EPTIEM Katowice 2022: RO24 groups, RO36 bracket announced - P…
| Forum Index > SC2 General |
|
Gina
241 Posts
| ||
|
Fango
United Kingdom8978 Posts
| ||
Poopi
France12750 Posts
On February 01 2022 01:59 Fango wrote: Damn Maru is the number 1 seed and I actually don't think he'll make it out of the group. Yeah if he loses to Zoun again, gets beaten by Clem (bo3 anything can happen) and by either Reynor or herO (let's be realistic, given how stacked group D already is, they gonna send the best ro36 player in it for giggles), which is realistic as well, he might not qualify. I think he has a decent shot though | ||
|
honorablemacroterran
188 Posts
| ||
|
Acrofales
Spain17735 Posts
On February 01 2022 03:13 honorablemacroterran wrote: I don't think there's a real risk of Maru not making it out of the group, but it will definitely be harder for him than most of the other top seeds despite him being #1. Usually, the higher your seed, the easier your path, but in this case it's the opposite. You think Maru has it harder than Trap? Or for that matter, Rogue? I mean... Maru in Trap's group would definitely crush the whole thing. But it's far from an easy group for Trap. The only first seed with an actual easy group is Serral, although I expect him to lose to Dark (they'll both beat everybody else, though). | ||
|
honorablemacroterran
188 Posts
On February 01 2022 03:52 Acrofales wrote: You think Maru has it harder than Trap? Or for that matter, Rogue? I mean... Maru in Trap's group would definitely crush the whole thing. But it's far from an easy group for Trap. The only first seed with an actual easy group is Serral, although I expect him to lose to Dark (they'll both beat everybody else, though). Yeah, Maru has the hardest group. I think it would be fairer to solve for the groups that minimize the variance of the total points across the groups. Edit: I didn't actually exhaust the problem space or do a real solver, but I ran my simulator a few times and this seems to be close to the minimum variance groups as I suggested above: {'a': ['Serral', 'Clem', 'TIME', 'Scarlett', 'Zoun'], 'b': ['Maru', 'Dark', 'Solar', 'Special', 'Showtime'], 'c': ['Rogue', 'Zest', 'HeroMarine', 'Neeb', 'Dream'], 'd': ['Trap', 'Cure', 'Reynor', 'Bunny', 'Lambo']} The variance of this configuration is 173.1875, whereas the variance of the actual groups is 453,794.1875 The group point totals would be thus: a 8960 b 8940 c 8923 d 8944 To me, this not only looks obviously fairer, but also more interesting. I have relatively little interest in watching the actual Group B tbh. Interestingly, it also distributed the Koreans and the foreigners more evenly as well. | ||
|
bela.mervado
Hungary366 Posts
444.75 8951 Maru,T,3185 Zest,P,1976 Solar,Z,1369 Scarlett,Z,1315 Zoun,P,1106 8924 Rogue,Z,2937 Dark,Z,2397 Heromarine,T,1573 Neeb,P,1265 Showtime,P,752 8944 Trap,P,2897 Cure,T,2153 Reynor,Z,1914 Bunny,T,1240 Lambo,Z,740 8948 Serral,Z,2670 Clem,T,2536 Time,T,1333 Special,T,1237 Dream,T,1172 | ||
|
honorablemacroterran
188 Posts
On February 01 2022 07:17 bela.mervado wrote: isn't this what you are looking for? 444.75 8951 Maru,T,3185 Zest,P,1976 Solar,Z,1369 Scarlett,Z,1315 Zoun,P,1106 8924 Rogue,Z,2937 Dark,Z,2397 Heromarine,T,1573 Neeb,P,1265 Showtime,P,752 8944 Trap,P,2897 Cure,T,2153 Reynor,Z,1914 Bunny,T,1240 Lambo,Z,740 8948 Serral,Z,2670 Clem,T,2536 Time,T,1333 Special,T,1237 Dream,T,1172 Yes but I think the variance is 111.1875 not 444.75? | ||
|
bela.mervado
Hungary366 Posts
uhh this line looks ugly i keep it | ||
|
honorablemacroterran
188 Posts
On February 01 2022 07:29 bela.mervado wrote: yes this is just the sum of square of gr sum pts diff from avg uhh this line looks ugly i keep it May I ask what you are using to do your analysis? Would you mind sharing the source code? | ||
|
bela.mervado
Hungary366 Posts
of course, here you are a bit messy g++ wat.cc ./a.out | ||
|
Poopi
France12750 Posts
| ||
|
honorablemacroterran
188 Posts
On February 01 2022 08:12 Poopi wrote: Are these proposals using the rules in place ie. tiers / buckets of 4 players? But without random draw Yes it is solving for the minimum variance possible under the existing rules, without random draws. | ||
|
Poopi
France12750 Posts
| ||
|
honorablemacroterran
188 Posts
On February 01 2022 08:38 Poopi wrote: I think random draws is fine / can add hype, and too low variance would be boring if favorites had very good shot of advancing. But they should make their process public so people don’t think rigged I think that loading the most hype matches into the Group of Death is actually bad for overall viewership tbh and makes the other groups less hype. I think the groups posted by myself and Bela are legitimately more interesting and the outcomes are less certain across more of the groups. | ||
|
Vindicare605
United States16004 Posts
On February 01 2022 09:42 honorablemacroterran wrote: I think that loading the most hype matches into the Group of Death is actually bad for overall viewership tbh and makes the other groups less hype. I think the groups posted by myself and Bela are legitimately more interesting and the outcomes are less certain across more of the groups. It almost guarantees an underwhelming Ro4 or Ro8 because all of the good matches are decided early on in Bo3s in the group matches. BTW all of this mess is going to look way worse if the rumors about the Chinese players having major Visa issues is true because that means some other low ranked player will get bumped up to Group B in place of TIME and that will only make the disparity in that group vs the others even MORE apparent. | ||
|
Acrofales
Spain17735 Posts
On February 01 2022 07:17 bela.mervado wrote: isn't this what you are looking for? 444.75 8951 Maru,T,3185 Zest,P,1976 Solar,Z,1369 Scarlett,Z,1315 Zoun,P,1106 8924 Rogue,Z,2937 Dark,Z,2397 Heromarine,T,1573 Neeb,P,1265 Showtime,P,752 8944 Trap,P,2897 Cure,T,2153 Reynor,Z,1914 Bunny,T,1240 Lambo,Z,740 8948 Serral,Z,2670 Clem,T,2536 Time,T,1333 Special,T,1237 Dream,T,1172 How is this better? Group A: interesting fight for 2nd place between Zest, Solar and Zoun. Maru all but guaranteed through unless he somehow loses to Protoss. Group B: Rogue vs Dark for the billionth time this year. Heromarine with an outside chance of causing an upset. Group C: interesting. Group D: Serral has an even easier group. Sure, you trade one angstgegner for another, but the rest still can't touch him. Clem has to TvT a bunch, so he should be comfortably second, but either Dream or Time can surprise. Vs the actual groups: Group A: interesting fight for 2nd place between Solar, Cure and Dream. Rogue all but guaranteed unless he decided to leave his A-game at home. Group B: Serral and Dark should stomp this. Only question is who is #1. Group C: interesting. Group D: interesting fight for second place. If Maru has an off day it's even possible that he falls to Clem, Zoun and/or Reynor. Unlikely given current forms. Seems like a lot of effort for no significant gain. | ||
|
RKC
2847 Posts
Putting all the top players in the same group or bracket will definitely guarantee more big clashes. Whether it generates more hype is debatable because casual viewers are more drawn towards the advanced stages by default. Speaking for myself, I'm not going to purposely tune in a group match just because it's a group of death. I plan to watch the knockout rounds simply because it's most certainly guaranteed to have the top players on form clashing. But if it turns out (hypothetically) that the final four is Serral against three other underdogs (because Maru, Rogue, Dark, and Reynor got knocked out earlier due to bad bracket luck and stacked groups), I won't bother watching at all.* * Of course, I would watch if the underdog is truly on form and deserving of advancing by beating the top players (and not fluked through by luck of the draw). | ||
|
deacon.frost
Czech Republic12125 Posts
On January 31 2022 20:59 Gina wrote: They might not be using a nice-looking draw like bits of paper out of a hat, but just put the names and points into a program and get the groups. I know my friends running Go tournaments had software for that sort of thing. So there may not be anything to show on video. That would be what? 30 seconds video? So why not show it? I mean there's literally no reason to not show it. You're not competing for views or to be interesting, you're doing it so you stop these stupid conspiracy theories. | ||
|
Schelim
Austria11528 Posts
| ||
| ||
