August League Tournament Results and Format Discussion

BO3 with map veto leads to some nice thinking and strategy. I like this part a lot : when you've spent a month playing the same guys, you have an idea of the maps they like, and the map they struggle with, so it adds some extra interest in having this system for a league tourney. It would be very sad if this interesting part was removed or reserved for the winner bracket. As an underdog, knowing that you will only play BO1 on maps you don't like, and then spend the rest of the tourney watching the best players on twitch ... not very thrilling.
Moreover BO1 raise the problem of : how do you chose the map ?

It's clearly a matter of waht do we prefer : BO3 but we meet less players, or BO1 where we meet all players.

Maybe some hybrid combination is better : keep it swiss style but we start with few (2,3,4 ?) BO1, which serve as "placement matches" where we can meet anybody, and then we have ... let's say 2 BO3 against players with similar results. That way we keep the best of both worlds.

Initial pairing according to league results adds the possibility of meeting some much better/worse players in the first round. I liked this, because I would never play the top players otherwise, and the tourney would just be me facing my usual ladder opponents. That's why I suggested above that the first BO1 would be "random opponent", that way during the placement matches, we can have unusual matchups, and then for the more "serious part" with BO3s we have an opponent of our level.

Aurico, I wholeheartedly agree it is a big sacrifice, but over 6 hours for an event is just too much. We are running the risk of losing players to begin with so if we keep the system as it is you will actually see LESS people.

So what about the idea of mixing BO1 and BO3 ? (third paragraph)

Aurico’s suggestion is the best way to continue the format while maintaining the original purpose. Double elimination is just complete garbage for this as it continues to hold the same game delay problem while failing to categorize players for LotS points. The game delay is especially a problem with BO3 and BO1 mixed in the proposed fashion.

GSL format is just 4 player double elimination, same problem.

Double elimination also makes seeding significantly more important, the question of seeding by ladder or league points is pointless for Swiss. League points were used because it fits thematically.

@auricocorico said in August League Tournament Results and Format Discussion:

So what about the idea of mixing BO1 and BO3 ? (third paragraph)

Okay I guess I am not really sure:

  1. Challonge Bracket #1 for placement: swiss-style with 3-4 BO1 rounds
  2. Challonge Bracket #2 for top 4 placements, 2-3 rounds?

It does not seem like this would really cut down on time as we were averaging nearly 2 hrs per BO3 round.

Currently : 4 BO3 = from 8 to 12 games per player
2 BO3 + 3 BO1 = 7 to 9 games
2 BO3 + 2 BO1 = 6 to 8 games

It's a slight reduction of the number of games. Could be enough imo. Plus, note that we would skip a few veto phases.

I agree that this will slightly reduce the amount of games, and therefore time, but in challonge you cannot adjust the amount of games played per round midway through a Swiss style.

This leads to a need to have 2, separate brackets (the qualifier and playoff matches). The likelihood of having situations where tie breakers won’t work is very high….

I hosted a 2v2 round robin in the past, the inability to break ties was awful, led to 2 extra days of holding the event and everyone was miserable.

Is there a way we can deal with so little rounds not leading to ties?

Managing by hand the bracket is doable i think since it's only 8 players ( it's a little more of work for the TD, i know, but if rules are clearly stated before the tourney it's not very hard).
Ties are indeed more probable, since there are only 2 BO3 where you can use the difference between wins and losses as a tie breaker .. It's not a problem for the pairings of the second BO3 (we can use ladder rating to rank people with similar score and similar W-L ratio), but it can be a problem for the final results.

There are probably some ways to make this work, the only think really is to find an acceptable way to break ties ... but it's surely a bit more complicated than what we have now I agree.

@morax said in August League Tournament Results and Format Discussion:

in challonge you cannot adjust the amount of games played per round midway through a Swiss style.

For matches that are meant to be BO1 instead of BO3, can't you just "double report" them to challonge? As in: report 2 wins instead of 1 win.

That would make a BO1 match equally significant to winning a BO3 2-0.

Or, if it makes a difference, you could report it as win-loss-win, so it would be like winning a BO3 2-1.

Reporting 2-1 would be best to keep the tiebreaker automated and not giving undue influence to the BO1s I think.

@ftxcommando said in August League Tournament Results and Format Discussion:

Reporting 2-1 would be best to keep the tiebreaker automated and not giving undue influence to the BO1s I think.

Yeah, still kind of a cheat, though, since there was no actual games played. I am convinced Aurico is right in that we would have to use a manual system for this, and it kind of sucks since that won't be as well tracked for reference in Challonge's history.

Still feeling most people like the BO1 with more rounds that I spoke to. I will have to do wider conversations rather rely on folks to post replies here it seems.

Assuming they are uses as substitutes, a 1-0 and a 2-1 are identical in how they impact tiebreakers and succeeding matchups. I don’t understand why you think manual systems are needed at all.

Swiss tournaments are risky because of one reason: variability in game length. While lowering game quantity per series addresses it in one fashion, the bigger problem is that people can pick whatever map they want and TDs aren’t going to really tell them that they can’t play Ditch while everyone has been waiting for 30 minutes for their round to finish.

This extends to making every round BO1 but then adding more rounds, you aren’t really helping solve the problem as much as you think you are because it still carries more risks of people picking a huge macro map that everyone now needs to finish.

I would FIRST say that the issue to solve is to create consistency between rounds by maybe making round 1 a list of 5 of the smaller/quicker maps from the pool. Then round 2 is the middle 5. Then round 3 is the big 5. Then you cycle back.

Less freedom for players but likely leads to less downtime and if that is still not enough, begin decreasing BO3s. In theory 4 BO3s is totally possible to do in 5.5 hours as that is what happens in single elimination tournaments, it’s the downtime that is killing it.

Or if you want more consistency and have people play one another on various map sizes, have game 1 in a series on a small map, game 2 medium, game 3 big. Or some variation of it.

FREEDOM TO THE PLAYERS !!!

Yea again, that's just my opinion, go ask the others, but i'd rather wait a bit and have more freedom than being forced to play a certain type a map ... "you're against ZLO, oh darn, too bad for you it's time to only play 20x20 maps .. !"

Now splitting the problem in half is fine to me. First BO1s can be fixed map (round 1 : 10x10, round 2: 5x5 round 3: 20x20 for instance), and then we have the BO3 where we are more free. With 2 BO3, there is just one moment where we might wait for players to finish their Ditch game ... Some time is saved, w can use challonge, TD is happy, Ricorico is happy, everybody is happy.

Waiting for players to veto makes the FAFLive experience significantly worse. Players are third-class citizens fighting in the pit for my enjoyment, they don't deserve any free will.

Shouldn’t impact the casting experience since there are several games that could be getting watched during deadtime anyway. Kind of why you need 2 casters as one is primarily handling talking while another can also deal with secondary duties like finding the next best games to cover if deadtime shows up and interacting with the audience.

I've talked with Rowey about using MapBan and it seems very doable. If the casters were casting live, they would be fed the UI + the whole banning process.