Proposal Swiss-style Tournaments

[false]

After reading much of the wide-ranging discussion surrounding seeding, ways to increase quality of tournaments, and the fact that Tamahome has had absurd run of classic opponents in every cup so far I thought I would throw my hat in the ring with an idea for I've had in the back of my mind for a while now; one that I discussed for a short while with some of the Classiest guys. I wanted to open a discussion about potentially using a swiss-style structure (specifically "x-3 = eliminated") for some of our current tournaments (specifically Classic cups and Grand Slam opens), or potentially a new tournament altogether.

Likely every single person who is reading this thread has, at one time or another, suffered elimination from a tournament following either a sequence of poor luck, a difficult matchup, or even just playing below their capability. Much in the same light; a brand new player might excitedly look forward to a new experience in the tournament scene, only to end up, to borrow the words from Triangles "instead facing Finchinator using stall for 2 games and then wanting to quit Pokemon forever."
Of course there is the double elimination-style tournaments, which aim to better handle some of the issues caused by single elimination brackets by giving users a second try if they happen to fall. However, double elimination tournaments require too much time to be a viable option for tournaments with a great many signups. If recent Classic cups or Grand Slam opens used double elimination bracket, the tournament would take 4.5 months to complete.
The idea of seeding, in order to avoid the situation where one top player is knocked out by another early-on, was also floated around, but seemed to receive mixed interest at best. Surely there must be a better way!

A swiss-style tournament could run a 512-man tournament in a little over 2 months, the same amount of time our current single elimination opens and cups are slated to last (and around half the time of an identically-sized double elimination bracket). Early rounds could even be extended to account for increased load without increasing the total tournament by more than a few weeks at most. The obvious main benefit being that a swiss-style bracket does a far greater job at reducing variance and rewarding consistency, while also being more accessible to newcomers, giving them more chances to play. In the case of Classic cups or Grand Slam opens, if still having an outright 'winner' is desired, which i assume it would be, the winner could simply be the player who remains undefeated for the greatest amount of time.

I understand that changing to a swiss-style tournament wouldn't bring only positive changes, so I decided to make a list of some pros and cons that I could think of in order to give a little of my perspective, as well as list some potential issues with the suggested format.


Pros:

• Rewards consistency/reduces variance; Stronger players are less likely to be eliminated early. It also doesn't potentially hamstring a player's playoff chances in the case of a stroke of misfortune/bad matchup/subpar series etc.
• Final results would give a more accurate picture of the overall standings/performance of each player, which would also lead to more reliable/accurate seeding of playoff brackets
• More games played also leads to potentially greater exploration of the metagames in question, and also allows more people to play in tournaments for longer, making tournaments a little more inclusive. They aren't stuck watching the DPP Cup for 2 months if they get knocked out early.
• No requirement of additional time to complete rounds, as we already use a 1 round per week system
• Creates more interesting matchups; spectators would be able to see more of stronger players with top records facing off, because stronger players would remain in the tournament for longer periods of time, and would play more series
• Newer players get to play more series before being eliminated, giving them a greater number of chances to win, and more chances to gain experience. They would also be slightly more likely to be matched vs opponents who are around their skill level, potentially giving some less unbalanced matchups
• Swiss-style brackets are much more flexible, and can technically work with any number of players. This could enable the removal of byes/subs and instead allow everyone who signs up to play

Cons:

• More games being played could be a negative for players who do not have ample time to lend towards prepping and playing, especially for stronger players who could find themselves having to juggle a wide number of important games as later rounds progress in the case of a tournament like Grand Slam or Classic
• Hosting becomes slightly more difficult. Hosts would have to monitor a wider range of players for a longer amount of time, and would also have to make multiple brackets for each record group. It would also be ideal to ensure there are no repeat matchups where possible
• Potentially lowers the stakes of some series as they are no longer do-or die
• Would require a different points system than the current ones in use, as tournaments would not end with a 3rd-4th, 5th-8th, 9th-16th etc. structure - this would also mean, at least in the short term, that current seasons would be hard to compare to previous since the format/points awarded would be different

I'm sure there are a great deal of other positives and negatives that come with this format, these are just the ones I was able to come up with in this short time. I acknowledge that a swiss-style tournament is not a perfect solution to all problems, but I do genuinely think it does the best a comfortable midground between rewarding consistency/strong play and keeping tournaments short, while still being accessible and fair to both newcomers and experienced players alike. In my mind I feel like a large component of this discussion will come down to whether people are willing/able to fit a potentially larger number of games into their schedule for the sake of considerably greater tournament integrity, but I accept that there may be other flaws with a system such as this which I haven't noted. Regardless, I hope this post can serve to drive some sort of quality discussion, all I want for Smogon tournaments is for them to continue to get bigger and more competitive.

I also recognise there are a number of variations on ideas that could work to potentially limit the load of games such as only taking the top 4 results from a given run, or potentially giving earlier rounds extended deadlines, or even using x-2 = eliminated; but they all come with their problems that would need to be discussed separately. My main goal with this post is just to explore the idea of potentially using a swiss format in an attempt to make smogon tours even better.

Even if this is ultimately seen as too difficult of a task for tournament circuits such as Classic and Grand Slam, which have many concurrent series (although I know many people are able to fit in all their r1/2 series, I understand them not wanting to have to deal with that great of a time for a longer period of time), I'd love to see if its possible to have a discussion about adding this to another tournament, or perhaps something new utilizing a swiss-style bracket. Thanks for reading

 

[Stratos]

DOU ran a swiss style tournament twice in our 2014 circuit and activity issues were insane. Someone went 5-2 and cut while going 1-2 in sets that they actually played. You got dropped from the bracket after two straight activity losses (or one week of no activity at all iirc) and by the end the bracket was less than half the size it was at the start. I highly recommend against running a swiss tour with standard week long rounds. Which is a shame, because I agree it's a theoretically good format.

VGC has seen some success with single day swiss tours running x-2 cut so the format might have more promise for live tours.

 

[Shurtugal]

A little side note but I highly recommend using Battlefy to setup a Swiss format, which should help host monitor sets much more easily should this route be taken.

 

[DaWoblefet]

VGC has been doing Swiss-style tournaments at live tournaments for a very long time now, and any serious online tournament for VGC is going to be using Swiss as well (think Champions Cup, VGC 17 Smogon tour, Trainer Tower circuit, old Nugget Bridge circuit, etc). I think false laid out a really solid list of pros/cons, having been a player who's participated in Swiss-style VGC events for over 7 years now, but I wanted to offer a couple more points and emphasize a few of his pros/cons myself:

• Determining seeding in top cut is almost certainly going to be decided on by resistance, which is how well your opponents do throughout the tournament as well. If two x-2 players are compared, the player who had the "tougher schedule" will take the higher seed. This leads to some interesting implications:
  • If you make a tournament all x-2 cut (typically what is considered optimal in VGC), players have no reason to continue participating after hitting x-3 in a tournament outside of pride (for example, saying you went 6-3 in a 9 round Swiss tournament can be a good accomplishment to newer players). At VGC live events, you may as well get your money's worth out of the tournament imo, but online the motivation to continue playing is going to be low unless there's like circuit point incentives at x-3. This could lead to severe activity issues.
  • You can resolve x-3 activity issues by instantly dropping any player who becomes x-3 at any point in the tournament. Then if you're x-3, you're literally out of the tournament. The problem with that is 1) it affects seeding in top cut for players who are still in (a player who started x-2 and won out will have very, very low resistance), 2) it can introduce byes late in a tournament life cycle, and 3) it prevents a player who really does want to play serious tournament games against opponents of their relative skill level from doing so.
• One of the advantages of Swiss is also that you don't immediately know who you are playing against before the round goes up. Unless you hide the bracket in double elimination (which to me has always seemed sketchy), you can know instantly who you will fight in subsequent rounds once matches complete, which means in theory you can have significantly more time to prep for a game, either by knowing "I play this opponent for sure" or "I play either of these opponents, I can do some initial scouting on both and then have groundwork laid when one wins in the round". Obviously once you hit single elimination, the bracket is public, but I think it's fine to have that information late in the tournament. Swiss fixes that issue by putting everyone on an even information playing field regarding their opponent's identity at the start of the round.

Cons:

• Potentially lowers the stakes of some series as they are no longer do-or die

This will vary from tournament to tournament. For example, if you are projected to have a top 25 cut after Swiss (25 players will have x-2 records), then in the first round of top cut, seeds 8-25 will play for their chance in top 16, and seeds 1-7 will have a bye. That makes it so a player who starts 7-1 in a 9 round tournament still has motivation to play his set for a chance at the bye. Realistically, the final x-0 vs. x-0 of Swiss won't have a match that matters, but you still have the entire top cut to play out which is sufficient for high-stakes match material.

Pros:

• Swiss-style brackets are much more flexible, and can technically work with any number of players. This could enable the removal of byes/subs and instead allow everyone who signs up to play

This is technically true, but it is important to note that x-2 cuts are fairly large at smaller levels of entrants. So if you had, say, 54 players enter your tournament, 18 or 19 players make top cut, which is 33-35% of the field. Meanwhile, if you have a tournament with 400 entrants, you'll have 35 or 36 players make top cut, which is about 9% of the field (numbers based on http://swisstriangle.net/). 33% vs. 9% is a big difference, so just keep in mind that x-2 cut in smaller tournaments allows a much higher percentage of the field through. This might not necessarily be a bad thing, depending on what's desired. In most situations like that for smaller tournaments in VGC, it'd just be a top cut of 8 players, so x-2 performers can miss moving on depending on how well their opponents did. For Smogon purposes though, there's probably some % threshold you could work out to say "if a tournament reaches x number of players, use Swiss; otherwise, use double elimination".

Cons:

• Would require a different points system than the current ones in use, as tournaments would not end with a 3rd-4th, 5th-8th, 9th-16th etc. structure - this would also mean, at least in the short term, that current seasons would be hard to compare to previous since the format/points awarded would be different

Basing your seeding on record in a tournament should be fine, so long as you don't instantly drop players once they reach x-3. All of top 8 gets x points, all of top 16 or x-2 gets y points (whichever is greater), all of x-3 gets z points, etc. VGC scales its CP distribution based on attendance, so top 64 wouldn't get points at a Regional unless they hit 200 players. You could do a similar thing for Swiss tournaments on Smogon, you'd just need to come up with numbers that make sense for record-based point distribution.

Pros:

• Rewards consistency/reduces variance; Stronger players are less likely to be eliminated early. It also doesn't potentially hamstring a player's playoff chances in the case of a stroke of misfortune/bad matchup/subpar series etc.
• Final results would give a more accurate picture of the overall standings/performance of each player, which would also lead to more reliable/accurate seeding of playoff brackets

This is far and away the best reason to switch to Swiss imo. Without proper seeding in double elimination, your run can be influenced by bracket luck far more significantly than in Swiss. In Swiss, it is considerably more probable for you to play players of your skill level as matches progress. Additionally, we all know Pokemon is Pokemon - being able to take a loss due to luck increases the odds that, on average, the better player will be able to proceed throughout the tournament. I prefer Swiss-style tournaments in VGC to the seeded Double Elimination tournaments in other tiers like Doubles OU significantly because I think it's substantially more fair to the players involved.

 

[Bughouse]

Swiss is excellent for live events where everyone has committed to play, and terrible for events where losers drop out and their opponents just get a free win. You lose all the benefits of increased accuracy due to the amount activity wins.

Stratos highlighted the issue well, but I’d also point to, as a more recent and pertinent example, OLT playoffs where people have dropped out at 0-2 even though by rule some of the 0-2s will make it to 3-2 and advance.

I‘m a lifelong chess player who is very, very used to Swiss style tournaments. It’s what’s used for virtually all tournaments, whether live online or in person, they’re great. Most of the time the automated online tournaments can automatically repair players when dropouts occur. This would be a large burden for any host however and doesn’t square well with weekly rounds. It could feasibly work with a live tour with an automated, live Swiss bracketmaker that’s fully integrated into PS that automatically makes the subs for dropouts. Short of this, though, I think Swiss is a bad idea.

 

[Boat]

I'm a huge fan of Swiss for cups and opens.

"boat10/28/2019
Btw its rly funny how classic cups are a textbook example of when to use swiss
And we dont"

Swiss is best used when having more games is preferred. In events like Smash, Swiss is good for new players that usually pay to go 0-2. In cups and opens, Swiss could serve a much more practical purpose. Having more games played will dramatically improve the quality of the seeding. I'd like to address a few of OP's concerns.

"Hosting becomes slightly more difficult. Hosts would have to monitor a wider range of players for a longer amount of time, and would also have to make multiple brackets for each record group. It would also be ideal to ensure there are no repeat matchups where possible".

I'm not convinced that hosting becomes more difficult. Swiss isn't a very complicated format, and there's not a lot of room for error. However, the earlier rounds of the tournament would certainly require substantially more time from the host, and adding a second host to each cup could resolve that.

"Potentially lowers the stakes of some series as they are no longer do-or die"

While individual series are no longer do-or-die, each series will still be important. Just because you aren't eliminated for losing a series doesn't mean its not high stakes; we see this with double elimination pretty frequently. Dropping into the loser's bracket of a large DE bracket is a huge detriment. In addition, this kinda feels like it's the point, right? Lower the stakes of individual matches and have people play more matches to demand a higher level of consistency.

"Would require a different points system than the current ones in use, as tournaments would not end with a 3rd-4th, 5th-8th, 9th-16th etc. structure - this would also mean, at least in the short term, that current seasons would be hard to compare to previous since the format/points awarded would be different."

This could be resolved pretty easily by awarding points for specific records i.e you gain 15 points for going 9-0, 13 points for 8-1, etc.