D8 Deb8

[Dinglesberry]

Re: D8 Deb8

Yeah, but League and Overwatch don't have "official tournaments" that try to place everyone into their own bracket and tell everyone that they can win if they play hard enough. I'm sure if there was a similar thing for League, people would complain that a Dia IV team would stand no chance against a Dia I team.

if you think of each competitive season as a tournament then they pretty much do that, and yes people complain that the borders between divisions are gross and hard to climb from

edit: also i think ur idea is pretty interesting actually, starting at max division or w/e and progressively moving to a weaker pool of players/songs while still having the potential to be eliminated if you arent at that level... i stand corrected that its impossible to make it fair cuz ur idea is a pretty solid idea

[SK8R43]

Re: D8 Deb8

I agree, that's a really neat idea and I think it would work out fine for the most part.

[xXOpkillerXx]

Re: D8 Deb8

Goods points Renegade. A few interesting perspectives in your last few posts.

How many files would that require ? I believe that's:
nbFiles = [sum 2^(i - 1), i = 1 to n] = 2^n - 1
where n is the number of rounds. That means if we assume the end condition (1 player with no strike) happens on round 8, we'd need a total of 255 files.

Since we're splitting the group in 2 for each round, and if we use the number or participants so far, that means a final split of 1 person in the top 50% will happen at 275 * 0.5^8 ~= 1. That assumes no tie happens. So 8 rounds is most likely the minimum that would be needed, which means 255 files. Therefore we'd need to have a way to recycle files somehow.

Edit: compared to a 7 divisions 8 rounds tourney, which is 56 files, that means we'd need to be able to recycle about 80% of the files without making it redundant for players, because there's no way we can improve the content quantity by 500%.

[justin_ator]

Re: D8 Deb8

Goods points Renegade. A few interesting perspectives in your last few posts.

How many files would that require ? I believe that's:
nbFiles = [sum 2^(i - 1), i = 1 to n] = 2^n - 1
where n is the number of rounds. That means if we assume the end condition (1 player with no strike) happens on round 8, we'd need a total of 255 files.

Since we're splitting the group in 2 for each round, and if we use the number or participants so far, that means a final split of 1 person in the top 50% will happen at 275 * 0.5^8 ~= 1. That assumes no tie happens. So 8 rounds is most likely the minimum that would be needed, which means 255 files. Therefore we'd need to have a way to recycle files somehow.

Edit: compared to a 7 divisions 8 rounds tourney, which is 56 files, that means we'd need to be able to recycle about 80% of the files without making it redundant for players, because there's no way we can improve the content quantity by 500%.

I know it doesn't solve this completely but the length of the tournament and required files could be alleviated some if the split between strike/no strike was a little more strict. Say bottom 60% or so each round, for example.

[RenegadeLucien]

Re: D8 Deb8

How many files would that require ? I believe that's:
nbFiles = [sum 2^(i - 1), i = 1 to n] = 2^n - 1
where n is the number of rounds. That means if we assume the end condition (1 player with no strike) happens on round 8, we'd need a total of 255 files.

No, we actually wouldn't.

In the first round, we'd need one file, since there's one group.

In the second round, we'd need two files: one for the zero-strike group, and one for the one-strike group.

In the third round, we'd need three files: one for the zero-strike group, one for the one-strike group, and one for the two-strike group.

And so on.

So in an 8-round tournament, we'd only need 1+2+3+4+5+6+7+8 = 36 files. For a tournament the size of the current OT (275 total signups), we'd actually need 9 rounds to get down to one player, but that still means we'd only need 45 files. That's less than the current OT requires. (7*8 = 56 files.)

[xXOpkillerXx]

Re: D8 Deb8

I know it doesn't solve this completely but the length of the tournament and required files could be alleviated some if the split between strike/no strike was a little more strict. Say bottom 60% or so each round, for example.

Indeed a 40% upper bracket cut would lower the rounds to 6 with 63 files needed. Which honestly sounds very reasonable. Nice one there.

[xXOpkillerXx]

Re: D8 Deb8

No, we actually wouldn't.

In the first round, we'd need one file, since there's one group.

In the second round, we'd need two files: one for the zero-strike group, and one for the one-strike group.

In the third round, we'd need three files: one for the zero-strike group, one for the one-strike group, and one for the two-strike group.

And so on.

So in an 8-round tournament, we'd only need 1+2+3+4+5+6+7+8 = 36 files. For a tournament the size of the current OT (275 total signups), we'd actually need 9 rounds to get down to one player, but that still means we'd only need 45 files. That's less than the current OT requires. (7*8 = 56 files.)

Hmmm didnt see it like that. How would you compute the difficulties needed at each round ?

Also doesnt that system makes it so the upper half of R1 which ends on lower half of R2 will play against lower half R1 that ends in upper half R2 ? Idk if that's good or bad, just making sure that's what you had in mind.

[RenegadeLucien]

Re: D8 Deb8

Admittedly that would be a bit tricky, since in each successive round there would be more and more uncertainty in the skill rating of the participants of each group. But here's what I would do, using the current 275 player signups as a model:

In the first round, I'd take the median of everyone's skill rating. In this case, that's low-mid D5, so around 72-74. Then I'd add a few to that, to avoid seeing 50% AAA's. So the first round would be around 78.

In the second round, we'd expect the groups to contain roughly the top 50%, and the bottom 50%. The median of the top 50% of players is in the mid D6 range, which is about 87-89. This group would get a song of about 93. The median of the bottom 50% of players is in the high D3 range, so about 56-58. This group would get a song of about 62.

The third round we'd see groups of around top 25%, middle 50%, and bottom 25%. We'd take the expected medians again and add a few to each to get song difficulties of around 99, 78, and 40, respectively. Though by this point, that 40 might need to be adjusted upwards a point or two to account for the likely skillboosting of many of the D1-D2 players.

And so on.

And yeah, that's the idea. In Round 3, the middle group (1 strike) would contain both people who got the strike in the first round and survived the second round, and people who survived the first round and got the strike in the second round.

[xXOpkillerXx]

Re: D8 Deb8

Admittedly that would be a bit tricky, since in each successive round there would be more and more uncertainty in the skill rating of the participants of each group. But here's what I would do, using the current 275 player signups as a model:

In the first round, I'd take the median of everyone's skill rating. In this case, that's low-mid D5, so around 72-74. Then I'd add a few to that, to avoid seeing 50% AAA's. So the first round would be around 78.

In the second round, we'd expect the groups to contain roughly the top 50%, and the bottom 50%. The median of the top 50% of players is in the mid D6 range, which is about 87-89. This group would get a song of about 93. The median of the bottom 50% of players is in the high D3 range, so about 56-58. This group would get a song of about 62.

The third round we'd see groups of around top 25%, middle 50%, and bottom 25%. We'd take the expected medians again and add a few to each to get song difficulties of around 99, 78, and 40, respectively. Though by this point, that 40 might need to be adjusted upwards a point or two to account for the likely skillboosting of many of the D1-D2 players.

And so on.

And yeah, that's the idea. In Round 3, the middle group (1 strike) would contain both people who got the strike in the first round and survived the second round, and people who survived the first round and got the strike in the second round.

That means someone could play a 78 on R1, then a 99 on R2, then an 84-ish on R3, then an 90-ish on R4, etc no ?

Idk if I like that or not yet. Needs some more thought I think.

[RenegadeLucien]

Re: D8 Deb8

78 on R1, 93 on R2 (by winning the round), then back to 78 on R3 (by losing the 2nd round and gaining a strike), then up to around 87 on R4 (by winning the 3rd round).

If a player were to follow that exact path, they'd probably be a high D5 or low D6 player with a skill level of 80-84, so by Round 4 this is pretty close to the songs they'd be playing in the current OT anyway. Though I admit it is true that lots of players would be playing songs way above/below their skill level for the first few rounds.

[xXOpkillerXx]

Re: D8 Deb8

78 on R1, 93 on R2 (by winning the round), then back to 78 on R3 (by losing the 2nd round and gaining a strike), then up to around 87 on R4 (by winning the 3rd round).

If a player were to follow that exact path, they'd probably be a high D5 or low D6 player with a skill level of 80-84, so by Round 4 this is pretty close to the songs they'd be playing in the current OT anyway. Though I admit it is true that lots of players would be playing songs way above/below their skill level for the first few rounds.

Not only people will play many files outside their range, but the feeling of progression also dies and I feel like that isnt suitable for a tourney as is.

[devonin]

Re: D8 Deb8

All of these ideas amount to basically the same real conclusion:

The flaw isn't divisions, or the size or range of the divisions. It's that our flagship tournament can be entered by literally anybody.

Trying to come up with a system that would make it so a group of 5 buddies who play basketball at the gym on Wednesdays could compete in the NCAA tournament is just a losing effort.

In basically every other place where there is competition, there is a season, or qualifiers, or seeding that takes place ahead of time and weeds out all the inferior competition.

If we put everybody into divisions by skill rating, and then did a preliminary event purely for qualifying for the ACTUAL tournament, and then made the actual tournament only be between the top 8 in each division, it would be pointful.

As long as the format of our tournaments is "Here is a song, play it and get the best score you can" then 95% of people who enter the tournament have 0 chance to win it. It would probably be more productive and interesting for people to come up with ways to do user-run or staff-supported events at other times, using songs already in the game, under novel constraints to give the general population something to do and have fun with. The FFR TCG is a -fantastic- example of something that gives you recurring content to take part in that doesn't punish you for being low/mid in a division.