definitely much more variable than round robins, that's without question. | ||
|
dainbramage
Australia1442 Posts
| ||
|
rift
1819 Posts
Ranking: All, eSF, KeSPA or foreign team (e.g. Prime, KT Rolster, Evil Geniuses) Records>History: Offline, Online, Both (the overall graph) | ||
|
TheBB
Switzerland5133 Posts
On June 23 2013 07:26 rift wrote: Ranking: All, eSF, KeSPA or foreign team (e.g. Prime, KT Rolster, Evil Geniuses) This is basically #46. On June 23 2013 07:26 rift wrote: Records>History: Offline, Online, Both (the overall graph) This is not likely to happen unless we start computing rating numbers for each of these. I don't see the point behind that really. It's a lot of extra computation and complexity for what I don't imagine will be a very insightful (set of) features. | ||
|
Tuczniak
1561 Posts
I think it's interesting, but I'm not sure if it would tell us something better than monthly winrates, since the ratings changes rather slowly and on top of that you would pick even smaller set of players. Problems are similar to 'op/up' stat, which I find not that useful. So it's just idea. | ||
|
Inimic
Canada153 Posts
Second, if you look at the graphs of all the Kespa players, you find situations like this one a lot: Stats' stats Stats appears to be on a 10 game/series winning streak against players like JYP, Rain, TY and Trap, yet only has a 1344 rating. Going to the most recent 25 games, he's won 19 of them, and his wins then include Soulkey, sOs and Parting. Its like the Aligulac rating system is relying on Soulkey, Innovation and Flash to go out across the earth and steal nerds Aligulac points, and then graciously lose a bunch of times in proleague so that the rest of the Kespa players can have some too- ie, get the reasonable ratings they should have. I wonder if you shouldn't give the Kespa players a bit of the benefit of the doubt and start all Kespa players with a score of 1400. I think its reasonable to suggest that, over time, most Kespa players will prove that they belong above the 1400 line anyways- in this way they can have accurate ratings sooner, their earning of points will upset the code A/code S players scores less, and they won't steal as many points from the point pool in the rest of the world as they climb up to where they're supposed to be at. This roughly equates them to the Aligulac level of Thorzain, who did play in proleague and generally got stomped with at 20% winrate until he left, or Zenio who has a 30% proleague win rate. The thing is, over time if the 1400 was too much it'll just go to stronger players elsewhere as the points gradually spread out over time. I don't think 1400 is too much though. | ||
|
Otolia
France5805 Posts
On June 27 2013 08:34 Inimic wrote: First recommendation: When browsing rankings, its really annoying to get to page 4+. You can have numbers by the arrows to make it more user friendly. Second, if you look at the graphs of all the Kespa players, you find situations like this one a lot: Stats' stats Stats appears to be on a 10 game/series winning streak against players like JYP, Rain, TY and Trap, yet only has a 1344 rating. Going to the most recent 25 games, he's won 19 of them, and his wins then include Soulkey, sOs and Parting. Its like the Aligulac rating system is relying on Soulkey, Innovation and Flash to go out across the earth and steal nerds Aligulac points, and then graciously lose a bunch of times in proleague so that the rest of the Kespa players can have some too- ie, get the reasonable ratings they should have. I wonder if you shouldn't give the Kespa players a bit of the benefit of the doubt and start all Kespa players with a score of 1400. I think its reasonable to suggest that, over time, most Kespa players will prove that they belong above the 1400 line anyways- in this way they can have accurate ratings sooner, their earning of points will upset the code A/code S players scores less, and they won't steal as many points from the point pool in the rest of the world as they climb up to where they're supposed to be at. This roughly equates them to the Aligulac level of Thorzain, who did play in proleague and generally got stomped with at 20% winrate until he left, or Zenio who has a 30% proleague win rate. The thing is, over time if the 1400 was too much it'll just go to stronger players elsewhere as the points gradually spread out over time. I don't think 1400 is too much though. Korean already get a headstart of 1200. You may believe that KeSPA players are godlike and deserve better rankings but there is nothing we can do if they only play a few games per month. The more games played, the better. Also they get offline bonuses other players don't always get. That being said, you can try to give them a bigger headstart and recalculate everything. The ressources are out there. | ||
|
Inimic
Canada153 Posts
MLG Spring Stats has my back  He even took down TY, who Flash named as one of the top 3 terrans in the entire world on LO3 a day or two ago.Again, I think players like him are currently extremely underrated in Aligulac because of the isolation of the Kespa progamers relative to the rest of the world. I think a larger adjustment would help to correct it, with no adverse effects in the long run <3 I can say one last thing though: the 1400 initial number for Kespa progamers could be used for the next year, then a new version revert all koreans to start with the same 1000 as everyone else. Thus, it would be a temporary fix while the ratings wait for the number of games to stack up enough to be accurate. | ||
|
Entirety
1423 Posts
For example, if Stats (or any other KeSPA pro-gamer) is truly underrated, then he will easily breeze through Code A qualifiers and steal points from eSF players. (He did in fact make it through the Code A qualifiers I believe). On the other hand, look at Zest. Zest performs finely against other KeSPA players in ProLeague. If Zest is truly underrated, why hasn't he qualified for Code A? | ||
|
stenole
Norway869 Posts
On June 22 2013 13:16 jalstar wrote: I'd like to thank Aligulac for offering evidence for a suspicion I've had since I started watching MSL with Lost Saga and Avalon: The dual tournament format is unfair to players and offers advantages and disadvantages based on initial pairings. (...) The chances to advance differ by 10% OR MORE depending on initial pairings. While this format is more entertaining than round robin to me and many others, it's definitely a completely unfair format and no better than flipping a coin to break ties. There will also be unfairness in the round-robin format based on what order games are played. In the last rounds of play, some players will already have advanced or have been knocked out. These players will most likely not play as well as they otherwise would. If placement affects how later brackets are arranged, it motivates players to throw games under given circumstances. This is something we saw a lot in WCG during Brood War times. Nothing is worse than watching a game where you know both players are better off losing while still pretending they are playing at their standard level. | ||
|
StarGalaxy
Germany744 Posts
What do you guys thing about that? I think to have "cleaned" winrates and average opponents rating would be really cool for analysing. On June 19 2013 08:30 StarGalaxy wrote: It would be really cool if you could filter the match history depending on the opponents rating. like the date option which is really awesome btw: (opponents rating between___ and____) An other interesting stats besides the displayed winrate would be the average opponents rating. Those two things would help to analyse the matchhistory tremendously. one other thing i noticed. not sure if it is bugged but his vZ rating seems abnormal high: http://aligulac.com/players/286-CupCake/ <3 this site! | ||
|
TheBB
Switzerland5133 Posts
On July 01 2013 00:52 StarGalaxy wrote: My post might have got lost (the last one on the last page). What do you guys thing about that? I think to have "cleaned" winrates and average opponents rating would be really cool for analysing. Noted  . I agree it would be cool, but I'd have to introduce some extra DB fields to do this in a clean and efficient manner, so it's not something I can just slap in there and get working in half an hour. | ||
|
ThunderGod
New Zealand897 Posts
Examples: http://www.teamliquid.net/forum/viewmessage.php?topic_id=141236 (scroll down to team NZ). http://www.teamliquid.net/tlpd/sc2-international/players/2575_Kowi | ||
|
MasterOfPuppets
Romania6942 Posts
On July 02 2013 20:18 ThunderGod wrote: Hi just wanted to point out a mistake. 'Kowi' is listed as Australian when he is a New Zealander. Examples: http://www.teamliquid.net/forum/viewmessage.php?topic_id=141236 (scroll down to team NZ). http://www.teamliquid.net/tlpd/sc2-international/players/2575_Kowi Fixed! Thank you for pointing this out. | ||
| ||
