Discussion of Stage Legality in Smash Bros. Ultimate

[dav3yb]

If you're going to advocate for hazards on Pokemon stadium, you might as well advocate for stage morph. No 2 stages you pick will be any worse than 4/5ths the PS1 forms. Just because it's legal in melee doesn't mean it's a good stage, it has like, 1 good transformation, and 1 decent one. The rest are absolutely trash.

 

[Skeleghost]

Ok here is a weird thing that happens if you use lists to decide stage 1, and counter-picks to decide future stages... it would make sense for the pool of stages used in counter-picks to be SMALLER than the initial stage 1 pool used for lists. Like... maybe 30 or more stages used in the lists, but 20 or fewer used in counter-picks. That's a total reversal from the system used now.

Edit: Pokemon Stadium 1 with hazards on is a great example of a stage that I would say makes sense in a stage list (because people can just put it on the bottom if they don't like it) but might be better off banned from counter-picks.

 

[Ulevo]

So I took the blast zone information compiled by Ruben and created a chart:

Stage Blastzone Data:​

Stage | Height | Depth | Outer Width | Total Width | Stage Length
Final Destination | - | - | - | - | -
Kalos Pokemon League | 192 | 118 | 165 | 490 | 160
Unova Pokemon League | 190 | 142 | 155 | 474 | 164
Pokemon Stadium 2 | 180 | 125 | 157 | 500 | 186
Pokemon Stadium | 180 | 149 | 162 | 500 | 176
Battlefield | 192 | 140 | 160 | 480 | 160
Fountain Of Dreams | 192 | 140 | 163 | 480 | 154
Dreamland | 192 | 140 | 173 | 500 | 154
Mdigar | 192 | 154 | 164 | 500 | 172
Yoshi's Story: Melee | 180 | 117 | 159 | 454 | 136
Smashville | 190 | 115 | 160 | 459 | 139
Yoshi's Island: Brawl | 180 | 125 | 150.5 | 440 | 139
Town & City | 195 | 118 | 147.5 | 460 | 165
Lylat Cruise | 180 | 125 | 150 | 460 | 160
Warioware, Inc. | 210 | 95 | 120 | 360 | 120
Castle Siege | 180 | 115 | 144.5 | 420 | 131
Brinstar | 185 | - | 134/130 | 380 | 116
Halberd | 215 | - | 138 | 460 | 184
Skyloft | 192 | - | 139/149 | 460 | 172
Mushroomy Kingdom U | 192 | 108 | 142 | 500 | 216
Frigate Orpheon | - | - | - | - | -
Wuhu Island | - | - | - | - | -
Rainbow Cruise | - | - | - | - | -
Duck Hunt | - | - | - | - | -
Wily Castle | - | - | - | - | -
Umbra Clock Tower | - | - | - | - | -
Pictochat | - | - | - | - | -
Dracula's Castle | - | - | - | - | -

Some interesting takeaways:

- Ceilings are either in the 180 club or the 192 club with the exception of T&C, WarioWare and Halberd. Halberd ironically enough has the highest ceiling in the game for a potentially legal stage.
- The longest legal stage in the game is Mushroomy Kingdom U for Doubles, followed by Pokemon Stadium 2. The Stadiums in general have long floors.
- Mushroomy Kingdom U has reasonably normal blast zones, so it isn't too large there. However along with having the longest floor, it has close blast zones from the ledge, though not as close as WarioWare. Probably fine for doubles.
- While WarioWare has narrow blast zones it also has smallest stage floor with the exception of Brinstar.
- Nothing ever exceeds 500 units across from left to right blast zone.
- Not only are the blast zones narrow on WW but spiking people for kills is stronger.
- Out of all the triplats available, Yoshi's Story is by far the most distinct from Battlefield. It is the only one in the 180 club for ceiling height as opposed to 192, making it the Dreamland of this game. It is also the easiest to spike on, has the narrowest blast zones, and has the shortest floor. It is not as small as Melee's Yoshi's Story, but it is the smallest variant here.

 

[Luigifan18]

So I took the blast zone information compiled by Ruben and created a chart:

Stage Blastzone Data:​

Stage | Height | Depth | Outer Width | Total Width | Stage Length
Final Destination | - | - | - | - | -
Kalos Pokemon League | 192 | 118 | 165 | 490 | 160
Unova Pokemon League | 190 | 142 | 155 | 474 | 164
Pokemon Stadium 2 | 180 | 125 | 157 | 500 | 186
Pokemon Stadium | 180 | 149 | 162 | 500 | 176
Battlefield | 192 | 140 | 160 | 480 | 160
Fountain Of Dreams | 192 | 140 | 163 | 480 | 154
Dreamland | 192 | 140 | 173 | 500 | 154
Mdigar | 192 | 154 | 164 | 500 | 172
Yoshi's Story: Melee | 180 | 117 | 159 | 454 | 136
Smashville | 190 | 115 | 160 | 459 | 139
Yoshi's Island: Brawl | 180 | 125 | 150.5 | 440 | 139
Town & City | 195 | 118 | 147.5 | 460 | 165
Lylat Cruise | 180 | 125 | 150 | 460 | 160
Warioware, Inc. | 210 | 95 | 120 | 360 | 120
Castle Siege | 180 | 115 | 144.5 | 420 | 131
Brinstar | 185 | - | 134/130 | 380 | 116
Halberd | 215 | - | 138 | 460 | 184
Skyloft | 192 | - | 139/149 | 460 | 172
Mushroomy Kingdom U | 192 | 108 | 142 | 500 | 216
Frigate Orpheon | - | - | - | - | -
Wuhu Island | - | - | - | - | -
Rainbow Cruise | - | - | - | - | -
Duck Hunt | - | - | - | - | -
Wily Castle | - | - | - | - | -
Umbra Clock Tower | - | - | - | - | -
Pictochat | - | - | - | - | -
Dracula's Castle | - | - | - | - | -

Some interesting takeaways:

- Ceilings are either in the 180 club or the 192 club with the exception of T&C, WarioWare and Halberd. Halberd ironically enough has the highest ceiling in the game for a potentially legal stage.
- The longest legal stage in the game is Mushroomy Kingdom U for Doubles, followed by Pokemon Stadium 2. The Stadiums in general have long floors.
- Mushroomy Kingdom U has reasonably normal blast zones, so it isn't too large there. However along with having the longest floor, it has close blast zones from the ledge, though not as close as WarioWare. Probably fine for doubles.
- While WarioWare has narrow blast zones it also has smallest stage floor with the exception of Brinstar.
- Nothing ever exceeds 500 units across from left to right blast zone.
- Not only are the blast zones narrow on WW but spiking people for kills is stronger.
- Out of all the triplats available, Yoshi's Story is by far the most distinct from Battlefield. It is the only one in the 180 club for ceiling height as opposed to 192, making it the Dreamland of this game. It is also the easiest to spike on, has the narrowest blast zones, and has the shortest floor. It is not as small as Melee's Yoshi's Story, but it is the smallest variant here.

When you say that no stage is more than 500 units wide, is this including gigantic stages like Temple, New Pork City, Great Cave Offensive, and Palutena's Temple? I know these stages are absolutely not tournament-legal, but if they had the same blast zone dimensions as the tournament-legal stages, I'm not sure if the stages themselves would fit within the boundaries!

 

[Untouch]

Brinstar having a really small ceiling is actually a surprise to me, always felt like it was large.
I think it should be locked as a counterpick if sharking isn't viable (which it doesn't look like it will but whatever).

Some more stage data:
Wuhu's length is 136 w/o the lips and 200 w/ the lips. So actually smaller than MKU.
Gamer's length is 182, bit larger than average.
Frigate's length 109 w/o the platform at the end, no data on it with the platform.
Pictochat's length is 160, making it identical in length to BF.
Big Battlefield's length is 260, yeah.
Arena Ferox's length is 180.
Pokemon Stadium 1 and 2's platforms are identical. Unova's are very slightly larger.
Green Greens' full length is 210.

No FD data yet, I've heard the size is the same as BF but I don't have evidence.

 

[Ulevo]

When you say that no stage is more than 500 units wide, is this including gigantic stages like Temple, New Pork City, Great Cave Offensive, and Palutena's Temple? I know these stages are absolutely not tournament-legal, but if they had the same blast zone dimensions as the tournament-legal stages, I'm not sure if the stages themselves would fit within the boundaries!

No. This is just involving potentially legal stages.

 

[Frihetsanka]

Brinstar having a really small ceiling is actually a surprise to me, always felt like it was large.
I think it should be locked as a counterpick if sharking isn't viable (which it doesn't look like it will but whatever).

The ceiling is actually slightly above "standard" (with "standard" being 180), but the sides are really small, just slightly larger than WarioWare. I wouldn't be too optimistic about Brinstar if I were you (people seem skeptical).

 

[Veggi]

If you're going to advocate for hazards on Pokemon stadium, you might as well advocate for stage morph. No 2 stages you pick will be any worse than 4/5ths the PS1 forms. Just because it's legal in melee doesn't mean it's a good stage, it has like, 1 good transformation, and 1 decent one. The rest are absolutely trash.

I think you're making the mistake of believing that desirability, randomness, and fairness are all the same word.

You are making the argument that because you believe the transformations are undesirable it is comparable to stage morph which is a problem because it is assigns one player an advantage in game one.

In the case of PS1 hazards, the players agree that the stage is right for them and play on it for game one or it becomes fair because the player who does not like PS1 hazards can ban it or receive a similar advantage on their turn to CP.

When PS1 hazards is legal, the players decide if it is fair. When stage morph is legal, players are forced into a situation that is always unfair. So they actually aren't comparable at all.

 

[Midoriki]

Regarding the CP process, I don't think it's a good idea to extend the lists past the first match. It puts much more emphasis on the lists themselves, which means players are going to spend a lot more time thinking about them. It also removes a little bit of excitement since the stages are all known from the start and the players simply have to go through them one-by-one.

One thing that needs to be considered with using ordered lists for the first match and another method for the rest of the matches is how much information the ordered list gives players about their opponent's stage preferences, since that can be used against them in the next stage picks.

If players are looking at each other's ordered lists to pick the first stage, they have complete information of which stages their opponents don't want to play on.
- This can be mitigated a bit if the list is only N/2 + 1 long, but your opponent will still always know your least favorite stages, just not how much you dislike them relative to each other.
- This makes counter-picks even more lopsided, and makes pXp1 and X-2-1 almost as bad since each player knows X stages their opponent would rather not play on.
- If players can look at each other's lists, I think either only the lists should be used to pick stages, or players should be given a large number of bans.

If TOs are picking the stage then ordered lists actually give players less information than striking does. You can still infer stages that your opponent would want to play on less than the first stage from the stages above it on your list, but that's balanced a bit by the fact that you get more information for getting a less preferred stage in the first match. In comparison, striking always gives players N/2 -1 stages your opponent would want to play on less than the first stage.
- Counter picking with bans is still an option since its a little less potent than it is with striking, but keeping track of a potentially large number of bans might be cumbersome and require an app or pen and paper to go smoothly.
- pXp1 and X-2-1 would both work very well! And since the TO handles the match 1 stage picking, players wouldn't need to worry about apps or pen and paper at all after submitting their initial list when they register.

I do agree that we should avoid having too much emphasis on lists. I don't want there to be a "list meta" where people try to game the system by putting commonly disliked stages at the top of their list so that their actual favored stages are more likely to be the first to appear on both lists, or intentionally putting some of their favored stages lower on their list to trick their opponent into selecting them in future matches.

I also think that having at least some element of counter-play so that the lists don't dictate everything if you realize that stages you normally prefer are less good in a specific match-up is probably a good thing, especially since similar stages will likely be ranked closely together on a list. It would be incredibly frustrating to realize during match 1 that a match-up is bad for you on stages that have low ceilings, and then have all of your matches be on stages that have low ceilings because the stages have already been decided by the order of the lists.

 

[ParanoidDrone]

A brief, possibly obvious realization.

In computer science, the concept of "big-O" notation generally refers to "time complexity", or the time required for an algorithm to completely execute over a set of n elements. For example, a "bubble sort" algorithm has a complexity of n^2, whereas a "merge sort" algorithm has a complexity of n * log(n). (These are on average. Best and worst case scenarios may differ.) As n increases, n^2 grows faster (a lot faster, in fact) than n * log(n), which means bubble sort takes much more time than merge sort (again, on average) to sort the same list.

We can probably apply similar principles to theorized stage selection algorithms. I'm far too rusty to sit down and calculate time complexities myself, but one thing I do recall is that iterating over each element in a list is slow, splitting the list into pieces and working on the pieces is faster, and bypassing the whole idea of iteration and working with some constant number of elements is the fastest. We already see shades of this with traditional striking being considered unwieldy at anything over 9 stages, novel methods such as FLiPS accepting a higher initial cost in favor of working with a subset of the list for faster ad-hoc selection, and counterpick procedures such as "loser picks X winner picks 1" requiring next to no additional time investment at all.

Food for thought.

 

[J0eyboi]

A brief, possibly obvious realization.

In computer science, the concept of "big-O" notation generally refers to "time complexity", or the time required for an algorithm to completely execute over a set of n elements. For example, a "bubble sort" algorithm has a complexity of n^2, whereas a "merge sort" algorithm has a complexity of n * log(n). (These are on average. Best and worst case scenarios may differ.) As n increases, n^2 grows faster (a lot faster, in fact) than n * log(n), which means bubble sort takes much more time than merge sort (again, on average) to sort the same list.

We can probably apply similar principles to theorized stage selection algorithms. I'm far too rusty to sit down and calculate time complexities myself, but one thing I do recall is that iterating over each element in a list is slow, splitting the list into pieces and working on the pieces is faster, and bypassing the whole idea of iteration and working with some constant number of elements is the fastest. We already see shades of this with traditional striking being considered unwieldy at anything over 9 stages, novel methods such as FLiPS accepting a higher initial cost in favor of working with a subset of the list for faster ad-hoc selection, and counterpick procedures such as "loser picks X winner picks 1" requiring next to no additional time investment at all.

Food for thought.

The real question is if we can find an algorithm to ban every counterpick in worst-case polynomial time

In all seriousness, Big O notation is weird to use in this situation, because multitasking exists. For example, the list system skeleghost(?) suggested has a worst-case at least as bad as striking if we use Big O, but it still saves time because the lists can be adjusted while other people are playing their matches, whereas striking always happens at the setup (side note: an app could absolutely be made that allows for striking to occur before both players sit down, assuming both have phones, which would solve the issue of striking being too slow).

 

[Akiak]

One thing that needs to be considered with using ordered lists for the first match and another method for the rest of the matches is how much information the ordered list gives players about their opponent's stage preferences, since that can be used against them in the next stage picks.

If players are looking at each other's ordered lists to pick the first stage, they have complete information of which stages their opponents don't want to play on.
- This can be mitigated a bit if the list is only N/2 + 1 long, but your opponent will still always know your least favorite stages, just not how much you dislike them relative to each other.
- This makes counter-picks even more lopsided, and makes pXp1 and X-2-1 almost as bad since each player knows X stages their opponent would rather not play on.
- If players can look at each other's lists, I think either only the lists should be used to pick stages, or players should be given a large number of bans.

If TOs are picking the stage then ordered lists actually give players less information than striking does. You can still infer stages that your opponent would want to play on less than the first stage from the stages above it on your list, but that's balanced a bit by the fact that you get more information for getting a less preferred stage in the first match. In comparison, striking always gives players N/2 -1 stages your opponent would want to play on less than the first stage.
- Counter picking with bans is still an option since its a little less potent than it is with striking, but keeping track of a potentially large number of bans might be cumbersome and require an app or pen and paper to go smoothly.
- pXp1 and X-2-1 would both work very well! And since the TO handles the match 1 stage picking, players wouldn't need to worry about apps or pen and paper at all after submitting their initial list when they register.

I do agree that we should avoid having too much emphasis on lists. I don't want there to be a "list meta" where people try to game the system by putting commonly disliked stages at the top of their list so that their actual favored stages are more likely to be the first to appear on both lists, or intentionally putting some of their favored stages lower on their list to trick their opponent into selecting them in future matches.

I also think that having at least some element of counter-play so that the lists don't dictate everything if you realize that stages you normally prefer are less good in a specific match-up is probably a good thing, especially since similar stages will likely be ranked closely together on a list. It would be incredibly frustrating to realize during match 1 that a match-up is bad for you on stages that have low ceilings, and then have all of your matches be on stages that have low ceilings because the stages have already been decided by the order of the lists.

Yeah. Another solution would be to use an app which figured out the stage for you, without revealing the lists. This means either using only one phone, which is passed from one player to the other to rank the stages, or, ideally, an app which connects with another phone and compares the two lists wirelessly (that way both players can prepare their lists in advance). Or both options could be available.

I suppose the only counterargument to the information issue is that maybe it's not that big of a deal? Most of the picks are going to be somewhat obvious if you know their character, especially if it's just N/2+1.

I personally think an app is the better option here. I'm not really on board with the TO involvement thing, since it removes the option to modify your list at any given time before a set. And in those instances where an app isn't available, players can simply accept that they're going to give away a little bit of information.

Edit: the 'list meta' problem is definitely there too, but again would only happen when both players don't have access to the app, which would hopefully not be very often.

 

[Midoriki]

I'm far too rusty to sit down and calculate time complexities myself

Big-O notation only really cares about asymptotic complexity, so you basically need to treat numbers as if they were 0, 1, or infinity. This is fine for stages and sets since we have so many of them we can kinda pretend they are infinite and still get something sorta useful. But with the number of matches its a little weird, treating them as 1 hides repeated costs that are actually a big deal, but treating them as infinite hides once per set costs completely.

In an attempt to deal with this and still calculate the time complexities I've split up complexities for picking the 1st stage, and complexities for picking stage 2 and onward. This is probably more useful as an example of why you should only use Big-O notation when your algorithm runs on really large inputs, but it was kinda fun

Methods for picking stage 1 ordered by Big-O time complexity on N the number of stages:

Ordered lists compared by TO ahead of time: O(0)
- stage selection is already done. The setup time before hand is pretty huge though.

Random: O(1)
- real fast! if only it was fair x)

Random with bans (e.g. FLiPS): O(N)
- linear setup for the first stage.

Static ordered lists compared before each set: O(N)
- If the lists are made before hand their creation time doesn't affect the time complexity here. The algorithm for selecting the stage has linear complexity in the worst case scenario.

Striking: O(N)
- plain old striking is also linear.

Ordered lists that can change between sets: O(N log(N))
- letting the players re-sort their lists between sets could take a while in the worst case scenario, although irl it would probably only be a few minor changes and be during player downtime anyways.


Methods for picking stage 2 and on ordered by Big-O time complexity on N the number of stages:

Ordered lists for all stages at once: O(0)
- if all the stages are picked at once using an ordered list, stage selection is already done at this point.

Random: O(1)
- still real fast! still not fair!

Random with bans: O(1)
- makes setup take O(N) time, but just as fast as pure random afterwards, and somewhat more fair.

pXp1: O(1)
- Picking a constant number of stages and then picking one stage off that list is constant time.

pXp2p1: O(1)
- same as pXp1, just with another constant time step.

Counter-picking with bans per set (e.g. FLiPS): O(1)
- makes setup take O(N) time, but has constant time for every match afterwards.

Counter-picking with bans per match: O(N)
- this would technically be O(1) if you only ban a constant number of stages no mater how long the stage list is, but to be actually fair the number of bans should grow linearly with the number of stages.

Ordered lists that can change between matches: O(N log(N))
- letting players change ordered lists between matches would be a very bad idea xD

EDIT: terminology fixes

 

[Untouch]

No frame drops anymore on Fountain of Dreams it seems... Tested with 5 Iceclimbers with items on.

— MrConCon (@mrconcon321) December 5, 2018

FoD is apparently saved.

 

[Frihetsanka]

No frame drops anymore on Fountain of Dreams it seems... Tested with 5 Iceclimbers with items on.

— MrConCon (@mrconcon321) December 5, 2018

FoD is apparently saved.

Fun fact: The frame drops apparently only appear when playing 1v1 (or maybe up to 4 players), so this doesn't prove anything. Should be easy to test 1v1 though.

 

[Skeleghost]

Guys, just so we're clear, I already created an app for stage list selection here: http://smashfilters.droppages.com/
I know it's rough around the edges but the basic functionality is all there. You can load player's lists, as well as rulesets, and do the calculations, all basically instantly as long as you have the lists beforehand. You can do all of that on a phone. You can only create player's lists on a computer right now because I haven't added the right touchscreen controls, but I'm planning to add that soon!
I'll probably rename it to 'smash lists' or something though, because that's what everyone has ended up calling it.

Akiak's suggestion of having a version of the app that hides the lists is something I was already planning. Creating a version that can transfer lists over wi-fi is also a good idea, although that's more work. It is doable though, and I'll plan to add that feature too if people are legitimately interested in giving it a try (which it sounds like at least a few people are so far?).

I also might eventually try to make the interface look a little prettier, but in my mind that's a lower priority in comparison to getting something that does the job quickly and well. I do hope that it's at least easy to understand. If the interface is so ugly that people can't even use it, that's not very useful at all.

 

[Akiak]

Guys, just so we're clear, I already created an app for stage list selection here: http://smashfilters.droppages.com/
I know it's rough around the edges but the basic functionality is all there. You can load player's lists, as well as rulesets, and do the calculations, all basically instantly as long as you have the lists beforehand. You can do all of that on a phone. You can only create player's lists on a computer right now because I haven't added the right touchscreen controls, but I'm planning to add that soon!
I'll probably rename it to 'smash lists' or something though, because that's what everyone has ended up calling it.

Akiak's suggestion of having a version of the app that hides the lists is something I was already planning. Creating a version that can transfer lists over wi-fi is also a good idea, although that's more work. It is doable though, and I'll plan to add that feature too if people are legitimately interested in giving it a try (which it sounds like at least a few people are so far?).

I also might eventually try to make the interface look a little prettier, but in my mind that's a lower priority in comparison to getting something that does the job quickly and well. I do hope that it's at least easy to understand. If the interface is so ugly that people can't even use it, that's not very useful at all.

My only advice atm would be to work towards having images of the stages instead of just text. Something like this website for striking: https://bans.page/

 

[Zenos]

As someone with no experience playing on Yoshi's Island Brawl competitively all that much, can someone explain why the slants at the ledges are such a big issue for some? I've observed this exact same sentiment multiple times now from different players.

 

[DarthLuigi36]

I've finished up work on a huge study on stage selection in Smash 4, covering over 1300 games from 2018. While this is about Smash 4, it was created with intent to use the data moving on to Ultimate. Here's the original thread: https://www.reddit.com/r/smashbros/comments/a3dc5g/i_spent_the_last_two_months_working_on_a_massive

Obviously Ultimate is a unique game, but we should be able to look at Smash 4 info as helpful.

The biggest takeaway is that choosing the stage offers no statistical advantage. This can be viewed a few ways. Possibly it is explained by the fact that the stage chooser just lost previously, so it is logical they could lose again. Personally, I think it is simply the fact that character matchups are massively more important than stages in Smash.

Take a look and see the more in depth stuff if you're interested!

 

[Untouch]

lol Smashville getting picked 40% of the time.
I really don't want this again, this is why I want stage variety, this killed spectating Smash 4 for me.

 

[Skeleghost]

My only advice atm would be to work towards having images of the stages instead of just text. Something like this website for striking: https://bans.page/

Well, images take up a lot of space on the screen, and I'm worried about making it look right on a phone. The images do add a lot to that website though, so I should probably at least give it a try. It will be harder to do it when there are lists of 30+ stages though... I was thinking, as a compromise, maybe series icons? But images shouldn't be impossible... I'll definitely at least make an attempt though.

 

[Alias Tex]

One suggestion for stage grouping:

In Smash 4 we have more of a need for stage grouping than ever.

Spoiler: Stages That Could Be Grouped Together

Tri-Plats
Battlefield
Battlefield Forms
Dreamland
Yoshi's Story
Midgar
Fountain of Dreams

Omegas
Final Destination
Omega Forms
Wily's Castle
Pictochat
Umbra Clock Tower

(Potentially Duo-plats, but those might be interchangeable enough to be "echo stages")

We have a lot of stages that would be great to have legal if only they weren't so similar to other stages, such as the tri-plats and omegas. Including them all can cause centralization, but to chose to exclude some and include others is both subjective and removes metagame nuance. Grouping the stages together fixes these problems. Although likely worth the tradeoff, grouping stages can result in smaller problems such as all of them being banned with one ban when each of the stages aren't all equally good/bad for your character, or only the most polarizingly favorable tri-plat being a factor during selection.

My suggestion is the following: Group the stages, but put them into two slots: "Player1's choice" and "Player2's choice." Then we end up with:

Spoiler: Grouped Stage Slots

Player1's Tri-Plat
Player2's Tri-Plat

Player1's Omega
Player2's Omega

When a player chooses a stage to ban or creates a list for their opponent to pick from, they might, for example, say "I ban your Tri-plat." In this example, the worst case scenario tri-plat for the winner is banned, leaving the best case scenario tri-plat represented. This way, the nuances between the stages can be preserved without giving them too many stage slots. This also increases the number of possible stages for spectators. It also probably won't fill our stage list with too many of one kind of stage, because tri-plats or omegas would only take up two out of potentially two dozen legal stages. Finally, if we end up going with a clause preventing you from returning to a stage that you already won on, it could be applied to the player's ability to pick that you won on. (E.G. I win on my pick, now my ability to pick is banned, but we can still play on whatever is my opponent's best pick. Now if I win on my opponent's best pick, then neither of us can choose one unless we make an agreement.)

This obviously isn't a full stage selection ruleset, but it could be implemented into many of the current proposed systems so I hope that we can give it some consideration.

 

[ParanoidDrone]

As someone with no experience playing on Yoshi's Island Brawl competitively all that much, can someone explain why the slants at the ledges are such a big issue for some? I've observed this exact same sentiment multiple times now from different players.

At the risk of ruffling a few feathers, I'm 99% sure it's just johns and/or people having a really freaking narrow view of what constitutes an acceptable stage feature.

 

[Veggi]

I'll probably rename it to 'smash lists' or something though, because that's what everyone has ended up calling it.

Name ideas:

SOB: Stage Operating Buddy
SSB: Stage Strike Buddy
USSR: Ultimate Stage Strike Robo
USA: Ultimate Strike Automatic

 

[Midoriki]

I've finished up work on a huge study on stage selection in Smash 4, covering over 1300 games from 2018. While this is about Smash 4, it was created with intent to use the data moving on to Ultimate. Here's the original thread: https://www.reddit.com/r/smashbros/comments/a3dc5g/i_spent_the_last_two_months_working_on_a_massive

Obviously Ultimate is a unique game, but we should be able to look at Smash 4 info as helpful.

The biggest takeaway is that choosing the stage offers no statistical advantage. This can be viewed a few ways. Possibly it is explained by the fact that the stage chooser just lost previously, so it is logical they could lose again. Personally, I think it is simply the fact that character matchups are massively more important than stages in Smash.

Take a look and see the more in depth stuff if you're interested!

That's so cool! Thank you for doing so much work!

The data seem to agree with the principal that it's more important to let players ban stages their character(s) do poorly on, than to spend a lot of time picking the stage each match.

That raises a big question that is hard to answer: "how many stages should a player be able to ban?"
From the data it looks like that number is around 2 in Smash 4, but no way to know how data on a 6 stage list will translate into a potentially 20+ stage list, especially with so many similar stages in Ultimate.

If that number is fairly small, then Full List Partial Striking works great and is nice and simple.

If that number is fairly large, then it would be faster for players to have pre-made lists with them. Either pre-made ban lists, or if you are using ordered lists for the starting stage anyways, a list of N - X stages effectively bans X stages if you are only allowed to counter-pick from your opponent's list. Either would speed things up by having the players finish thinking about bans before the tournament (or during downtime if they're allowed to change between sets) and by removing all the back-and-forth and rock-paper-scissors needed for striking.

 

[Munomario777]

Haven't posted a list in a while... here's the stage list I'm considering right now, as well as a morph-based procedure that I'd like to experiment with.

 

[SuperDoom1]

I just wish people accepted what Smash is and didn't try to mold it into what they wanted it to be. In my eyes, all stages should be legal, tier lists shouldn't hold people back from training with other characters (as far as I know Kirby could have the potential to bring down Fox if we actually tried him out enough), and all items should be set to high. This is what Smash is. Not what the competitive scene says it is.

 

[DarthLuigi36]

I just wish people accepted what Smash is and didn't try to mold it into what they wanted it to be. In my eyes, all stages should be legal, tier lists shouldn't hold people back from training with other characters (as far as I know Kirby could have the potential to bring down Fox if we actually tried him out enough), and all items should be set to high. This is what Smash is. Not what the competitive scene says it is.

And there is absolutely nothing wrong with that. But when we want to have a competition to see who is the best player, some things are not conducive to that. Having a 1-on-1 game on Temple where you can just run away the whole time does not show who the better player is.

Look at it this way. Nobody is "molding" Smash to what they want. Those options to turn off items, change stages, switch from Time to Stock - they all exist because everyone is free to play the way they enjoy.

 

[SuperDoom1]

switch from Time to Stock

Everybody does that, competitive or not.

 

[boysilver400]

I don’t think all stages should have hazards off. Smashville is a good stage with or without them, but FOD, Melee’s Yoshi’s Story, And Dreamland 64 just become second battlefield. The later being hazards off is understandable, as I’m no fan of Whispy, but the other two were completely fine with hazards on.

Anywats, here’s my two cents on what the list should be:
Starters(All hazards off)
Battlefield
FD
Brinstar
Pokémon Stadium(both)
Wario Ware
Frigate
Halberd
Lylat
Castle Seige
Smashville(I think both versions will work just fine as starters)
Unova/Kalos Pokemon League
Skyloft
Delfino Plaza

Counterpick(hazards on)
Yoshi’s Story Melee
Yoshi’s Island Brawl
FOD
Town & City
Dreamland 64

 

[Inking2003]

Technically, there are over 200 legal stages. Each stage now has both a Battlefield and an Omega form. If all of the Battlefield and Omega forms are allowed to be picked, this opens up many different possibilities. Counter picking might not just be picking a stage that is hard for the opponents play style. For example, a dark stage might be picked if the opponent's character is light in color while the picker's is dark. This could make it so that it is harder for the opponent to see the picker's character, which would make it harder for hits to be landed and hits to be blocked.

 

[Frihetsanka]

I don’t think all stages should have hazards off.

There has already been tournaments (at least one) where people tried this and messed up (accidentally played with hazards on instead of off) and had to restart. This will keep happening, making hazards mixed not really a good option.

Smashville is a good stage with or without them, but FOD, Melee’s Yoshi’s Story, And Dreamland 64 just become second battlefield.

All of these are pretty unviable with hazards on. FoD has frame drops still, Yoshi's Island has Shy Guys bringing food (for some reason), and Dream Land 64 has had its blastzones changed to BF, making it redundant (more redundant than Yoshi's Story with hazards off).

Anywats, here’s my two cents on what the list should be:
Starters(All hazards off)
Battlefield
FD
Brinstar
Pokémon Stadium(both)
Wario Ware
Frigate
Halberd
Lylat
Castle Seige
Smashville(I think both versions will work just fine as starters)
Unova/Kalos Pokemon League
Skyloft
Delfino Plaza

Counterpick(hazards on)
Yoshi’s Story Melee
Yoshi’s Island Brawl
FOD
Town & City
Dreamland 64

Every stage with hazards "on" should be "off" or removed (remove at least FoD for lag and DL 64 for being basically BF). Brinstar, Halberd, and Delfino Plaza look really suspect (especially Delfino, which isn't really considered by many). Other stages are worth testing.

Unova and Kalos are very different, both could potentially be legal.

 

[boysilver400]

I've seen a couple tournaments myself. What's the official stage list if there is one?

 

[Androobie]

Is it confirmed that Yoshi's Story's Shy Guys still drop food when items have been turned off?

 

[Gunla]

Is it confirmed that Yoshi's Story's Shy Guys still drop food when items have been turned off?

It is unfortunately the case still, same goes with FoD's framedrops. I think at some point the former will be fixed, but people are probably too wary of FoD or Lylat at this time.

 

[Gunla]

Stagelist discussion will be moving to this thread:
https://smashboards.com/threads/ssbu-stagelist-discussion.464895/​