Tournament DOU Tournament Format Discussion

Smudge

NatDex Doubles TL
is a Site Content Manageris a Forum Moderatoris a Community Contributor
Breaking down how this would work out with last year's results:

6x Auto Invites:
ADV Cup: Lhions
DPP Cup: qsns
BW Cup: JRL
XY Cup: Z Strats (Banned)
SM Cup: Zoe (Declined Invite)
SS Cup: Papiloco

Doubles Classic (going to substitute the Vintage and Classic slots here for the 2 finalists to simulate. Alternatively, knock these out and take 2 more circuit points, which would be Spurrific&Meminger I think.)
JRL, Z Strats

Best DPL
Memoric (7-0)
Nails (6-1)

Best Derby Slots
GMA (8-0 ADV)
Tie for 2nd Farfromani/Croven (DPP 4-2)

List of qualified & not banned players so far:

Lhions
qsns
JRL
Papiloco
Memoric
Nails
GMA
Farfromani/Croven (we'll assume croven wins this, since farfromani qualifies off an future model of circuit points anyways)

for the rest of the slots, based on classic points, this leaves:
Yoda2798 Declined Invite
Actuarily
Toxigen Declined Invite
SingleThunder
Mishimono
Ann
MADARAAAA (Declined Invite)
zee
Akaru Kokuyo
Fran
zaaya, or farfromani if you consider DPP kickoff as a cup for the sake of it.

Comparison of lists:

Actuarily
Akaru Kokuyo
Ann
Croven
DaAwesomeDude1
Farfromani
Fran
Grandmas Cookin
JRL
Lhions
Memoric
Mishimono
Nails
papiloco
qsns
SingleThunder
zee

In our OldGens Invitationals, team tours were added to circuit points, rather than guaranteeing invite slots. I still prefer this method, however you can see that the list here is not significantly different proving that our system did in fact work to get a similar result. Our math was: Wins + (W-L) unless WL was negative. For example, GMA earned 16 points for an 8-0 score in Derby, but it wasn't necessarily a perfect system.

It could be argued that the Nails inclusion may be a little less representative of good performances across the year, given that dad1 had 2 strong performances across DPL and Derby, posting a 5-2 and 4-2 record + other circuit participation, vs 1 6-1 record and no other circuit participation.

The one caveat is that you can't account for the ban affecting this so much, so hopefully we don't see this happen again.

As for the rest:

1. I support zee's method for assigning points, as an improvement on the above.

2. I think 16 slots is pretty perfect, but I'm not really opposed to 24 w/ byes either.

3. If team tour auto qual slots are kept, you could give an extra one to DPL, rather than reducing Derby.
 

Amaranth

is a Site Content Manageris a Forum Moderatoris a Community Contributoris a Top Tiering Contributoris a Top Contributoris a Dedicated Tournament Hostis a Tournament Director Alumnusis a Social Media Contributor Alumnusis a Community Leader Alumnusis a Past SPL Champion
UPL Champion
Should the championship remain at 16 players? Or should it be expanded to 24 or 32?
personally I think 32 is too much and I think giving people byes is too big of an advantage. i think 16 is perfectly fine.

Should DPL & Derby have the same number of auto-qualification spots?
It could be argued that the Nails inclusion may be a little less representative of good performances across the year, given that dad1 had 2 strong performances across DPL and Derby, posting a 5-2 and 4-2 record + other circuit participation, vs 1 6-1 record and no other circuit participation.
I think since DPL has twice the oldgen slots as Derby, it makes sense to assign the 4 spots as:
Best two DPL records
Best Derby record
Best cumulative DPL+Derby performance other than the three above players

I feel like this should do a good job catching crazy single-tournament performances, with the fourth slot ensuring that if you go, like, x-2 both tournaments you probably make it (which I think is fair & deserved, if you're a high-end teamtour player in two separate oldgens you should be in there even if you didn't quite have the immaculate run either time)

How do we determine best Oldgens record?
i think the general fundamental question to answer here is: should 5-0 be considered better than 6-2, or worse, or the same?
you can design systems however you want to get what you need based on how you answer that question. all other comparisons are generally pretty intuitive to draw, but when there's a mismatch between "who has more wins" and "who has better win differential" it gets funky - that's the area our systems need to address
(win% is an awful stat for tons of reasons, let's not waste time considering it)

my personal feeling is that 5-0 should be considered equal to 6-2. an easy way to achieve this would be scoring players based on (wins + differential). 5-0 has five wins and five differential, so it scores 10. 6-2 has six wins and four differential, so it also scores 10.
this system is equivalent to "+2 for a win and -1 for a loss".
zee's proposal would score these 15 / 14 (assuming both players played every available game). which, like, that's fine if you believe that's the correct way to compare these records.
i dunno how i feel about the "-1 for not playing" thingy. i feel like it overcomplicates things, you are already punished for not playing more weeks because you're missing on a chance to get a win every time you skip a week. not having another +3 is punishment enough. i also don't get practically how it really helps with any specific scenario

I still prefer this method, however you can see that the list here is not significantly different proving that our system did in fact work to get a similar result.
this is by design. i mean, fundamentally, any two systems that are trying to determine a set of top 16 performers will end up with a similar list, if both systems are good.
i'm not 100% on what your full method last year was, but like I hinted at in my original post, the way circuits points are usually assigned means that the winners of the individuals will usually just have enough points to qualify anyway. the system I proposed is very deliberately quite similar in function to a traditional circuit, just reskinned to add a bit of artificial tension to the finals of the individual tournaments, and "limit the damage" of teamtour unfairness to only a few slots. if you don't believe that the closedness of team tours is unfair you of course won't really agree with this, but it felt like a necessary middle ground given the concerns of the TLs
 

Amaranth

is a Site Content Manageris a Forum Moderatoris a Community Contributoris a Top Tiering Contributoris a Top Contributoris a Dedicated Tournament Hostis a Tournament Director Alumnusis a Social Media Contributor Alumnusis a Community Leader Alumnusis a Past SPL Champion
UPL Champion
So the forum mod team has been talking about what changes we could make to the Oldgens circuit, and here’s a rough outline we’re proposing, and would like the community’s feedback on.

This structure would have guaranteed spots in the invitational for the following:
  • Winner of each oldgen Swiss (so a max of 6 spots taken)
  • Winner of each Classic & Vintage (max of 2 spots taken)
  • 2 (?) of the best records in oldgen tiers in each DPL and Derby (max of 4 spots taken) (ties being decided by a best-of-5, similar to how the tour is played).
  • The remaining spots filled by whoever has the most circuit points that doesn’t automatically qualify. (So a minimum of 4).
  • The tournament would still be seeded based on circuit points.
What I mean by max & minimum is that there may be users who meet the criteria to auto-qualify multiple times, and thus free up more spots for users to qualify via the circuit.

Other than the above outline (of which all of it is up for discussion), there’s also some key discussion points we’d like to hear feedback on:

1. How do we determine best Oldgens record?
Some options are winning percentage, most wins in the regular season, most wins overall, etc.
Winning percentage can be flawed at such small samples, for instance if someone takes one week off but goes 5-1 (83.3%) did they really do better than someone who played all through finals and went 7-2 (77.8%)?
Most wins would balance this, but is it fair that someone who’s team doesn’t make the playoffs gets punished, or should we reward winning in the postseason/tiebreakers, even if it means a worse overall record?

2. Should the championship remain at 16 players? Or should it be expanded to 24 or 32?
Expansion could offer much more meaning to seeds, like a 1 vs 32 would theoretically be a much more lopsided matchup, while still allowing much more players to participate. 24 could offer byes to the top 8 seeds.
If expanded, the number of auto-qualifications could also be expanded.

3. Should DPL & Derby have the same number of auto-qualification spots?
As Derby theoretically would only have half of the number of Oldgens slots (gens 3 & 4) compared to DPL (5-8), does it make sense to have the same number of auto-qualification slots?

Let us know your thoughts!
things have died down a bit but the circuit is underway so it would be good to have final word on this. and maybe announce it here when it's done. just a thought :)
 

zoe

Tragic Decision
is an official Team Rateris a Social Media Contributoris a Community Leaderis a Community Contributoris a Smogon Discord Contributoris a Tiering Contributoris a Contributor to Smogonis a Site Content Manager Alumnus
DOU & Discord Head
hihi we're back with an update on all this after all the prior discussion, thanks to everyone who shared their thoughts and sorry for the wait!

So after all the discourse before and after we stated our intentions to change the plan for the oldgens invites we have decided to do the following:

1. The invitational spots will be split up as follows:
  • 8 Slots for all the tournament winners (i.e swiss as well as Classic / Vintage winners)
  • 2 slots for best DPL record, 1 for Derby for 3 total slots
    • This will be decided by a points system, with 3 for each win and -2 for each loss (No punishment for not playing as sometimes circumstances can be entirely out of the player's control and players who don't slot every week into oldgens will likely not have enough points regardless)
    • DPL has a 2:1 ratio over Derby's oldgens slots, so it makes sense to represent that with how many qualify from each
    • Playoffs will be counted as well
  • Remaining slots and ones that don't get filled (i.e someone wins all swiss tours) will go to total circuit points, which will also likely be used to determine seeding
  • Total of 16 slots
2. Future swiss tournaments will go on for an extra week and require a 2-2 record to make cut, as opposed to a 1-2
  • Idea behind the old format was to allow everyone to play at least 3 sets per tournament but ended up allowing people to win a coinflip in order to cut, which is far from desirable (also it has been wildly unpopular).
    • To preserve the guarantee of everyone being able to play at least 3 sets but providing a safeguard against people getting an act win and cutting, a 2-2 record is now the minimum needed to make cut (at the cost of the swiss portion being 4 rounds instead of 3).
  • This will not apply to any ongoing tournaments but will apply to the rest of the swiss tournaments that have yet to begin.

That's all for now - if anyone has any further grievances with any changes (or lack thereof) feel free to air them here and i'll do my best to respond in a timely manner c:
 

Users Who Are Viewing This Thread (Users: 1, Guests: 0)

Top