Ranking OCT 2020 [ELO] - Page 2
Forum Index > BW General |
GruGloG
Netherlands114 Posts
| ||
MaTRiX[SiN]
Sweden1282 Posts
On November 12 2020 04:17 GruGloG wrote: - I really dont think so, the sole goal of a series is to win the series. I dont think it would make any sense to count each game. - Not sure what you mean, should a player have a different mmr for different races? What race someone plays to win or lose doesn´t really matter. - It seems odd to me that winning a series 4-0 counts the same as a 4-3 or 1-0 but I guess that's a matter of opinion then. - I meant that race picking/random is not taken into account for matchup winrates. Rush played vs Flash random but it counts towards his vs terran winrate. For the korean players this doesnt change much but notably a lot of NA players race pick so for the foreigner stats I think it's fairly significant. | ||
Poegim
Poland261 Posts
| ||
b0lt
United States786 Posts
On November 12 2020 04:17 GruGloG wrote: - I really dont think so, the sole goal of a series is to win the series. I dont think it would make any sense to count each game. If someone played flash in a best of 99, going 49-50 would be very different from 0-99. | ||
ox.tQ
792 Posts
Our planned future round robin (“best of the best”) tournament, which features 12 top foreign players and a five game format (330 total games), might prove to be of interest to your independent ranking system...we designed it so that every game will count as almost $10, so each five game showmatch would be worth a total of $50 minimum (with 3k prize pool). In this way, even “losers” of showmatches who won up to two games will be compensated financially. There would be 66 such showmatches in a 12 player round robin tournament. | ||
GruGloG
Netherlands114 Posts
On November 12 2020 13:15 b0lt wrote: If someone played flash in a best of 99, going 49-50 would be very different from 0-99. When we see a competitive BO99 I promise you to personally put every game in there. On November 12 2020 05:09 MaTRiX[SiN] wrote: - I meant that race picking/random is not taken into account for matchup winrates. Rush played vs Flash random but it counts towards his vs terran winrate. For the korean players this doesnt change much but notably a lot of NA players race pick so for the foreigner stats I think it's fairly significant. This is a valid point, not sure how much work it would be to code in race picking in player profiles or what priority it should have but maybe Poegim could fit this on his never ending to do list. | ||
phase
United States399 Posts
On November 11 2020 06:29 ox.tQ wrote: Myfleetingdream created a showmatch rankings sheet for us a while back (see masters result list). Some data there, not sure if it's useful and how much. Unfortunately it hasnt been updated recently. https://docs.google.com/spreadsheets/u/1/d/1r7yIkVXm1t5fgCBtmqzXYy3ddC15602wTfFLMB4GFCI/edit#gid=0&fvid=148457067 It's been updated! On November 12 2020 13:15 b0lt wrote: If someone played flash in a best of 99, going 49-50 would be very different from 0-99. I absolutely think it makes sense to count each game, since it generates additional data for you and helps with storylines. E.g. X has never beaten Y in a head-to-head, but on paper X is stronger than Y because his wins in that matchup are more dominant and his losses are closer. Unfortunately, I can understand if it's the case that the data for tournaments from years ago isn't as clean and doesn't have that granularity. For example, I calculated a similar ELO calculator and Ranking list for the Think Quick showmatch series (https://docs.google.com/spreadsheets/d/1r7yIkVXm1t5fgCBtmqzXYy3ddC15602wTfFLMB4GFCI/edit?usp=sharing), and to improve accuracy, I calculated ELO updates based on individual games as opposed to entire showmatches. I can share the github repo if anyone is interested (the code is super hacked together and messy though lol) Are you guys crawling Liquipedia to get your data? Like Ox mentioned, we might enter all of the ThinkQuick showmatch results into Liquipedia some day, since each match is for prize money. | ||
Poegim
Poland261 Posts
| ||
Duke_nk
38 Posts
| ||
martoto088
32 Posts
Dewalt, Bonyth, parfumetoss higher than ZeLoT/Shine? Specially Shine, he will beat them same fashion as Larva, playing with his foot. Something is very wrong with the algorithm you're using or at least it should be called ELO rankings. | ||
Incomplete..ReV
Norway598 Posts
On November 20 2020 04:45 martoto088 wrote: ... or at least it should be called ELO rankings. It says so in the title and in the OP? | ||
Poegim
Poland261 Posts
| ||
Leonix
161 Posts
How comes Soma i under all these other playerS? I mean will we just ignore the fact that he got 2nd place in the ASL ? and 4rd in the previous one ? ASL RESULTS NEED TO COUNT THE MOST# How comes Shine is under Dewalt ? Is this realistic? Bonyth under Dewalt ? Didnt he just won BSL again? The biggest tournament for the foreign scene? | ||
FakeFin
Germany386 Posts
On November 20 2020 08:33 Leonix wrote: First of all, it needs to be one ranking system for the top players. Division into foreign and korrean can be a nice thing to see the difference, but it needs at least the same point system, if not this is just waste of time for the casual people at least. I get that this is based on results,etc. , but the rankings seem off on both parts... How comes Soma i under all these other playerS? I mean will we just ignore the fact that he got 2nd place in the ASL ? and 4rd in the previous one ? ASL RESULTS NEED TO COUNT THE MOST# How comes Shine is under Dewalt ? Is this realistic? Bonyth under Dewalt ? Didnt he just won BSL again? The biggest tournament for the foreign scene? 1) This only takes games until end of october into account, so later rounds of asl10 and bsl10 aren't in there 2) Just tell Soma to do better in tournaments besides asl 3) Just tell Bonyth to do better in tournaments besides bsl | ||
Jealous
9974 Posts
On November 20 2020 08:33 Leonix wrote: First of all, it needs to be one ranking system for the top players. Division into foreign and korrean can be a nice thing to see the difference, but it needs at least the same point system, if not this is just waste of time for the casual people at least. I get that this is based on results,etc. , but the rankings seem off on both parts... How comes Soma i under all these other playerS? I mean will we just ignore the fact that he got 2nd place in the ASL ? and 4rd in the previous one ? ASL RESULTS NEED TO COUNT THE MOST# How comes Shine is under Dewalt ? Is this realistic? Bonyth under Dewalt ? Didnt he just won BSL again? The biggest tournament for the foreign scene? Hey man, math, rankings, and dates are hard concepts, so let me try to make it a little easier to understand. Let's say I have 10 apples. I've been buying an apple every day for the past 10 days. Boy, I sure love apples! Let's say you aren't a big fan of apples, but your mommy went to the store today and you got FIVE apples! That's huge! Certainly way more apples than I got today, or even the past few days! But, who has the most apples right now - is it you, or me? I have 10 apples, and you have just 5 apples, so I have more apples - even though recently you got a lot of them! Just because I have more apples than you now, it doesn't mean that you can't ask your mommy to buy 5 apples for the next two days and have even more apples than I do! Now, keep in mind, that no matter what we do - our apples are worth the same. After all, they are just apples! It doesn't matter if you got your apples from your mommy or if I bought them myself, they are just apples, no matter how easily we got them or in what way. I hope this helps! | ||
Poegim
Poland261 Posts
| ||
LaStScan
Korea (South)1288 Posts
On November 20 2020 04:45 martoto088 wrote: That is so unaccurate. Lol'd at Piano TOP 7 KR Also Scan top non kr? Dewalt, Bonyth, parfumetoss higher than ZeLoT/Shine? Specially Shine, he will beat them same fashion as Larva, playing with his foot. Something is very wrong with the algorithm you're using or at least it should be called ELO rankings. poegim is doing a hard work and i like seeing his effort organizing datas. But since you don't like his data analysis with rankings, I highly recommend you to look at here then. This is excluded offline tournament record. Only online+sponbbang matches | ||
martoto088
32 Posts
[QUOTE] [img]https://i.imgur.com/EuWivK8.png[/img] [img]https://i.imgur.com/1rIrNg0.png[/img] [img]https://i.imgur.com/FhfwT9W.png[/img] [img]https://i.imgur.com/y17y5bJ.png[/img][/QUOTE] Yeah, that helps a lot :D Any chance to get that in english or some notable translated by you? I apreciate a ton of poegim work and i like seeing his effort organizing datas, but there is no way on earth someone to convice me Shine won't beat Dewalt, Bonyth and parfumetoss playing with his foot Flash 1070.95 ELO, ajuk12 or whoever 1165.13... come on [QUOTE]On November 20 2020 08:33 Leonix wrote: ASL RESULTS NEED TO COUNT THE MOST# [/QUOTE] [QUOTE]On November 20 2020 09:44 FakeFin wrote: [QUOTE] 2) Just tell Soma to do better in tournaments besides asl [/QUOTE] [QUOTE]On November 20 2020 09:44 FakeFin wrote: 1) This only takes games until end of october into account, so later rounds of asl10 and bsl10 aren't in there [/QUOTE] Only qualifing for ASL should count like winning BSL 10 times in a row | ||
FakeFin
Germany386 Posts
On November 20 2020 22:42 martoto088 wrote: Flash 1070.95 ELO, ajuk12 or whoever 1165.13... come on Your elo number just shows how high you're above average, the average player in the korean scene would someone like YSC or Mong, because only the top level competition is taken into account. The average player in the foreign scene would be someone like ClickMer or VinGzil. The reasone why these rankings have to be seperated is because there are not enough matches between the scenes and people would complain that Scan, Dewalt and Bonyth are ranked way higher than Flash in the same ranking | ||
FakeFin
Germany386 Posts
On November 20 2020 22:42 martoto088 wrote: Only qualifing for ASL should count like winning BSL 10 times in a row It does because they both count nothing! Elo ranking is not based on results but based on matches played (it works like MMR but for only tournament matches) | ||
| ||