Proposal Abolish Round Robin Finals in Non-Live Tournaments

[kaori]

Hello!

I come here with a proposal given the headaches presented by round robin finals in the current unofficial SM Cup (part of the ROA circuit), which can be found here.

As you can see, the 3 finalists are on their 4th run of finals, which was posted on April 18th. This stage of the tournament has not been resolved in well over a month, and while this isn't a failure of the hosts or players, it's a failure in how we currently make our brackets. Smogon Tour uses this system, and while it is far from perfect, the live tournament format supports a bracket reset way better than one where players have to carve out time again and again to play long, often Bo3 sets. This issue has come up in the past in the 3rd DPP Cup, featuring 3 rounds of finals. Unofficials are one thing, but the issue taking place in SM Cup is something that could happen to two onging cups in RBY and GSC, as we're headed for 3 finalists in each of those. Imagine having to delay Classic or Slam playoffs by weeks at a time because of bracket reset issues like this, it's frankly ridiculous and giving more byes in the initial rounds should be a non-issue in comparison to the headaches infinite resets can cause.

This topic was also covered in another thread found here in relation to double-elim tournaments, but I think this should absolutely be extended into single elimination tournaments as well. The burden on players is ridiculous and in the case of the SM Cup above, the players themselves barely want to finish themselves but have to out of fear for an infraction.

And to be clear since somebody will bring it up, I don't believe Smogon Tour needs to implement this change and round robin finals in that rarely cause as many issues as we've seen in non-live tournaments.

I'll also tag the current finalists in SM Cup to offer any thoughts they might have Aurella umbreon098 beatiful

 

[Aurella]

Something to point out is that none of us in unofficial SM Cup finals actually need the extra points from winning: reaching the final stage alone has guaranteed each of us enough points for ribbon playoffs at the end of the year, no matter how we perform during the rest of the SM circuit. The extra points only affect our seeding.

We've almost doubled the length of the whole tour from resets alone, if something like this were to happen in a cup/open where the points awarded to its winner could actually determine a sizeable portion of the playoffs roster and qualifier sequences it would be a nightmare. Just do us all a favour and do away with Round Robin formats, please.

 

[MrAldo]

Regardless of the need for points I want to support the idea of abolishing the concept of round robin finals in tournaments were a bigger playoffs picture is taken into account (Classic Playoffs, Slam Playoffs, you get the idea).

Applying byes to the closest magic number to avoid round robin finals is an ideal scenario tbh, to stop potential delays in case of constant bracket resets on a round robin finals (which are always a possibility), and this is something I always try to do in the few tours I host from time to time cause even for fun side tournaments it is just not a fun thing to happens to a finals. Prime examples can be seen above and I think it would be a good idea to just remove the concept as a whole. The magic number is done by the simple exponential function 2^n, n being a number bigger than zero so like:

2
4
8
16
32
64
128, and so on, or numbers not divisible by 3.

Now there is the issue about the amount of subs required on like 256+ signups tours tho, but we should discuss a potential alternative when the amount of subs is almost equal to the amount of signups in the previous cut-off if the round robin numbers are abolished.

 

[Diophantine]

Hold a preliminary round. If there are n=(2^k)+m players (where k is the greatest index such that we have 0 <= m < 2^k < n) then give n-2m players byes and let 2m players play each other. This way, m players get eliminated, so there are n-2m+m=n-m=2^k players in the next round, which ensures no byes or round robins for any future round as it's a power of 2.

For example, if n=69=64+5, then let 10 people play each other. 5 get eliminated, so in our next round we have 64.
If n=420=256+164, then let 164*2=328 people play each other. 164 get eliminated, so in our next round we have 256.

 

[teal6]

Okay I don't know why my post was deleted because it was sincere, but maybe there just wasn't enough words, so I'll write more to make it clear.

Diophantine's idea in the post above is clearly the route we should take as it involves very little change to our system and remains incredibly fair. It also gets rid of RR finals which are virtually universally hated. We should implement what Dio said above in order to improve the quality of our tournaments.

 

[Stratos]

Okay I don't know why my post was deleted because it was sincere, but maybe there just wasn't enough words, so I'll write more to make it clear.

Diophantine's idea in the post above is clearly the route we should take as it involves very little change to our system and remains incredibly fair. It also gets rid of RR finals which are virtually universally hated. We should implement what Dio said above in order to improve the quality of our tournaments.

I mean, that's literally just what byes are. "328 man play in round" or "92 byes" is the same shit. Either way I support though.

 

[Boat]

When I wrote the post about round robin in Double Elimination, I initially wanted to include Single Elimination, but eventually decided against it. Here's why.

When you increase the Max Bye %, you also increase the number of subs that can exist on the upper range of the previous bracket size. If you want to extend my proposal to Single Elimination, you have to consider bigger tournaments, namely OST.

Size 16 : 12 - 23
Size 32 : 24 - 47
Size 64 : 48 - 95
Size 128 : 96 - 191
Size 256 : 192 - 383
Size 512 : 384 - 767
Size 1024 : 768 - 1535
Size 2048 : 1536 - 3071

Double Elimination tournaments on this site are pretty small, and so you don't really need to consider bracket sizes larger than 256. No such guarantee exists for Single Elimination. At the signups deadline for OST XVI, there were 1550ish signups, which is pretty damn close to 1535. If these were the bracket ranges, then we would have been very close to a tournament with over 500 substitutes. That's 500 people that don't get to play in the tournament because they signed up in the last 33% of all signups. Unless we are confident that we get at least 1536 signups in every future OST, this could seriously damage a trophy tour. Personally, I would love to see Round Robin go but I wanted to make sure everyone was aware of the consequences of this.

 

[Wigglytuff]

At the signups deadline for OST XVI, there were 1550ish signups, which is pretty damn close to 1535. If these were the bracket ranges, then we would have been very close to a tournament with over 500 substitutes. That's 500 people that don't get to play in the tournament because they signed up in the last 33% of all signups. Unless we are confident that we get at least 1536 signups in every future OST, this could seriously damage a trophy tour.

The situation described is obviously a problem (500 subs yo), but it assumes that the TDs stick rigidly to signups count guidelines for OST. They haven't; an example being OST XVI, which was solidly a 1536 size tournament. TDs (rightfully) started with a 2048 size bracket with 300 more byes than would otherwise be permissible in the interest of not having a RR final and that was that. A flexible approach when it comes to OST is fine and I trust the TDs to use their best judgment for future OSTs.

I don't have anything to say w.r.t abolishing RR finals, though I would be hesitant to apply it generally. They should certainly be avoided for tournaments that don't exist in vacuums (Classic and Slam were two prime examples brought up), but it's not a big deal for most unofficial tournaments, as also mentioned in the OP.

 

[Perry]

I have had my fair share of experience with Classic and Slam, both as host and sheet guy. I agree with the sentiment echoed above that RR finals are not what we truly want for those tournaments, but I still have my share of doubts regarding changes to the "system" we've known for so long.

While Dio raises a good method, I can't help but wonder how that would actually work re: points. Some players would eventually play a "preliminary" round to advance into the actual cup and, as such, would it be fair if we did not count it as a win towards the overall sheet? would it be fair towards the other (huge) group of players that would not get this "chance" of getting an extra point? I'm aware of the WCoP qualifying round and how a parallel could be theoretically drawn between it and a possible "preliminary" for Classic and Slam (mainly how they don't count toward the "official tournament data" such as WCoPs sheet and the Tournament's overall one), but ultimately I feel that this argument doesn't really hold its weight due to WCoP intrinsic characteristics (team tour, relegation system, new teams, etc).

I do see potential with Dio's proposal, however. It could fit nicely with Eo Ut Mortus's proposed policy of signup deprioritization that was brought up some days ago and could end up killing any chances of infinite resets that could delay playoffs considerably! But, to be fair, throughout the three years I've done sheet for Classic (and potentially) Slam, this was never too much an issue... so yeah, there's that. I think I'd be more inclined to fully support it if there was a test run somewhere, but I think thats unlikely at this point

 

[Sapientia]

I don't like the idea of "preliminary rounds", at least we should not call it that way. I totally agree with what Perry said regarding points, especially if we have very little people playing "play ins". Despite points it also feels wrong to let people compete about competing in a tournament, that should be open for everyone. How do we want to decide who gets to play play ins and who has a bye? First come first serve? Random? Seeding? I can also see the first complaints if we got two top players facing each other in that stages.

I know that increasing the number of byes is basically the same, but wording matters. RR finals suck and increasing the number of byes is the only viable option we have, but please don't call it play ins

 

[Boat]

I'd like to bump this thread, since it really never reached a conclusion. I think being able to finally kill round robin on the site would be an objectively good change to tournaments. Just like the Double Elimination situation, I think extending the Max Bye % and adopting the following ranges, as described in the OP, would be good.

Size 16 : 12 - 23
Size 32 : 24 - 47
Size 64 : 48 - 95
Size 128 : 96 - 191
Size 256 : 192 - 383
Size 512 : 384 - 767
Size 1024 : 768 - 1535
Size 2048 : 1536 - 3071

There are basically only 11 tournaments that this is relevant to. 5 Opens, 5 Cups, and OST. Any other single elimination, non-seeded tournament is almost certainly smaller than CG OU Seasonal. For smaller tournaments like those, this problem has been reduced to one already solved, since the TDs obviously find it acceptable that tournaments of that size have these ranges.

The tricky part is tournaments of size 384 and higher. I'll address OST first; Wigglytuff made a post here responding to my initial concerns regarding OST that I think is pretty good. As we saw last year, there are clearly situations where ignoring the Max Bye % is what the community wants, and I think most people trust the TDs with the level of flexibility to make that decision.

Opens and Cups can take a similar approach, but with a bit more structure. Hosts should be more willing to close signups early. Using projections from previous editions of the tournament, you can get a pretty good idea of how many signups a tournament will have by Thursday or Friday. If you project that you will end slightly above a magic number, close early. If it's close to the next magic number, let signups stay open for longer. If it's in the middle, that's a judgement call but I would lean towards closing early. Some people might have issues with extending signups that I'll address. The potential "time loss" of extending signups is much much less than the potential "time loss" that a resetting round robin can have on the overall schedule of Classic / Grand Slam, especially if the tournament resetting is LC / PU / GSC / RBY.

As a side note, since this change would mean that both Single Elimination and Double Elimination no longer use the 3*2^N Size brackets, that means that the bracketmaker can be reprogrammed to represent this. When the double elimination change was implemented, it used a creative solution by Eo Ut Mortus to avoid reprogramming the bracketmaker. Since that would no longer be the case, there shouldn't be any concerns about this making hosting harder, and actually makes it easier for Double Elimination.

 

[Quantum Tesseract]

Going to add my support to this, such as it is. As a frequent live tour host, I try to do what I can to avoid 3 way finals because they can make it just stretch on and on, and as a player they're even less fun. For instance, I just had a 3 way finals that dragged on for well over an hour just on the final rounds, because we had to reset the bracket three times before it finally came down to a win. It's not like this is to the benefit of more high level games, either; by the time I'm playing round 4 of the finals my enthusiasm for each match has gone down a lot and my playing skill to an extent with it.