Point card format sounds fun as hell I must say.  I've amassed loads of random divisions in my head that i've been apart of / helped come up with over the years.  Here's one of my favs from the DDRStorm circuit circa...'04?

'Challenge' format.
No limit player cap

Players all compete per round doing a 'challenge' that steps away from standard score attack.  Top half of placers per round get points *Example 16 players = 8 for 1st down to 1 for 8th* with the bottom half recieving a strike.  3 strikes = eliminated from tournament.

Example of challenges
*MY SPOON IS TOO BIG* - Play song while balancing 2 ping pong balls on 2 spoons.  Dropping ping pong ball = - 25.00% score.  Highest score wins.
*Ass backwards* - Play song facing the bar while holding a mirror
*Deaf&Distracted* - Play song on 1x with machine volume off and misc other song blasting.  Mellow is suddenly hard to FA with Lil John or Beiber blasting your ears.
*I CAN'T READ!* - Play song on stealth with a copy of the stepchart in your hands.

...etc etc etc, there's loads of them and the various methods i've seen this run RARELY have the regular tourney winners also win it.

I posted one in the level capping thread, but I'll summarize the core idea here:

We set a core difficulty range (say, 9-11), and allow players to opt in - upon registration - to wider ranges (say 8-12 or 7-13), which extend equally from both caps.  Matches are played based on the narrower range claimed between the two players.

Taking a wider range allows players with a specialty outside the core to exploit it, but open themselves up on the other end of the spectrum as a result, so wanting to play 12's means you feel prepared to play 8's as well.

This still sets a reasonable bar for all entrants (say, competency in the 9-11 range), and offers this core range as a safe zone to players who may not be strong enough to compete across a wider range, but allows players open to (or wanting to) play songs outside that range to do so, provided they accept the risk.  As the tournament progresses, the core range could be widened, and any player using a smaller range will automatically take the new core range, forcing top players to be competent outside the the core range.

@Neko: We actually do these sometimes, and I try to always run them at Sakuracon! Last year, one of the challenges was to FA a chart of just left right jumps backwards while doing jumping jacks - so basically you had to sync totally by ear while also doing a proper jumping jack or something similar.  Harder than it looks.  Also, it was to Rick Astley.  Gerrak demonstrates below:

Sakura Con Dance Games Challenge tourney (Allan)

This two just sound like potential mindfucks that someone should make happen.

The printed stepchart has been done before, it ended badly.  Very badly.

How we did it was everyone watched the chart get played once, and then by seeding played it out.

I was horrible at it at first, but it was hilarious watching people that you wouldn't even consider tourney threats pull it off without much problem.

The following might be a little too abstract to explain in words alone; I'll try my best to make it clear, but I might have to post a diagram if enough people are confused.  I am assuming players have a reasonable understanding of how double elim. brackets and seeding work - if you aren't fully sure how they work in the first place, this will probably go right over your head, unfortunately.  I'm highlighting the parts that are most important - they generally do not require deep understanding of the bracket structure.

Triple elimination is exactly what the name indicates: adding a "Last Chance (LC) bracket" onto the bottom of a double elim. bracket to catch players bumped from the Loser's Bracket and implementing some resolution rules for the winners of the three brackets.

I mapped out brackets for 8- and 16-player brackets (there's one on Wikipedia somewhere, but it uses weird schemes for both the Loser's and LC brackets).  With 16 it's a contrived mess, but the 8-player LC bracket (8-LCB) can be cleanly attached to any larger double elim. bracket.

The 8-LCB that I mapped, assuming perfect seeding (that is, seed rank corresponds exactly to final rank - in double elim., this gives 1st-4th, with tied pairings for 5th/6th and 7th/8th), makes the following (very interesting) setup:

1. The 8-LCB starts with 8 vs. 5 and 7 vs. 6, then the respective winners play each other.  Perfect seeding makes 5 the winner of this playoff
2. The other two brackets play to completion (including the Winner-Loser round - with 2 matches required for an upset - that would normally be the final match), dropping 4, 3, and 2 into the LC bracket.
3. The 8-LCB survival chain: 5 plays 4, winner plays 3, winner plays 2, winner plays 1.  1 must be beaten twice to upset the championship (technically, 1 could have 0 losses at this point and should take 3 matches to truly lose, but that would be extremely tedious).

Back when I was in youth bowling leagues, we had an annual singles tournament with a similar survival chain to determine final rank, so I know that it can be an exciting twist to a more traditional format.  The full sequence for the tournament worked like this (everything was divided into 5 male divisions and 4 female divisions, so there were actually 9 champions who never played against each other when all was said and done):

1.  Getting the highest series within your league was the weekly qualifier for a monthly (Sept-May) qualifier tournament.
2.  Best 3-game series in the monthly tournament qualified for the annual tournament.
3.  At the annual tournament, all 9 players throw a 3-game series.  Players are immediately ranked and the bottom 4 drop.
4.  The "Roll-Off": 5 plays 4, winner plays 3, and so on, using single game scores to determine the winner.  There is a short warm up period (both players can throw 4 balls) between games, but otherwise, the games are played without breaks.

8-LCB yields the same effect as the Roll-Off, there's a final event for the top slice of players to climb to the top, but the further back you are, the more it becomes a brutal challenge to do so - as well as playing against the players who have already performed better than you under nearly identical conditions, the further back you start, the longer you need to sustain an undefeated streak to advance in the ranks.  Upsets during the survival chain can only knock a player back by one rank, but the potential to climb up (possibly all the way to 1st) is extremely exciting at the time of.

An 8-LCB has a "cost" of 7 or 8 additional matches (due to the multi-win upset rule in the final round).  For reference, with N players single elim. takes N-1 matches (for 7 for 8 players, 15 for 16 players, etc.) and double elim. takes 2N-1 (15 for 8 players, 31 for 16).  When Laura and I budget time for Sakuracon's ITG tournament, we usually figure about 10 minutes per match will buy more than enough time for matches, repairs, hunting down distracted players, and so on, so attaching the 8-LCB to an event can be assumed to use an extra 1 to 1.5 hours.

@Ancsik - hah, you just brought back memories.  Minus the winners finals winner Only having to lose twice instead of 3 times, you just described the standard format for eastern PA's tourney circuit back when they were active.  I was a fan of it until one particular tourney where I had a pair of pretty rough matches and in a 16 man format ended up in 'loser losers' by the 3rd round.  I battled back to the grand finals and had to beat flash 2/3 in 3 straight sets.  When all was said and done I think I had played like... 11 tourney rounds?  It's definitely not for one with any form of stamina issues and invites preying on people bring fatigued by a larger than normal amount of matches.


Also, it gives a horrendously large amount of downtime to those that win through it if run on one machine, which is likely the only reason I managed to win whatsoever.  One guy in grand finals will be battling his endurance due to more rounds, the other probably hasn't been on the machine in 90 minutes depending on how the bracket is ran through.

dear god this alone blows my mind

play and have fun

I was trying my best to think of a tournament format that would use teams of 3-5 people, but I can't seem to come up with anything that would be fair and still be enjoyable.

I pose a challenge to everyone to try and come up with a team tournament format. Preferably, I would like to see it allow players of all ranges to participate, from beginners to experts. So, how about it? Anyone want to show us what they've got?

One thought I had awhile ago regarding teams is to have a player pick a chart ("attack") and the other team elects their best suited player for that chart ("defend").  The defender now attacks and the first team defends.  All players attack and defend exactly twice.  There are some sequences leading to a defender who cannot attack afterward (having already attacked twice) - if this happens, the team elects an eligible player to attack.

A match would look something like this (I'm using teams of 3 to save some typing, it works the same 2, 4, or 5):

Team 1 is A, B, C - Team 2 is X, Y, Z
A -> Y (A attacks, Y defends)
Y -> B
B -> Y
Y -> C (Y has attacked and defended twice)
C -> X
X -> A
A -> Z
Z -> A (A has attacked and defended twice, so C is elected to attack)
C -> X
X -> B
B -> Z
Z -> C

As soon as one team reaches 7 wins, the match can end.  If it goes to a 6-6 tie, a random tiebreaker is chosen and each team elects a player to play it.

I think this format does a good job of generally balancing skill - teams can chose to send their best player as a defender to crush the other team's worst player, but then that player was wasted one of his chances to defend against the first team's better players.  It definitely encourages relatively balanced amounts of play for team members and lets players pick songs according to their own strengths, which should reduce the need to restrictions on what charts can be picked, meaning few toes will be stepped on when fleshing out the details - so long as teams are fairly balanced, somebody picking a 12 or 13 could be matched by an appropriate player on the other team; it bears mentioning that if the teams are stacked, any team format is going to fall apart.

As for the teams themselves, they could either be self-organized ahead of time (though there's a risk of stacking the best players in one team and we lack a ranking system that could define what "too good" means) or they could be seeded via qualifier.  If seeded by qualifier, we would look at the turnout, figure out "X teams of size Y" appropriately (4 of 4 for 16, 5 of 3 for 15) and randomly sort the top X players into different teams, then the next X, and so on.  If a team comes up one player short, the match format can be adjusted for them (allowing two players to play one extra song would do handle the problem).

Given the small number of teams likely involved, round robin is probably the way to go.  With 3 teams of 5, the worst case for a match is 21 songs, so the entire tournament would be the qualifier plus 63 songs, which would be comparable to the qualifier plus matches for 12 player double elim.  With 4 teams of 4, it jumps a bit (since round robin is quadratic with regard to the number of teams) and we get a worst case of 102 songs, which is closer to 16 player double elim.  A system other than round robin would have to be used for 5 or more teams.

I like what you've got so far. But I have some questions.

In your examples, the defending player ALWAYS initiates the next round's attack. Is this required, or could any member of the team initiate the next round?

I'm assuming all difficulty levels would be available for play in any match?

What do you think about mods (non-standard, like blinker, hidden, etc) being issued as part of the attacker or defenders strategy?

I had been debating a set rotation (either ABC-ABC or ABC-CBA) or the sequential version (like I posted).  Having the order completely open wouldn't be good because it would encourage prioritizing the best players as attackers.  I like the sequential version for simplicity, but a set rotation may have other benefits.

As I noted, unless full teams are overspecialized, team members should be able to pick up slack for one another, meaning that we could play with loosened caps or completely open selection rules and not have it ruin anyone's fun.

Mods are an interesting option, but most players never touch mods anymore, so it's hard to say what would be a fair way to allow them.  If mods were allowed, I could see a stockpile system working - for example, each team gets 3 chances to force a mod on the other team's player within a match (attacking or defending) and the decision to do so is made after a defender is selected.  It would allow mods be used in a more limited (but strategic) way, since people just aren't used to them.  Allowing mods would definitely require discussion beforehand.

I really like this idea. Reminds me of having "bombs" in a SHMUP! =]

Btw, what do you feel is the best way to balance out teams in regards to distributing players?

Team balancing is the trickiest part of all of this.  We originally came up with this format in the context of "Team America wants revenge on Team Canada," which made picking teams unnecessary - it was supposed to be a "however many Americans want to participate get together and challenge Canada to send us that many of their players" kind of thing.

I'd say the fairest way would be to seed and then distribute the players, though - so for 3 teams, seeds 1, 2, and 3 would be on different teams, and then... probably snaking around would be fairest, so team 3 would have 4, team 2 would have 5, and team 1 would have 6, and so on and so forth.

I had suggested dividing out teams by seed order, but with random assignments for each 3 players.  Snaking is probably a more balanced approach.  Either way, sandbagging will be a huge concern with team assignments.

[uses Phoenix Down on thread]

What if we were to fully utilize the vast number of custom songs added in the last year?  
Each player chooses a song [any song on the machine]. The winner is determined by whoever wins by the greatest difference in scores across both choices.

example:
The first song is Holy Orders. Player1 gets 80% and Player2 gets 40%
The second is Under Arrest. Player1 gets 20% and Player2 gets 90%.
Player2 wins.

Failing your song choice would result in a 0%, and a random tie-breaker would only occur if both players won by the exact same percentage. It would be a troll-tournament, filled with gimmicky picks or Dragonforce, but I think things would balance out, since less stamina-intensive players have a wide range of choices, and most people can't play 14s back-to-back. Victory would go to whoever studies or trains the most.

All it takes is a stamina player to chose a 12 or 13 against a non-stamina player and the whole match is done.

Example; Gerrak and I go up. He chooses a 12.
Results are:
G=90% S=70%

Second match, I choose a 9.

G=98%
S=99.8%

I am still ~20% behind and unless I choose something that I can quad and Gerrak breaks a leg or something, there's almost no way for me to pull ahead. That's the thing about accuracy - it takes a LOT of time and effort to get good at it, but it really doesn't push you leaps and bounds past the other players on most songs. It gives you an edge when playing against similarly skilled players on songs of similar difficulty, but it's not worth much in a set of 13, 6, 10 difficulty songs without the stamina to back it up.

I tip my hat to players like Keby, Gerrak and Tuan (I know there're others too, sorry if you weren't mentioned) who possess both awesome levels of stamina and accuracy. These are truly exceptional players. Unfortunately some of us (myself included) have been trying to improve our stamina for years and have made very little, if any progress.

I'm going to pitch my "pure random" tournament idea again, since I really liked the difficulty limits of Lucky Start and this would take the "being a good all-around player" idea to an extreme.

No difficulty ranges.  No set songlist.  Final destination.

Just what it sounds like.  Any song, any difficulty is in the range of the tournament.  The final match could come down to a 1, or Dumpstream could come up in the first match.

As for song selection, the DDR Magic way of doing it has still been my favorite all these years.  Cards of 5 songs each, and then each player eliminates one, playing the remaining three.  This would alleviate the ridiculous scenarios above somewhat, but still force players to be good at not just the high-level songs but the low to mid-level songs too.

Well hey, "unusual" is in the subject title.  What did you expect?  This would hardly be serious, and non-serious is just the way I like it.

i like this one, allows more freedom of choice and whatnot.

I also like that "DDR Magic way" method. We just need to turn off Items, so it's fair.  

Final Destination

eh, what do items have to do with it?

DDR Magic was a series of tournaments way back in the day.  Not to be confused with that mode on the Disney Rave version of DDR.