Official Stage Legality Discussion: Philosophical

Raziek · Aug 9, 2010

Jack pretty much nailed what I was thinking. You really need to get off the "Super Smash Bros: Street Fighter Edition" concept.

He's the better player if the opponent doesn't know how to fight MK on Norfair lol. Learn the stages.

That said, MK on a lot of stages is pretty absurd, so I personally feel he ought to be banned for the sake of a more balanced game.

sunshade · Aug 9, 2010

ADHD said:
No, it's about having the better player win..

And the better player isn't the one that chose metaknight on norfair. We play the game how we want to because we feel it's best competitively, so I don't see how YOUR view being "the game" (even though it's not, when you still ban stages and items) matters when the game allows us to be shaped to our desires.

And heck, most of you in here don't even compete.

Why is the better player not the one that utilizes every option to maximize his potential odds of winning? Find the most powerful tactic and cashing in, is the essence of all competitive fighting games. Often times we find something that seems insane but then we find a counter and the game resets. If we dont find a counter the game is bad and we should not play it.

I am not yet convinced brawl is a bad game so I am not in support of people preemptively banning that which is yet to be proven to be broken.

Budget Player Cadet_ · Aug 10, 2010

ADHD said:
No, it's about having the better player win..

And the better player isn't the one that chose metaknight on norfair.

Scrub. Scrub scrub scrub scrub scrub. You sound like that scrubby ganondorf johning about losing, and how he only lost because you couldn't beat him with a "real" character. Seriously, that's exactly how you sound. It's the classic "you only won because you chose a cheap <character in most cases, sometimes stage, etc.>".

Let me ask you something. DLA vs. (random scrub X). DLA is obviously a really ****ing good player-he beat Ally's MK in friendlies FFS. Random Scrub X picks sheik, DLA stays ganon. Random Scrub X gets the lead by 1%, and then whips out the chain on a normal starter stage. Did the better player win? Should we ban sheik so that ganons have a chance? What about ICs, Ganon literally cannot beat their desynched blizzards. Should we ban ganon so that situations like that don't happen?

Another example. Let's keep using DLA. DLA vs. (random scrub Y), Temple is the only legal stage. ScrubY chooses falco, shoots one lazer, and circle camps the rest of the match. did the better player win?

It doesn't matter how genius of a player you are if you already throw in the towel at the CSS/SSS. You could be literally the perfect brawler; if you go Ganon, you will probably lose. And it's entirely your fault, your opponent was the better player in the match. MK is the only viable choice in the game unless we severely sculpt the ruleset/stagelist around him? Your loss for not picking him.

Also, ironic that you say MK on Norfair; without any extraneous, MK-centric rules, Norfair would probably be MK's worst reasonably legal stage, except for perhaps PTAD.

We play the game how we want to because we feel it's best competitively, so I don't see how YOUR view being "the game" (even though it's not, when you still ban stages and items) matters when the game allows us to be shaped to our desires.

See also: originalist vs. ~~scrub theory 101~~ constructivist.

And heck, most of you in here don't even compete.

Irrelevant and false.

DunnoBro · Aug 10, 2010

So uhm, why are we still taking ADHD seriously?

Budget Player Cadet_ · Aug 10, 2010

powerhouse said:
So uhm, why are we still taking ADHD seriously?

Because he's a top-level player. This should probably be proof that this is not a good reason to take someone seriously.

ADHD · Aug 10, 2010

powerhouse said:
So uhm, why are we still taking ADHD seriously?

Suck my ****ing **** you random *** scrub, ROFL.

Raziek · Aug 10, 2010

ADHD said:
Suck my ****ing **** you random *** scrub, ROFL.

This is exactly why nobody takes you seriously.

Luxor · Aug 10, 2010

*awkward pause*

So, I made this post in a different thread. Do you guys agree? Disagree?
I would actually like to get every stage ordered in a continuum like this, if you'd like to help.

sunshade · Aug 10, 2010

I like the idea but I need to ask, what match-up is it referring to? Every match-up is different and in every match up the spectrum will change.

This is why I support full list stage striking. You end up in the stage in the center of the spectrum for that match-up (only instead of aerial and grounded it is advantaged and disadvantaged). If the final stage ends up in one character's favor that means they are a better character.

Luxor · Aug 11, 2010

sunshade said:
Every match-up is different and in every match up the spectrum will change.

I always saw the spectrum as a universal upon which different characters were placed at different relative positions. My idea is that the more grounded character in a matchup will always want to CP FD/SV and the more aerial char will always want to CP Brinstar/RC except in the case where some factor other than air/ground balance influences the stage dynamic of the matchup, such as Frigate's ledgeless side or Delfino's walkoff or something like that.

Blah that came out bad

New analogy
Starter list

A B C D E F G

Character "X" is more grounded than Character "Y." The dark, left stages are more grounded than the light, right stages.

Now, it's clear what happens next- X strikes ABC, Y strikes EFG, Game 1 is on D. Imagine that X, however, is Olimar, and C is Frigate. As far as air/ground balance goes, Frigate is good for Olimar, especially the second transformation. However, the external influence of the ledgeless side is enough to make Olimar need to strike it, so X strikes CFG, Y strikes ABD, and Game 1 is on E.

I don't know where I'm going anymore lol, but I think that's the only way the spectrum is "matchup dependent." The air/ground spectrum is a universal, and is mostly the same as the matchup strike continuum, but the matchup strike continuum is slightly different for a few reasons.

Using the above analogy, the "air/ground continuum" is A B C D E F G, but the "matchup spectrum" is A B D E F G C, since C happens to be bad for the left-side character.

ignore this post btw, rambling except for the idea of "air/ground continuum" vs. "matchup spectrum"

sunshade · Aug 11, 2010

Think of it this way. The platform alignments and placements on different stages while generally promoting of aerial play will aid a specific character more than others.

Look at the different between a campy metaknight and a marth. Marth is going to be in the air alot via short hopped fairs. metaknight is going to be in the air alot but he is going to be floating around swinging at the open air. They are both going to be in the air and aerial stages will have a tendency to help both of them but they will be doing completely different things in the air.

As a result the spectrum will shift more. This is why I said there is really only "advantaged and disadvantaged" because that takes into account everything.

If player A loves final destination and player B hates it then that stage regardless of how it helps them will most likely not be the best stage for player B.

If character A is great on ground stages but has a partcular AT only on stage Y then he is going to prefer that stage even if the characters typically aided by the stage are based around a different style than him or her.

*another example which you will read despite already having got the point.*

Luxor · Aug 11, 2010

That's kind of what I was trying to say- some stages plat layout, lack of grabbable edges, stage-specific ATs, walkoffs, or blastzones might move it to a spot in the "matchup spectrum" other than where it was in the "air/ground continuum." I'll call such a stage an "outlier."

However, I don't really get what you meant in your MK v. Marth example. In that matchup, one of the characters is more grounded (Marth I think), and so Marth will prefer to go to FD (BF in actuality since it's favorable plat layout makes it an outlier).

So the "air/ground continuum" may be
FD SV YI BF Lylat

But the "matchup spectrum" is
BF(marth's outlier) FD SV(might actually be an outlier in MK's favor) YI Lylat

...
I don't see the "air/ground continuum" moving. It's a universal, immovable Fact.
The matchup spectrum shifts on stage quirks though, since they can't be described by just an "air/ground ratio" number- they have other factors too like blastzones, etc. Player preference too.

AMKalmar · Aug 11, 2010

People keep saying that a good way to start a set would be to strike from the whole stage list, resulting in the median stage. This was my initial thought when the I first started thinking about the topic of starters, but Overswarm convinced me otherwise in another thread.

Consider the MU ratio between two characters (A and B). In a set, A will counter-pick certain stages, and B will counter-pick other stages. The MU ratio will be better for A on his counter-picks, and better for B on his. The mean of these ratios accurately represents their MU because it considers stages that would actually be played in that MU. Therefore, these two characters should start on a stage that produces this MU ratio when facing each other in a set. It is unlikely that the median of all the stages will produce the same MU ratio as this.

Agree, disagree, comments, thoughts?

ADHD · Aug 11, 2010

AMKalmar said:
People keep saying that a good way to start a set would be to strike from the whole stage list, resulting in the median stage. This was my initial thought when the I first started thinking about the topic of starters, but Overswarm convinced me otherwise in another thread.

Consider the MU ratio between two characters (A and B). In a set, A will counter-pick certain stages, and B will counter-pick other stages. The MU ratio will be better for A on his counter-picks, and better for B on his. The mean of these ratios accurately represents their MU because it considers stages that would actually be played in that MU. Therefore, these two characters should start on a stage that produces this MU ratio when facing each other in a set. It is unlikely that the median of all the stages will produce the same MU ratio as this.

Agree, disagree, comments, thoughts?

"But versatility should be rewarded."

Luxor · Aug 11, 2010

AMKalmar said:
People keep saying that a good way to start a set would be to strike from the whole stage list, resulting in the median stage. This was my initial thought when the I first started thinking about the topic of starters, but Overswarm convinced me otherwise in another thread.

Consider the MU ratio between two characters (A and B). In a set, A will counter-pick certain stages, and B will counter-pick other stages. The MU ratio will be better for A on his counter-picks, and better for B on his. The mean of these ratios accurately represents their MU because it considers stages that would actually be played in that MU. Therefore, these two characters should start on a stage that produces this MU ratio when facing each other in a set. It is unlikely that the median of all the stages will produce the same MU ratio as this.

Agree, disagree, comments, thoughts?

So you're saying something along the lines of "Grounded characters shouldn't have to start on the median stage cuz they have one really good stage called FD so if you average it they start somewhere more grounded than the median" in effect.

A. How do we determine the mean, or the ratios? 100% subjective and therefore meaningless.
B. Versatility on stages is a quality of the character. You're arbitrarily buffing Falco & Co.

Full list stage striking (without this "mean" stuff, it's silly) is good though, ask BPC.

AMKalmar · Aug 11, 2010

Luxor said:
So you're saying something along the lines of "Grounded characters shouldn't have to start on the median stage cuz they have one really good stage called FD so if you average it they start somewhere more grounded than the median" in effect.

A. How do we determine the mean, or the ratios? 100% subjective and therefore meaningless.
B. Versatility on stages is a quality of the character. You're arbitrarily buffing Falco & Co.

Full list stage striking (without this "mean" stuff, it's silly) is good though, ask BPC.

I really wasn't saying anything about grounded characters. I'm saying the mean stage should be the goal, and the mean stage is not the median stage.

A - the MU ratios should be calculated by actual win/loss data in tournaments. MU ratios will be different between characters on different stages. 100% real tournament data and therefore very meaningful.
B - I agree that versatility is a quality of the character. But not every stage is not used in every set (or any set).

Edit, I just realized I am completely off topic. What I'm saying relates to which stages should be used as starters, not which stages should be legal.

Luxor · Aug 11, 2010

AMKalmar said:
Not sure I understand what you mean.
I really wasn't saying anything about grounded characters. I'm saying the mean stage should be the goal, and the mean stage is not the median stage.

In effect you said "let's buff Falco & Co." because as it is with stage striking, they go to a stage they can't adapt to that well. With your system, you could end up on SV or something. That's a CP for them.

The mean of one char's best CP(s) and another char's best CP(s) = matchup on "neutral?"

AMKalmar said:
A - the MU ratios should be calculated by actual win/loss data in tournaments. MU ratios will be different between characters on different stages. 100% real tournament data and therefore very meaningful.
B - I agree that versatility is a quality of the character. But not every stage is not used in every set (or any set).

Edit, I just realized I am completely off topic. What I'm saying relates to which stages should be used as starters, not which stages should be legal.

Starter stage philosophy is a big part of these threads, it's fine.

I think what you're really saying is that the Smash powers-that-be should determine (by the method you described etc.) the ONE stage that is "neutral" for the matchup. First of all, making the matchup 50-50 neutral by force is stupid, so I'll assume you meant neutral in the sense that the stage hands out even helps and hurts. It would take a lot of testing just to get that "perfect" stage for even ONE matchup (you'd have to get a large sample size of matches on every single stage), and there are IIRC 39 * 19 = SEVEN HUNDRED AND FORTY-ONE different matchups to do that for.

Not to mention the fact that this system arbitrarily takes away grounded characters weakness in versatility. If they do bad on 2/3 of the stages and good on 1/3, as it is they go somewhere baddish for them. With your system if the matchup is, say, 55:45 on FD but 45:55 on RC, they end up on a 50:50 stage, when in reality they should go to something like 47/53 because they are bad at the middle stages.

You don't just hand out buffs and nerfs like party favors, man.

Amazing Ampharos · Aug 11, 2010

I think PS1 is more aerial than Castle Siege; I'd flip those two with both definitely being more grounded than Lylat and Battlefield. Lylat vs Battlefield is microscopic in terms of how close it is. I can't make a real judgment.

It's an interesting suggestion, and it makes for an interesting approach to building a starter list though you'd find many stages that fit in the middle as counterpicks. For instance, PictoChat is very close to Smashville on that scale. You'd also find that a large number of cp stages go below Delfino but no stage goes above Final Destination (except Bridge of Eldin, if legal).

Character quality obviously shifts with any stage list you use or any procedure you use. Eventually you have to define a way to determine the "true" match-ups and then make your stage list to reflect those. ADHD relies upon an unspoken premise that the match-ups as reflected on the most flat and non-interactive stages reflect the "true" game, and if you take that premise, then his position is a simple logical conclusion. Of course, many people disagree with that premise, and in classic logical form, you only need to knock out one premise to knock out the whole argument.

It's a harder question than you might think though. Imagine stage striking from every stage in the game. There are no banned stages; we just don't play on customs and that's it. This is an even number, but if we allow Mushroomy Kingdom to pick a random form, we can make it an odd number so the procedure works. What would be the balance implications of this? What if we struck down to three stages and played matches on that?

I'm actually currently intrigued with that possibility. I'm unsure of the overall balance implications of this:

1. Double blind character pick (or simultaneous pick if not contested)
2. 41 stage stage striking in 5-6-11-10-1-2-2-2-1 order. Remaining stage is used for game 1. Last two stages to be struck are noted (last four stages in a best of five).
3. Loser of previous game either picks stage from the last stages struck in immediately previous striking procedure or changes character.
4. If loser changes character, winner may change character.
5. If winner changed character, loser may change character again.
6. If any character changing occured at all, stage striking is repeated to select a stage and last two stages (or last four) are noted for future games in the set.
7. Steps 3-6 are repeated for any subsequent games. When players are picking a stage instead of changing characters, they may not pick a stage they have won on since the last stage striking procedure.

Assume Spear Pillar is always on a random form. Mushroomy Kingdom is on a random form game 1 and on a form dictated by the loser of the previous match in subsequent games. The odds of either of these stages actually being played are slim at best regardless; a major premise of this procedure is that a lot of stages would essentially always be struck (such as Temple) and are only included so as to maximally reflect the game's natural balance. For the sake of the mental exercise, please neglect the real life time such a procedure would take; it is designed 100% to reflect the game's natural balance with no concern for practical application at the moment. Hence it's a philosophical thought!

Luxor · Aug 11, 2010

^ That would be interesting, especially if we found where the "air/ground balance" really lies.

My shot at the continuum using BBR 3.1 stagelist, starter and CP:

Grounded
Final Destination
Smashville
Pictochat
Yoshi's Island (Melee)
Yoshi's Island
Luigi's Mansion
Frigate Orpheon
Castle Siege
Pokémon Stadium 1
Pokémon Stadium 2
Lylat Cruise
Battlefield
Port Town Aero Dive
Distant Planet
Green Greens
Jungle Japes
Pirate Ship
Halberd
Delfino Plaza
Norfair
Brinstar
Rainbow Cruise
Aerial

A lot of that is reallyreally iffy since those CPs mostly affect characters in ways outside of air/ground balance, like Frigate's ledgeless side, etc.
Interestingly, by this list, Lylat and BF share the perfect middle slot, which sounds pretty reasonable to me.

AMKalmar · Aug 11, 2010

Luxor said:
In effect you said "let's buff Falco & Co." because as it is with stage striking, they go to a stage they can't adapt to that well. With your system, you could end up on SV or something. That's a CP for them.

The mean of one char's best CP(s) and another char's best CP(s) = matchup on "neutral?"

Starter stage philosophy is a big part of these threads, it's fine.

I think what you're really saying is that the Smash powers-that-be should determine (by the method you described etc.) the ONE stage that is "neutral" for the matchup. First of all, making the matchup 50-50 neutral by force is stupid, so I'll assume you meant neutral in the sense that the stage hands out even helps and hurts. It would take a lot of testing just to get that "perfect" stage for even ONE matchup (you'd have to get a large sample size of matches on every single stage), and there are IIRC 39 * 19 = SEVEN HUNDRED AND FORTY-ONE different matchups to do that for.

Not to mention the fact that this system arbitrarily takes away grounded characters weakness in versatility. If they do bad on 2/3 of the stages and good on 1/3, as it is they go somewhere baddish for them. With your system if the matchup is, say, 55:45 on FD but 45:55 on RC, they end up on a 50:50 stage, when in reality they should go to something like 47/53 because they are bad at the middle stages.

You don't just hand out buffs and nerfs like party favors, man.

Haha, I didn't even say "neutral," I don't know where you're getting this. Perhaps I phrased it poorly if you interpreted it that way.

What I mean to say is that the first stage in a set should accurately represent the matchup between two characters (achieving that is another matter). Because of counter-picking and the limited number of games in a set, the matchup is not quite accurately represented if you include all of the stages. Finally, if you strike from the entire stage list, you may not end up with a stage that accurately represents the matchup.

For example, Mickey Mouse vs Donald Duck:
Let's say there are 21 stages total to choose from.

Mickey wins vs Donald:
65% of the time on 6 stages,
60% of the time on 6 stages,
55% of the time on 1 stage,
50% of the time on 2 stages,
and 45% of the time on 6 stages.

Using stages that would be realistically be counter-picked in a set, the matchup ratio is 55% (~56% if you include all) in Mickey's favour. Therefore, the set should start on a stage that achieves this same ratio - the stage where Mickey wins 55% of the time. However, if you choose the starter by striking from the entire stage list, you will end up with a stage on which Mickey will win 60% of the time. This does not accurately represent the matchup between these two characters.

Basically:
Mean = (65 + 45) / 2 = 55% Accurate representation of the matchup, the set should start here.
Median = 60% Inaccurate representation of the matchup, the set should not start here.

fkacyan · Aug 11, 2010

You know...

Defining versatility as just air and ground game is very convenient and all, but uh

there are a lot of other kinds of versatility.

- Being able to abuse walls
- Being able to abuse walkoffs
- Being able to circle camp

etc

If we're looking to let the most versatile characters win, there is no reason to disable stages other than those like Warioware with intrisically random results.

Luxor · Aug 11, 2010

AMKalmar said:
Haha, I didn't even say "neutral," I don't know where you're getting this. Perhaps I phrased it poorly if you interpreted it that way.

What I mean to say is that the first stage in a set should accurately represent the matchup between two characters (achieving that is another matter). Because of counter-picking and the limited number of games in a set, the matchup is not quite accurately represented if you include all of the stages. Finally, if you strike from the entire stage list, you may not end up with a stage that accurately represents the matchup.

For example, Mickey Mouse vs Donald Duck:
Let's say there are 21 stages total to choose from.

Mickey wins vs Donald:
65% of the time on 6 stages,
60% of the time on 6 stages,
55% of the time on 1 stage,
50% of the time on 2 stages,
and 45% of the time on 6 stages.

Using stages that would be realistically be counter-picked in a set, the matchup ratio is 55% (~56% if you include all) in Mickey's favour. Therefore, the set should start on a stage that achieves this same ratio - the stage where Mickey wins 55% of the time. However, if you choose the starter by striking from the entire stage list, you will end up with a stage on which Mickey will win 60% of the time. This does not accurately represent the matchup between these two characters.

Basically:
Mean = (65 + 45) / 2 = 55% Accurate representation of the matchup, the set should start here.
Median = 60% Inaccurate representation of the matchup, the set should not start here.

What makes the mean the ideal measure of central tendency? Explain why, other than your totally subjective opinion, we should use the mean rather than the median.

At a loss? Me too. The difference between the median and the mean is that the mean helps out those with stronger CPs (as % points from the median). The matchup may be in Mickey's advantage, but out of the stages, the ones that help Donald help him out more than Mickey's help him. However, Mickey has lesser help from a larger number of stages. What your "mean" idea boils down to is simply rewarding characters for having some relatively strong CPs rather than favoring them for doing well on a larger number of stages. All this does is end up favoring Falco/Diddy/ICs for having very strong CPs in FD/SV/(Japes, Picto, or BF depending on character) and MK/Wario for their "autowins" on RC/Brinstar. Shouldn't a character's starter stage reflect some advantage to them based on their inherent advantage of being more versatile on a greater number of stages? The answer isn't even debatable. Yes! If you say no, you're taking away tools built into the character, and you're as scrubby as someone saying "ban d3 the ch@!ngr@b!!!", which is just the same: an attribute built into the character. Your system arbitrarily hands out buffs and nerfs rather than following the natural system of allowing characters with good stage versatility as one of their attributes to use that to their advantage.

You keep coming back to the idea that some stage "accurately represents the matchup." Matchups involve stages too, and character viability is affected by stage performance.
tl;dr DONALD DOESN'T DESERVE A 44:56, MICKEY IS A BETTER CHARACTER

Thiocyanide said:
You know...

Defining versatility as just air and ground game is very convenient and all, but uh

there are a lot of other kinds of versatility.

- Being able to abuse walls
- Being able to abuse walkoffs
- Being able to circle camp

etc

If we're looking to let the most versatile characters win, there is no reason to disable stages other than those like Warioware with intrisically random results.

I'd list "recovery versatility" as well, for stages like Frigate, and "platform use versatility," since some characters (Marth, Oliman) just camp plats soooo well.

But to get to the meaty half-serious point of your post, obviously we sacrifice some of the game's integrity for a measure of character balance. We could let Brawl be "Fox only, no items, Hyrule Temple" or "DDD only, no items, Bridge of Eldin" but we choose not to because we're ~~scrubs~~ slightly constructivist.
I'm cool with walls though, I'd be fine with Onett legal at least.

BSP · Aug 11, 2010

Luxor said:
^ That would be interesting, especially if we found where the "air/ground balance" really lies.

My shot at the continuum using BBR 3.1 stagelist, starter and CP:

Grounded
Final Destination
Smashville
Pictochat
Yoshi's Island (Melee)
Yoshi's Island
Luigi's Mansion
Frigate Orpheon
Castle Siege
Pokémon Stadium 1
Pokémon Stadium 2
Lylat Cruise
Battlefield
Port Town Aero Dive
Distant Planet
Green Greens
Jungle Japes
Pirate Ship
Halberd
Delfino Plaza
Norfair
Brinstar
Rainbow Cruise
Aerial

A lot of that is reallyreally iffy since those CPs mostly affect characters in ways outside of air/ground balance, like Frigate's ledgeless side, etc.
Interestingly, by this list, Lylat and BF share the perfect middle slot, which sounds pretty reasonable to me.

I would put Japes in the more nuetral range, or tied with tied with Distant P. You might even be able to put Brinstar as the most aerial stage. Brinstar has extremely uneven ground, permanantly, while RC has the boat for a bit, and most of the latter part of the stage is even. Plus, Brinstar's got the lava which gives aerial based characters an easier time avoiding it, and better platform pressure, etc.

Raziek · Aug 11, 2010

I like that list, and I agree with mariobrouser's suggestion to put Brinstar over RC.

sunshade · Aug 11, 2010

My feelings in regards to full list stage striking is fairly simple. If a character is better on a majority of stages they are a better character.

If you are to strike from the full list the final product is the stage in which both players are an equal distance from their most advantaged or disadvantaged stage on spectrum.

This stage can easily be better for one character than the other however if the stage aids character A more than B then character A is better than B.

Raziek · Aug 11, 2010

Oh, adding something I forgot to state that should be implied but bears stating anyway: When I dicuss full striking I consider only the legal stages. None of this wasting strikes on 75m or Mario bros.

sunshade · Aug 11, 2010

Raziek said:
Oh, adding something I forgot to state that should be implied but bears stating anyway: When I dicuss full striking I consider only the legal stages. None of this wasting strikes on 75m or Mario bros.

I am the same.

I should start saying "full list striking among legal stages".

Amazing Ampharos · Aug 11, 2010

Raziek said:
Oh, adding something I forgot to state that should be implied but bears stating anyway: When I dicuss full striking I consider only the legal stages. None of this wasting strikes on 75m or Mario bros.

Is there any particular reason we shouldn't have to "waste strikes" on such stages? If they are extremely powerful stages for one side in a particular matchup, isn't it unfair to that side to remove them from striking under and originalist mindset? You don't end up with a broken result when they're struck, but you do end up with the power of characters on these stages not being completely removed from the game (thus we avoid changing the game more than we have to).

I'm mostly just seeing how well thought out and self-consistent you guys are with this position.

AMKalmar · Aug 11, 2010

Luxor said:
What makes the mean the ideal measure of central tendency? Explain why, other than your totally subjective opinion, we should use the mean rather than the median.

To calculate the matchup ratio between two characters, take the mean of their matchup ratios for stages that would be counter-picked by the characters in a set. It's not even my opinion, it's just math.

At a loss?

Nope.

Me too.

Glad to see we aren't making preemptive assumptions of each-other's responses.

The difference between the median and the mean is that the mean helps out those with stronger CPs (as % points from the median).

A stage with the mean matchup ratio wouldn't "help" anyone, it would accurately represent the matchup ratio... because the mean is the matchup ratio. The median stage is just a stage that is not necessarily relevant to the matchup.

The matchup may be in Mickey's advantage, but out of the stages, the ones that help Donald help him out more than Mickey's help him. However, Mickey has lesser help from a larger number of stages.

I gave an example of a matchup between two characters, the win percentages are data for the matchup between those two characters only. For all we know, Donald could be a strong character and Mickey could be a terrible character that just happens to do well versus Donald.

What your "mean" idea boils down to is simply rewarding characters for having some relatively strong CPs rather than favoring them for doing well on a larger number of stages.

This "mean" idea would accurately represent the matchup for what it is.

All this does is end up favoring Falco/Diddy/ICs for having very strong CPs in FD/SV/(Japes, Picto, or BF depending on character) and MK/Wario for their "autowins" on RC/Brinstar.

Again, this wouldn't favour anyone, it would give an accurate representation of the matchup between two characters.

Shouldn't a character's starter stage reflect some advantage to them based on their inherent advantage of being more versatile on a greater number of stages? The answer isn't even debatable. Yes!

As long as the advantage that character has over the other character is an accurate representation of the total matchup between those characters, then yes. Otherwise, no.

If you say no, you're taking away tools built into the character, and you're as scrubby as someone saying "ban d3 the ch@!ngr@b!!!", which is just the same: an attribute built into the character.

The characters still have their tools. The matchup between the two characters is just accurately represented on the first game in a set.

Your system arbitrarily hands out buffs and nerfs rather than following the natural system of allowing characters with good stage versatility as one of their attributes to use that to their advantage.

If accurately representing the matchup is buffing/nerfing characters then you aren't accurately representing the matchup.

You keep coming back to the idea that some stage "accurately represents the matchup."

The starter stage should accurately represent the matchup between two characters in a set because the matchup on that stage is an accurate representation of the matchups on all the stages that would be counter-picked in that set. I keep saying it because you're not getting it. I continue to repeat myself, hoping that the next time your read it, you might understand it.

Matchups involve stages too, and character viability is affected by stage performance.

Yeah, never said nor implied otherwise.

tl;dr DONALD DOESN'T DESERVE A 44:56, MICKEY IS A BETTER CHARACTER

The matchup between Mickey and Donald is 55:45 (56:44 if you consider stages that would never even be played in a set between them). Mickey is not necessarily a better character, he just has the advantage in this matchup. The matchup should be accurately represented. A stage that accurately represents their matchup is one where Mickey wins 55% of the time.

In summary, Mickey doesn't deserve a starter stage that results in him winning 60% of the time versus Donald. Likewise, Donald doesn't deserve a starter stage that results in him winning 50% of the time versus Mickey. Both characters deserve to have the matchup ratio accurately represented on the first game of a set - a stage where Mickey wins 55% of the time.

Luxor · Aug 12, 2010

prepare for caps rage post to answer AMKalmar
open only if you dare

YOU FAIL TO ANSWER WHY THE MATCHUP IS TRULY 55:45 RATHER THAN 60:40. YOUR ANSWER TO WHY IS THAT "ACCORDING TO MY REASON IT IS 55:45." WELL, ACCORDING TO MY REASON IT IS 60:40. WHOOPITY DOO.

YOU THINK THE MEAN IS THE ANSWER. YOU DIDN'T EVEN ANSWER MY POINT ABOUT WHY YOU THINK MEAN > MEDIAN. YOU JUST TOLD ME HOW TO CALCULATE THE MEAN, WHICH I'VE KNOWN SINCE I WAS SEVEN.

ALL YOU'RE SAYING IS "MEAN REPRESENTS THE MATCHUP FOR WHAT IT IS." WHAT I'M SAYING IS "NO IT DOES NOT, HERE'S WWWWWHHHHHHHHHYYYYYYY!!!!" PLEASE DO ONE OF TWO THINGS FOR ME.

1. ANSWER WHY THE MEAN IS SUPERIOR TO THE MEDIAN AS A MEASURE OF CENTRAL TENDENCY (DO YOU KNOW ANYTHING ABOUT STATS?)
2. YOU KEEP SUPPORTING MEAN > MEDIAN BY SAYING "IT ACCURATELY REFLECTS THE MATCHUP." WHEN I ASK "WHY" YOU SAY "IT DOES." TELL ME WHY IT 'JUST DOES,'
BECAUSE
IT
DOESN'T PERIOD.
USING THE MEDIAN IS SUPERIOR BECAUSE IT GIVES MORE VERSATILE CHARACTERS THE ADVANTAGE THEY ARE BUILT INTO THE GAME WITH

ALL YOU'RE SAYING IS "IT'S MY OPINION THE MEAN REPRESENTS THE TRUE MATCHUP RATHER THAN THE MEDIAN" WHILE I'M ACTUALLY SUPPORTING MY STATEMENT. YOU DIDN'T EVEN EXPLAIN WHY THE MATCHUP SHOULD BE ON A 55:45 STAGE, ALL YOU SAID WAS "THAT'S WHAT THE MEAN IS SO IT'S CORRECT." ENLIGHTEN ME, WHY IS THE MEAN CORRECT? IN THAT LAST POST, LITERALLY ALL YOU SAID WAS "NO, I'M RIGHT." ABSOLUTELY ZERO EXPLANATION AS TO WHAT REASONING MADE YOU THINK THAT. PLEASE EXPLAIN

WHY?

that was pretty cathartic, aaaahh
ignore the caps everybody, it just felt good
oh and btw AMKalmar read sunshade's second most recent post, sums up this stuff pretty good

Apart from that, I'm going to make a separate thread all about air/ground spectrum and stuff.

AMKalmar · Aug 12, 2010

Because sets never go past 7 (maybe 9) games. Any stages that wouldn't be counter-picked are irrelevant.

If someone gave me a list of 21 numbers and asked me what the average is, I would add up the numbers and divide by 21. It would not occur to me to take the median.

sunshade · Aug 12, 2010

Lets say we have character A and B.

A goes 90-10 with B on A's best stage and 80-20 on his second best.

B goes 60-40 with A on B's best stage.

Every other stage they go 55-45 B's advantage.

Do the middle stages still not matter?

AMKalmar · Aug 12, 2010

In that specific scenario they do because those stages are likely to be counter-picked by B in a set. However, all those 55:45s would add up to the weight of only a few counter-picks when considering them since each individual stage will be chosen less frequently.

X vs Y on five stages: 1, 2, 3, 4, 5. They play four games on each stage. X wins three times on 1, twice on 2, once on 3, once on 4, and once on 5. X wins a eight games of twenty - 40% of the games. The median says he only wins 25% of the games... but he won 40%!

Luxor · Aug 12, 2010

AMKalmar said:
Because sets never go past 7 (maybe 9) games. Any stages that wouldn't be counter-picked are irrelevant.

...What the 4asterisks does this have to do with what I said? You're saying that because the average set of 3 games doesn't occur on a whole lot of stages, the character's viability on other stages doesn't matter? That's terribad logic.

AMKalmar said:
If someone gave me a list of 21 numbers and asked me what the average is, I would add up the numbers and divide by 21. It would not occur to me to take the median.

*facepalm*
Well duh! Asking for the average Is Exactly The Same Thing As Asking For The Mean! Check the dictionary!
Also, what you're saying about why mean > median is "mean popped into my head first."
Personal bias =/= legit reasoning.

AMKalmar said:
In that specific scenario they do because those stages are likely to be counter-picked by B in a set. However, all those 55:45s would add up to the weight of only a few counter-picks when considering them since each individual stage will be chosen less frequently.

Here you flat out that a character's performance on neutrals is less important than their performance on their own counterpicks. It doesn't matter how good you are on your counterpick; if the matchup is evenish with regard to both player skill and character matchup, a CP is as good as an autowin since it gives you one extra edge your opponent lacks. It's really only performance on starters that matters AT ALL in a standard 3-game set.
Let's talk about Ice Climbers. Let's say (more or less correctly) that they beat every other char on FD, their CP. Let's also say (more or less correctly) that they suck on absolutely every other stage. Only a slight oversimplification. Are these hypothetical ICs good? No! They suck lollipops on the starter Game 1. It doesn't matter that they get an instant win on FD game 2 (ignoring stage bans) because they'll get wrecked even worse Game 3. They are screwed hard because they suck on stages.
In reality where the ICs have a few decent stages, your idea is still terrible. The "true" matchups occur on the mesian stage- the Xth worst IC stage and the Xth worst opponent stage. Let's say they're playing Marth, who will beat them by a sizable margin on all but 2 or 3 legal stages. However, the ICs win even more on their CPs, and the mean drags the matchup down to an even stage. That's stupid! Using the median puts it on a good Marth stage, rewarding Marth for not-being a one-trick pony and for actually being VIABLE on 3/4 of the legal stages in the game. Using the mean just arbitrarily buffs ICs/Diddy/Falco, as I said before.

AMKalmar said:
X vs Y on five stages: 1, 2, 3, 4, 5. They play four games on each stage. X wins three times on 1, twice on 2, once on 3, once on 4, and once on 5. X wins a eight games of twenty - 40% of the games. The median says he only wins 25% of the games... but he won 40%!

All that's saying in effect is that X played an abnormally large number of games on strong CPs for him but sucks otherwise- on the majority of those stages his win rate is only 25%, and he can only even go even on his second strongest stage. X IS trash, 25%. Out of five stages he only has a chance on one and an advantage on one while he gets r@@@@@ped on 3. Using the mean RANDOMLY BUFFS HIM BY 15%. THAT'S A SIGNIFICANT DIFFERENCE.

Besides, you're interpreting the data wrong. The median means the percent of the time he will win on the halfway stage- 25% because X really sucks on almost all stages and therefore has a major weakness as a character. The mean means the percent he will win if all stages are played equally- 40% because he has one decent and one good stage.

I still haven't heard a reason why mean > median besides "it was my idea" and "it's my opinion CP performance should matter more than overall performance," both of which are absolutely TERRIBLE justifications. L2Rgue.

Budget Player Cadet_ · Aug 12, 2010

What Luxor is saying makes a lot of sense... And in fact, it seems to make perfect sense when you look at what gets rewarded.
With stage striking from a huge list, we reward characters that do well on a lot of stages. (Metaknight, who does very well on pretty much every stage for example)
With counterpicking, we reward characters who do extremely well on a few stages. (Ganondorf, whose 90-10 matchups go down by a LOT on his favorite counterpick stages, norfair and brinstar, for example)
That actually seems really smart to me.

fkacyan · Aug 12, 2010

But to get to the meaty half-serious point of your post, obviously we sacrifice some of the game's integrity for a measure of character balance. We could let Brawl be "Fox only, no items, Hyrule Temple" or "DDD only, no items, Bridge of Eldin" but we choose not to because we're scrubs slightly constructivist.
I'm cool with walls though, I'd be fine with Onett legal at least.

If we're going to balance the game through stages, we can do a lot more than arbitrarily deciding that air games are the most versatile, as we've seen to have done.

We can't say "Let's reward versatility" and then balance the game around a particular type. We either don't balance it at all and let true versatility win, or we actually try to balance the game.

Luxor · Aug 12, 2010

Thiocyanide said:
If we're going to balance the game through stages, we can do a lot more than arbitrarily deciding that air games are the most versatile, as we've seen to have done.

We can't say "Let's reward versatility" and then balance the game around a particular type. We either don't balance it at all and let true versatility win, or we actually try to balance the game.

Except we didn't decide- we simply banned stages with degenerate gameplay. Aerial characters then emerged as the most versatile on the remaining, worthwhile stages the game came with.
As for going full constructivist and actively balancing the game, that means that the stage list would have to shrink, since short of custom stages (which would be great) we can't add stages to Brawl. I think APEX reminded us that some characters do get buffed by smaller, more "traditional" 5-starter rulesets, but MLG still showed respectable character diversity as well with its "lolwut" liberal ruleset. Actively balancing the game through stagelists is an iffy endeavor at best, since we really don't have a lot of concrete data about how stagelists affect results. Individual TOs should be encouraged to experiment with rebalancing stagelists, but I think the Smash scene as a whole should hold back for now.

AMKalmar · Aug 12, 2010

Please avoid saying "your logic is bad." We obviously both think the other's logic is wrong. If we didn't we wouldn't be arguing.

I'm on the fence about the CP point because if we want to accurately represent the MU between two characters in Brawl, all stages should be considered. I was thinking about how to accurately represent the MU between two characters in Tournament Brawl - a completely different game. However, there are some stages that just wouldn't be played in certain MUs. Regardless, this should have little effect on actual statistical wins.

The reason I do not understand your point of view on selecting the starter is this: even if all stages were played equally, the average win/loss ratio is still not necessarily the same on the median stage (as seen in the Mickey/Donald example, it becomes ~56%). I don't see why you would use the median stage (or any stage) as the starter if it does not represent the statistical average.

Going back to the Mickey vs Donald example:
Lets assume all stages are played equally. Of all the matches between Mickey and Donald, Mickey wins 56% of the time. Mickey is better on more stages, but this does not change the fact that Mickey wins 56% of the time. A player doesn't win every match he has advantage on. On average, Mickey wins 56% of the time versus Donald. He just does. That is how much he wins, that is how much he should win. I can't understand why the number of stages he has advantage on should be made to increase how often he wins.

Just making a statement to sum it up: On average, Mickey wins on more stages. This doesn't mean Mickey never loses on those stages, and this doesn't change how often Mickey wins.

fkacyan · Aug 12, 2010

Luxor said:
Except we didn't decide- we simply banned stages with degenerate gameplay. Aerial characters then emerged as the most versatile on the remaining, worthwhile stages the game came with.
As for going full constructivist and actively balancing the game, that means that the stage list would have to shrink, since short of custom stages (which would be great) we can't add stages to Brawl. I think APEX reminded us that some characters do get buffed by smaller, more "traditional" 5-starter rulesets, but MLG still showed respectable character diversity as well with its "lolwut" liberal ruleset. Actively balancing the game through stagelists is an iffy endeavor at best, since we really don't have a lot of concrete data about how stagelists affect results. Individual TOs should be encouraged to experiment with rebalancing stagelists, but I think the Smash scene as a whole should hold back for now.

If aerial characters (And this is limited to two, MAYBE three characters at best when it comes to true aerial versatility) are so dominant on their strong counterpicks, I don't see how this is any less degenerate than characters capable of circle camping (Fox, Falco, and Rob to a much lesser extent) on stages like Hanenbow and Temple.

Not having the data is no excuse not to do it. That just means that we go and get the data.

AMKalmar · Aug 12, 2010

After spending enough time in this thread, I now know how to respond to this.

ADHD said:
"But versatility should be rewarded."

No it should not. No characters should be rewarded for anything. Rewarding characters for stage versatility is like rewarding characters for weight - it doesn't accurately represent how much they actually win. One character will win a certain amount versus another character. That is how much they should win because that is how much they do win. No more, no less.

Official Stage Legality Discussion: Philosophical

Charging Limit All Day

Smash Ace

Smash Hero

The Free-est

Smash Hero

Smash Hero

Charging Limit All Day

Smash Champion

Smash Ace

Smash Champion

Smash Ace

Smash Champion

Smash Ace

Smash Hero

Smash Champion

Smash Ace

Smash Champion

Balanced Brawl Designer

Smash Champion

Smash Ace

Smash Hero

Smash Champion

Smash Legend

Charging Limit All Day

Smash Ace

Charging Limit All Day

Smash Ace

Balanced Brawl Designer

Smash Ace

Smash Champion

Smash Ace

Smash Ace

Smash Ace

Smash Champion

Smash Hero

Smash Hero

Smash Champion

Smash Ace

Smash Hero

Smash Ace

Information

Network