EPTThese observations are in general much closer to what we've seen from the 3 non-joke TIs, though TI1 would arguably have the same 100/20 rating adaptivity that motbob proposed due to the fact that 1/2 the teams had never played dota2 and were unable to actually set hotkeys
Winner's Advantage in Grand Finals - Page 2
| Blogs > motbob |
|
Kupon3ss
時の回廊10066 Posts
These observations are in general much closer to what we've seen from the 3 non-joke TIs, though TI1 would arguably have the same 100/20 rating adaptivity that motbob proposed due to the fact that 1/2 the teams had never played dota2 and were unable to actually set hotkeys | ||
motbob
United States12546 Posts
EDIT: Using Kupon's original suggestion of 2.5/1 for adaptation, the original result resurfaced: the "best team" as defined by the highest static ELO with the " good adapter" value won more often under the 1-0 format. So the result depends on the strength of the adaptation variable. | ||
|
Kupon3ss
時の回廊10066 Posts
Overall the differences are so minute and in Dota at least the full bo5 appeals to spectators more. Honestly given what we've seen in TI2/TI3/TI4 we've always seen a more stagnant team in the winner's bracket than the loser's bracket. | ||
|
Salazarz
Korea (South)2591 Posts
It's always going to be a compromise between 'fair' and 'viewable', and since we can't even honestly say that the team coming from the winners bracket of a tournament is always the 'better' team, to me 1-0 starts make little sense. | ||
|
Faruko
Chile34171 Posts
problem solved | ||
|
MetalMercury
United States1161 Posts
On March 14 2015 23:47 Salazarz wrote: There is literally no such thing as a 'best' team in anything, not just Dota, and any kind of a model created to find the best team is always going to be flawed. There is always the question of form, mental factors, preparation, and a ton of other variables that simply can't be quantified in a number - this is a big part of the reason as to why any kind of power ranks were always a hotbed of debate. That aside, tournaments are not, and never were a tool to find the best team and award them with the prize. Tournaments are made to provide a spectator experience, and a rather large majority of viewers seem to agree that 'clean slate' starts in the finals tend to provide more entertaining series. On top of that, literally any bracket format is already massively flawed at finding 'the best' team, and you could also argue that things like playing in front of a crowd, having to adhere to various live schedules, having to travel to the location etc etc also bring in external variables that may be counter-productive to the best team winning. It's always going to be a compromise between 'fair' and 'viewable', and since we can't even honestly say that the team coming from the winners bracket of a tournament is always the 'better' team, to me 1-0 starts make little sense. Someone earlier mentioned that the winner's team always stagnates more than the losers team, which is actually incorrect as far as I'm concerned. In International history, the winner's bracket champion has won the grand finals in EVERY SINGLE TOURNAMENT. This idea of lower bracket teams adapting more on the largest stage is bullshit. I find that people are largely terrible at judging athletic performance, and this is exactly why we need numbers. I'm actually really surprised at some of the comments in this thread chain; they show a little bit of a misunderstanding of how statistics function when it comes to athletic performance. Let's start with a bit of a basic premise, and work our way down from there. Are there factors that we cannot measure that impact team performance? YES. But the question after that becomes: if we can't measure it, why do we care about it? Even if it impacts performance, it's not predictable. If it were predictable, we could model it it. And the answer to that seems to never be satisfying from a statisticians point of view. However, just because something is intangible doesn't mean we can't measure it. Say a team always loses in certain situation: they can't win in front of crowds, they suck at traveling, whatever it might be. Then that gets accounted for into their ELO or their performance statistics when they actually lose the game. In fact, for all teams that have a sufficient amount of games played, these factors are ALWAYS accounted for in principle by mathematical ranking systems. I've heard people counter by saying that these don't happen enough to impact the mathematics much, which could be true. BUT, if that is true, then why do you care about it? If the mathematical ranking system doesn't have enough sample to judge it, you certainly don't have enough anecdotal evidence to judge it either. | ||
|
SKC
Brazil18828 Posts
| ||
|
MrCon
France29748 Posts
| ||
|
MetalMercury
United States1161 Posts
On March 15 2015 04:48 SKC wrote: The winner's bracket champion has not won the grand finals in every single TI. You're right I think, iG was the team that won from the losers bracket. My fault. | ||
|
SKC
Brazil18828 Posts
That's actually a good question, what would be the odds of the best team winning if the finals were completelly random? It's at least a decent baseline to know. It should depend a lot on the kind of teams you supply the system. If the teams that reach the finals were ussually evenly matched, it wouldn't even change that much. In the end the goal of a tournament is not strictly for the best team to win, so whathever feels fair is more important than whathever little advantage you can math it out of simulations. Fighting games have always felt true double elim is the best. A lot of people that watch Dota or SC2 would hate it. | ||
|
zdfgucker
China594 Posts
 Makes me ignore any results.Loser bracket teams play way more games, so it's a loss-loss situation. Start at 0-1 disadvantage or play (possibly) another game when you're burned out? Either way you're fucked, so I don't see any good argument speaking against WB advantage. | ||
|
nojok
France15845 Posts
I think most of people prefer bo5 no advantage so it's fine the way tournaments run it. | ||
|
Buckyman
1364 Posts
| ||
|
Badjas
Netherlands2038 Posts
The upper bracket final's winner (team A) should play against the lower bracket final's winner (team B) in a BO3 or BO5. If team B loses, they have lost twice and should be eliminated, team A is the winner. If team A loses, both team A and B have lost once in the playoffs. Now they are even. A BO1 or BO3 should be played to see who will lose a second time, so that a winner can be declared. (There are already three "finals" named. UB finals, LB finals, grand finals... this finals should have a proper name. And ultimate final isn't it I think.) If you support this concept but are against winner bracket's advantage... BO3 grand finals followed by BO1 "ultimate finals" is exactly the same as BO5 with 1 game winner's advantage, save for presentation (and a rest period that would be influential actually). | ||
|
eonrulz
United Kingdom225 Posts
| ||
|
Sr18
Netherlands1141 Posts
Offtopic: this doesn't change the fact that I really dislike watching tournaments with winner bracket advantages as it kills the grand finals hype for me. | ||
|
DucK-
Singapore11447 Posts
| ||
|
Bigtony
United States1606 Posts
On March 15 2015 16:21 Badjas wrote: Tournaments should get rid of winner's advantage. Because people complain about it. Alternatively they should play it out like this: The upper bracket final's winner (team A) should play against the lower bracket final's winner (team B) in a BO3 or BO5. If team B loses, they have lost twice and should be eliminated, team A is the winner. If team A loses, both team A and B have lost once in the playoffs. Now they are even. A BO1 or BO3 should be played to see who will lose a second time, so that a winner can be declared. (There are already three "finals" named. UB finals, LB finals, grand finals... this finals should have a proper name. And ultimate final isn't it I think.) If you support this concept but are against winner bracket's advantage... BO3 grand finals followed by BO1 "ultimate finals" is exactly the same as BO5 with 1 game winner's advantage, save for presentation (and a rest period that would be influential actually). This is the perfect solution but it doesn't work because of the unpredictable nature of how long the matches will take and how many matches you will have. Tournaments cannot afford to under-deliver or need an extra 3 hours for the final BO3. | ||
Elyvilon
United States13143 Posts
On March 15 2015 21:28 DucK- wrote: The fairest way to deal with Winner's Bracket advantage is to have the team from the Loser's Bracket win TWICE. After all, the Winner's Bracket team never ever had a second chance. I agree that is this is the fairest way, but that's not necessarily the same as being the best way. | ||
|
Badjas
Netherlands2038 Posts
On March 16 2015 02:18 Bigtony wrote: This is the perfect solution but it doesn't work because of the unpredictable nature of how long the matches will take and how many matches you will have. Tournaments cannot afford to under-deliver or need an extra 3 hours for the final BO3. That is indeed a problem. I do like the idea of having a proper break before the "ultimate final" as it will give the losing UB winner a chance to come up with a new strategy (they haven't lost before). If the Upper bracket winner wins undefeated, there will be no "ultimate final". Filling that space with a second/third placer match doesn't really work as the lower bracket finals already provided this. A show match is an option with various methods. Show matches can also be much more loosely scheduled... Will it be a good trade-in for the lost hype? (question purposefully left unanswered) | ||
| ||
