|
France12904 Posts
On November 20 2020 06:48 dbRic1203 wrote:Show nested quote +On November 20 2020 05:34 yubo56 wrote:On November 20 2020 04:58 Dave4 wrote:On November 20 2020 04:21 lostusername123 wrote:
In this Thread: people don't understand how aligulac works, or how these ratings are self correcting.
You could give a 1000 Aligulac Rating lead to koreans when they start off, Serral would still have taken #1.
No Korean should be ahead of Serral result wise, none of them come close to Serral or Reynor for that matter.
Agreed. Aligulac has 10 years worth of cross regional play that is built into the ratings inter- and intra-region. There is no objectively better metric than aligulac. Aligulac measures pattern of form extremely well but of course nothing can predict how someone will play 'on the day' so its never 100pc accurate. But much better than just randomly grouping players on gut feel or who your unconscious bias considers to be better than who. In any other sport if gold, silver and bronze went to one country oe region you would consider that country or region to be the best. Admittedly it is a bit different in SC2 because we tend to compare one country against all others but nethertheless the 3 current best players live outside of the home of Starcraft. "Best" is a fuzzy word, stating that as a fact is bound to raise controversy. An alternative interpretation is that the best player in the world and his two biggest enemies live outside of the home of Starcraft, or that the two best players and their designated killer live outside the home of Starcraft. All of these would have pretty similar signatures on Aligulac, simply for the fact that beating Serral gets you a lot of points there. Aligulac also doesn't account for offline-only, preparation-driven anomalies like TY (and Maru to a lesser extent). I think this is not as big of a problem as people think when predicting cross-region tournaments, since the big cross-region tournaments are weekenders and not preparation-driven, but that always does have to be mentioned as well. People will weigh IEM and GSL success differently in defining "best", some favoring the scale and fame of IEM and others the tested prestige of the GSL. Also, the leap of "Aligulac objectively is the best metric we have" to "Aligulac objectively is the best metric" is not very fair, just because it's the best measure in existence doesn't mean it has to do its job well, just that it has to its job better than others. Big proclamations feel good and drive hype, but to claim them as fact is little better than trolling. Particularly calling Clem one of the three best players of SC is speculation at this point. I believe it is true as well, based on his unmatched TvZ, but that's a forecast and not a fact. Facts require evidence, and Aligulac is nothing more than educated forecasting when it comes to matchups it has not yet seen (notably, Clem vs in-form Koreans). Agree with this. When Clem is playing at his best, his micro is better than that of every other Terran, currently playing this game, probably the best ever in SC2. But 1) micro isn t everything 2) he s realy momentum based Player, who still has to work on out-of-game aspects as well, wich come with experience. Despite him beeing older than Reynor, bis Z counterpart is way more experience in high stack matches die to his faster rise in 2018 and 2019
I am a big fan of Clem and French as well but not really.
I would not put his micro above Maru / peak ByuN for example.
|
On November 20 2020 16:46 Dave4 wrote:
To clear up a few misconceptions:
1) No I am not a troll. I truly believe Serrals dominance over 3 years is an uncontested feat in Starcraft 2 in a period when the game may not be its deepest, but certainly is the most figured out. Serrals peak is the highest and has been the longest in both time duration and number of tournaments, and has been repeated consistently against foreign players and Korean players alike. If you have a different opinion you are entitled to it but in any other sport a player this far ahead of the competition is usually acknowledged for the facts, which is what i base my opinion on.
2) Korean level of play is currently the highest it ever has been. Again, it may not be the deepest, but the game is far more figured out now than it was in any other era. Any assertion that this is not the case is assuming an alternate universe where the players of that era "got stronger over time".
3) As i mentioned Aligulac is not perfect at predicting results. It measures form over time using objective statistical measures. It is the best available level measure of skill. Applying subjective beliefs on top of that is fine as long as you acknowledge you are exiting mathematic evidence.
4) In an era where offline events can no longer exist, using this as an argument is no longer relevant.
And as I previously mentioned - any other sport in the world where gold silver bronze belong to one country/region, people almost unanimously agree that region is on top. I certainly have believed it about Korea when the reverse applied. You cant have your cake and eat it too - if you want esports to be legitimately considered a sport in this way you need to be willing to concede to the facts available, or acknowledge your unconscious bias.
P.s i am not trying to be obnoxious or anything I am trying to just state the facts. I appreciate others have a different view of the world but when the evidence contradicts that view you have to accept that people will bawk at it.
As one of the people who originally worked on Aligulac and organizing the matches, I believe your claim about Serral is warranted considering his placement at the top of the "Hall of Fame":
http://aligulac.com/records/hof/
|
On November 20 2020 04:21 lostusername123 wrote:
In this Thread: people don't understand how aligulac works, or how these ratings are self correcting.
You could give a 1000 Aligulac Rating lead to koreans when they start off, Serral would still have taken #1.
No Korean should be ahead of Serral result wise, none of them come close to Serral or Reynor for that matter.
Koreans actually start with a rating of 1200 instead of 1000 (or at least they did back during the start of HotS when I last worked on the site).
|
On November 20 2020 17:32 Grovbolle wrote:Show nested quote +On November 20 2020 16:46 Dave4 wrote:
To clear up a few misconceptions:
1) No I am not a troll. I truly believe Serrals dominance over 3 years is an uncontested feat in Starcraft 2 in a period when the game may not be its deepest, but certainly is the most figured out. Serrals peak is the highest and has been the longest in both time duration and number of tournaments, and has been repeated consistently against foreign players and Korean players alike. If you have a different opinion you are entitled to it but in any other sport a player this far ahead of the competition is usually acknowledged for the facts, which is what i base my opinion on.
2) Korean level of play is currently the highest it ever has been. Again, it may not be the deepest, but the game is far more figured out now than it was in any other era. Any assertion that this is not the case is assuming an alternate universe where the players of that era "got stronger over time".
3) As i mentioned Aligulac is not perfect at predicting results. It measures form over time using objective statistical measures. It is the best available level measure of skill. Applying subjective beliefs on top of that is fine as long as you acknowledge you are exiting mathematic evidence.
4) In an era where offline events can no longer exist, using this as an argument is no longer relevant.
And as I previously mentioned - any other sport in the world where gold silver bronze belong to one country/region, people almost unanimously agree that region is on top. I certainly have believed it about Korea when the reverse applied. You cant have your cake and eat it too - if you want esports to be legitimately considered a sport in this way you need to be willing to concede to the facts available, or acknowledge your unconscious bias.
P.s i am not trying to be obnoxious or anything I am trying to just state the facts. I appreciate others have a different view of the world but when the evidence contradicts that view you have to accept that people will bawk at it. As one of the people who originally worked on Aligulac and organizing the matches, I believe your claim about Serral is warranted considering his placement at the top of the "Hall of Fame": http://aligulac.com/records/hof/
While I have no objections to Aligulac's mathematical integrity or the HOF's validity with given definition, I see it important to note that the representation of the HOF list can be little bit misleading if only looking at pure numeric values. Past top HOFers' appearance on the list is disfigured due ever-increasing number of points in the system, causing continuous rating inflation and making meaningful cumulative average rating gaps that bigger later a player achieved top 7 spot in the list. Those bigger gaps of course make disproportionally bigger sums compared to previous eras with smaller average gaps. If the numeric values could be somehow renormalized for all players relative to their corresponding eras of activity and to then current total points in the system and/or average rating gaps among the top 7, the HOF list would give better image of that group of players over entire history of the data set.
But, as said, matematically there are no faults. Its only representation issue of the list. In a future a top 7 player in Top 7 will get ever increasing amounts of points from their ever increasing rating gaps, over each list. Absolute numeric values must be thus looked at under the light of this systemic feature.
Edit: Lol. I wrote HHOF, when I meant HOF. Corrected.
|
On November 20 2020 17:32 Grovbolle wrote:Show nested quote +On November 20 2020 16:46 Dave4 wrote:
To clear up a few misconceptions:
1) No I am not a troll. I truly believe Serrals dominance over 3 years is an uncontested feat in Starcraft 2 in a period when the game may not be its deepest, but certainly is the most figured out. Serrals peak is the highest and has been the longest in both time duration and number of tournaments, and has been repeated consistently against foreign players and Korean players alike. If you have a different opinion you are entitled to it but in any other sport a player this far ahead of the competition is usually acknowledged for the facts, which is what i base my opinion on.
2) Korean level of play is currently the highest it ever has been. Again, it may not be the deepest, but the game is far more figured out now than it was in any other era. Any assertion that this is not the case is assuming an alternate universe where the players of that era "got stronger over time".
3) As i mentioned Aligulac is not perfect at predicting results. It measures form over time using objective statistical measures. It is the best available level measure of skill. Applying subjective beliefs on top of that is fine as long as you acknowledge you are exiting mathematic evidence.
4) In an era where offline events can no longer exist, using this as an argument is no longer relevant.
And as I previously mentioned - any other sport in the world where gold silver bronze belong to one country/region, people almost unanimously agree that region is on top. I certainly have believed it about Korea when the reverse applied. You cant have your cake and eat it too - if you want esports to be legitimately considered a sport in this way you need to be willing to concede to the facts available, or acknowledge your unconscious bias.
P.s i am not trying to be obnoxious or anything I am trying to just state the facts. I appreciate others have a different view of the world but when the evidence contradicts that view you have to accept that people will bawk at it. As one of the people who originally worked on Aligulac and organizing the matches, I believe your claim about Serral is warranted considering his placement at the top of the "Hall of Fame": http://aligulac.com/records/hof/
Thanks, and also great work on it. Its a wonderful resource that isnt appreciated enough by some on these forums.
|
On November 20 2020 18:03 UnLarva wrote:Show nested quote +On November 20 2020 17:32 Grovbolle wrote:On November 20 2020 16:46 Dave4 wrote:
To clear up a few misconceptions:
1) No I am not a troll. I truly believe Serrals dominance over 3 years is an uncontested feat in Starcraft 2 in a period when the game may not be its deepest, but certainly is the most figured out. Serrals peak is the highest and has been the longest in both time duration and number of tournaments, and has been repeated consistently against foreign players and Korean players alike. If you have a different opinion you are entitled to it but in any other sport a player this far ahead of the competition is usually acknowledged for the facts, which is what i base my opinion on.
2) Korean level of play is currently the highest it ever has been. Again, it may not be the deepest, but the game is far more figured out now than it was in any other era. Any assertion that this is not the case is assuming an alternate universe where the players of that era "got stronger over time".
3) As i mentioned Aligulac is not perfect at predicting results. It measures form over time using objective statistical measures. It is the best available level measure of skill. Applying subjective beliefs on top of that is fine as long as you acknowledge you are exiting mathematic evidence.
4) In an era where offline events can no longer exist, using this as an argument is no longer relevant.
And as I previously mentioned - any other sport in the world where gold silver bronze belong to one country/region, people almost unanimously agree that region is on top. I certainly have believed it about Korea when the reverse applied. You cant have your cake and eat it too - if you want esports to be legitimately considered a sport in this way you need to be willing to concede to the facts available, or acknowledge your unconscious bias.
P.s i am not trying to be obnoxious or anything I am trying to just state the facts. I appreciate others have a different view of the world but when the evidence contradicts that view you have to accept that people will bawk at it. As one of the people who originally worked on Aligulac and organizing the matches, I believe your claim about Serral is warranted considering his placement at the top of the "Hall of Fame": http://aligulac.com/records/hof/ While I have no objections to Aligulac's mathematical integrity or the HOF's validity with given definition, I see it important to note that the representation of the HOF list can be little bit misleading if only looking at pure numeric values. Past top HOFers' appearance on the list is disfigured due ever-increasing number of points in the system, causing continuous rating inflation and making meaningful cumulative average rating gaps that bigger later a player achieved top 7 spot in the list. Those bigger gaps of course make disproportionally bigger sums compared to previous eras with smaller average gaps. If the numeric values could be somehow renormalized for all players relative to their corresponding eras of activity and to then current total points in the system and/or average rating gaps among the top 7, the HOF list would give better image of that group of players over entire history of the data set. But, as said, matematically there are no faults. Its only representation issue of the list. In a future a top 7 player in Top 7 will get ever increasing amounts of points from their ever increasing rating gaps, over each list. Absolute numeric values must be thus looked at under the light of this systemic feature. Edit: Lol. I wrote HHOF, when I meant HOF. Corrected.
Yeah - this is a common problem in FIDE rankings as well. Another good way is to just look at the Length of the period, in this ranking I believe Serral is top 5.
|
1. No pros really refer to the rank of aligulac. It's always wrong with prediction.
2. No 2020 tourny has done with offline. (except IEM) I hope foreigners can win a trophy like Serral in offline vs Koreans.
3. Koreans clearly got worse than before, even according to SpeCial, they do not practice much as foreigners currently. And according to Solar and Serral, no Koreans really practiced for Dreamhack for break after GSL s3 was done.
4. Our hero Serral is a smart guy. He knows people are easily overhyped or overreacted by small results. Let's see if anyone else can do well as our true hero Serral in offline premier events as well.
5. Let's wait until Reynor/Clem wins IEM/GSL/Blizzcon several times. Before that, only Serral is better than Koreans.
|
On November 20 2020 19:36 Grovbolle wrote:Show nested quote +On November 20 2020 18:03 UnLarva wrote:On November 20 2020 17:32 Grovbolle wrote:On November 20 2020 16:46 Dave4 wrote:
To clear up a few misconceptions:
1) No I am not a troll. I truly believe Serrals dominance over 3 years is an uncontested feat in Starcraft 2 in a period when the game may not be its deepest, but certainly is the most figured out. Serrals peak is the highest and has been the longest in both time duration and number of tournaments, and has been repeated consistently against foreign players and Korean players alike. If you have a different opinion you are entitled to it but in any other sport a player this far ahead of the competition is usually acknowledged for the facts, which is what i base my opinion on.
2) Korean level of play is currently the highest it ever has been. Again, it may not be the deepest, but the game is far more figured out now than it was in any other era. Any assertion that this is not the case is assuming an alternate universe where the players of that era "got stronger over time".
3) As i mentioned Aligulac is not perfect at predicting results. It measures form over time using objective statistical measures. It is the best available level measure of skill. Applying subjective beliefs on top of that is fine as long as you acknowledge you are exiting mathematic evidence.
4) In an era where offline events can no longer exist, using this as an argument is no longer relevant.
And as I previously mentioned - any other sport in the world where gold silver bronze belong to one country/region, people almost unanimously agree that region is on top. I certainly have believed it about Korea when the reverse applied. You cant have your cake and eat it too - if you want esports to be legitimately considered a sport in this way you need to be willing to concede to the facts available, or acknowledge your unconscious bias.
P.s i am not trying to be obnoxious or anything I am trying to just state the facts. I appreciate others have a different view of the world but when the evidence contradicts that view you have to accept that people will bawk at it. As one of the people who originally worked on Aligulac and organizing the matches, I believe your claim about Serral is warranted considering his placement at the top of the "Hall of Fame": http://aligulac.com/records/hof/ While I have no objections to Aligulac's mathematical integrity or the HOF's validity with given definition, I see it important to note that the representation of the HOF list can be little bit misleading if only looking at pure numeric values. Past top HOFers' appearance on the list is disfigured due ever-increasing number of points in the system, causing continuous rating inflation and making meaningful cumulative average rating gaps that bigger later a player achieved top 7 spot in the list. Those bigger gaps of course make disproportionally bigger sums compared to previous eras with smaller average gaps. If the numeric values could be somehow renormalized for all players relative to their corresponding eras of activity and to then current total points in the system and/or average rating gaps among the top 7, the HOF list would give better image of that group of players over entire history of the data set. But, as said, matematically there are no faults. Its only representation issue of the list. In a future a top 7 player in Top 7 will get ever increasing amounts of points from their ever increasing rating gaps, over each list. Absolute numeric values must be thus looked at under the light of this systemic feature. Edit: Lol. I wrote HHOF, when I meant HOF. Corrected. Yeah - this is a common problem in FIDE rankings as well. Another good way is to just look at the Length of the period, in this ranking I believe Serral is top 5.
Of course in certain sense the HOF is 'renormalizing' and absolute numeric values meaningful, as we are still in the context of about one decade, meaning that it can be assumed that for every guy (alive and active) at the top 7 in the beginning of HOF relevant era had/still have every chances to be there now and benefit from those bigger rating gaps themselves too. Only requirement is being enough good in the game to be in the Top 7, preferably with enough consistency to stay there for prolonged periods of time. Level of competition of an era doesn't affect to that as most of guys appearing on the list are still playing the game and some long time veterans still accumulate points benefiting from bigger rating gaps equally with "newcomers" reached later on the HOF list. The list is automatically 'renormalized' always against the players' peers of that time, and that is exactly one of most important parts of any good HOF using statistics as their inclusion criteria.
Direct comparisons with the past are difficult, and this is the problem of the listed values, but it can be easily argued also that if some past HOFer X would've had ability and necessary skill to stay on the Top 7 then, he would highly probably appear there also now. Older HOFer cannot ever use bigger rating gaps as excuse against younger HOFer as long as there are reasonable time frame he could still himself benefit from the "perks" of the future. He can only look at the mirror and ask from himself, why I am not still playing there near or on the top?
In this sense absolute numbers are objective over the entire history of Aligulac's data set: and this criteria is also openly and directly stated in the HOF page.
|
Fluffy Friday for ya'll.
European takes the podium, blizzard washes their hand. We've done a great deal in balancing starcraft 2. No more updates patches for ya'll. GG no RE.
|
It's certainly interesting to see South Korea fall to the Europeans though regardless of which players you support. Will be interesting to see if they can regroup, or if this is the beginning of the end for them?
|
There is no way Serral is GOAT. That's silly. His 'legacy' is built on the backs of beating his fellow EU players, and the occasional korean playing overseas.
Wouldn't make it to ro8 in GSL
|
On November 21 2020 12:10 rogzardo_ wrote:
There is no way Serral is GOAT. That's silly. His 'legacy' is built on the backs of beating his fellow EU players, and the occasional korean playing overseas.
Wouldn't make it to ro8 in GSL
I have reported this troll 8-post hero noting that Serral has won an array of cross regional tournaments online and offline, including blizzcon, gsl vw, and most recently dreamhack.
|
ALLEYCAT BLUES50530 Posts
|
Serral is ofcourse one of the all time greats, but he wont be considered the greatest until he conquers GSL, Koreans have dominated in GSL and Blizzcons Serral has not.
|
On November 21 2020 13:37 Dedraterllaerau wrote:
Serral is ofcourse one of the all time greats, but he wont be considered the greatest until he conquers GSL, Koreans have dominated in GSL and Blizzcons Serral has not.
Why would winning GSL prove that exactly? The 3 best players in the world currently arent even in that tournament.
|
On November 20 2020 16:46 Dave4 wrote:
2) Korean level of play is currently the highest it ever has been.
I don't know which kind of mental gymnastics you have to do to come to this conclusion........
no words
|
Czech Republic12129 Posts
On November 21 2020 23:18 Charoisaur wrote:Show nested quote +On November 20 2020 16:46 Dave4 wrote:
2) Korean level of play is currently the highest it ever has been. I don't know which kind of mental gymnastics you have to do to come to this conclusion........ no words
Don't forgtet he's not trolling!  Therefore he has to be Serral, damn, cereal, serial? I mean c'mon, everybody does this. You can't be the GOAT if everybody around you is worse than players before! They have to be at the peakest everst levels so he can be GOAT even without winning multiple great titles like Rogue, or big Korean S tier titles like Maru(OSL, SSL, Code S).
|
On November 21 2020 11:36 Dave4 wrote:
It's certainly interesting to see South Korea fall to the Europeans though regardless of which players you support. Will be interesting to see if they can regroup, or if this is the beginning of the end for them?
This would have been a good way to end your initial post, I think people on both sides would agree, and far fewer people would have taken issue with your post. I certainly agree, despite disagreeing (as I stated earlier) with much of your post, and maybe we can leave it at that We all root for good games, right?
|
If Korean level isn't as high as it is, how it is possible there is still some Koreans in TOP 10.
If Korean level wouldn't be that high it is there would be more Foreigners in TOP 5.
If Korean level will be so high and foreigner usurpers will be so efficient, then it will be so that the TOP 10 will statistically show the parity very soon.
If 50/50 parity in the TOP 10 happens, then we can be sure that the parity is 50/50 considering GSL/whatever Koreans-choose, tournament participation. Something that would immediately justify abolition of all kinds of region locks. Maybe some minimum quotas could be still reserved for global representation.
In Aligulac's context it doesn't matter anyway. The site and the system measure only who is the best in the game, totally regardless of a context of those SC2 games itself. Funny. The fact comes with a delay.
|
Czech Republic12129 Posts
On November 22 2020 08:24 UnLarva wrote:
If Korean level isn't as high as it is, how it is possible there is still some Koreans in TOP 10.
If Korean level wouldn't be that high it is there would be more Foreigners in TOP 5.
If Korean level will be so high and foreigner usurpers will be so efficient, then it will be so that the TOP 10 will statistically show the parity very soon.
If 50/50 parity in the TOP 10 happens, then we can be sure that the parity is 50/50 considering GSL/whatever Koreans-choose, tournament participation. Something that would immediately justify abolition of all kinds of region locks. Maybe some minimum quotas could be still reserved for global representation.
In Aligulac's context it doesn't matter anyway. The site and the system measure only who is the best in the game, totally regardless of a context of those SC2 games itself. Funny. The fact comes with a delay.
What would the abolition or region lock factically changed? Why is everyone so obsessed with this? It's way too late to change that anyway...
Anyway, Korean level dropped. You can dodge it how you can, but the Korean scene lost shit ton of talent. Soulkey, Rain, Flash, Classic, Hero, herO... just few names to drop. We never got any proper replacements.
So, what many are saying is, that Korean level dropped. If you're at 10/10 and you drop to 8/10, while the rest of the world still sits on their 5/10 while some players got to 7-8/10, would that show anything? Hack no. Even in their prime time Koreans were beatable, SC2 is volatile and many know this, there are moments where you lose the game with 1 mistake. FFS even BO3 is considered volatile, that's why some ask for BO5 in the RO16 groups.
Let's add better example. Imagine having a car with 400 hp. After some while some horses will run away and you end with a 350 hp engine. It's less than before, but it's still enough to overtake majority of the normal cars. Except those few new ones which are operating at similar power and the time didn't take their power down.
TL, DR - Koreans had a reserve in skill so they could lose some and still be good.
Edit> TBF I'm more on the side of skill pool shrinking, but a good example is IMO Innovation, when was he in his machine mode the last time and how long it lasted? Do you remember how he used to perform under the KeSPA regime? Also the biggest example of them all - sOs. We used to have a dozen or more serious contenders for the best player. Now? 8 Koreans if Rogue is on fire, Maru's shoulder is fine, TY isn't playing BW, Innovation is in accidental Machine mode, Dark is fine and... well I was going to add Trap and Stats but then I realized they rarely win any tourney 
|
|
|
|
|
|
EPT
