Frost | Odds
Puddings: 1 /// Odds: 0
Seems like a relevant topic now that we know 3.5 drops in a few days.
Do you think Allstar could form the basis of a viable tournament ruleset? Would you prefer such a ruleset over single-character? Maybe only as a side event? Why, or why not? What should such a ruleset look like in your opinion?
Would you attend an Allstar tournament - even if only to try it out?
My own thoughts, from further down in the thread:
Do you think Allstar could form the basis of a viable tournament ruleset? Would you prefer such a ruleset over single-character? Maybe only as a side event? Why, or why not? What should such a ruleset look like in your opinion?
Would you attend an Allstar tournament - even if only to try it out?
My own thoughts, from further down in the thread:
Alright, here we go. I apologize for the double post, but it seems like the most sensible way to distinguish my personal opinions from the queries in the OP.
Reasoning: Current rulesets are already designed to minimize variance, and allow players to compensate for bad maps and matchups, via the counterpicking/banning systems for stages and characters. Unfortunately, these systems are unable to compensate for variance within matches, and can place an undue amount of importance on playing the matchup game, rather than the other player.
Here's an example of such a situation [in 3.02]. Because I'm locally known as a Bowser player, everyone always bans the illuminati stages against me (Yoshi's Island, FoD, WarioWare), because Bowser is disproportionately strong on those stages, particularly against characters that are dependent on camping. Unfortunately for the other player, Bowser isn't the only character I can play - I also have a (relative to the local scene) strong Toon Link, who tends to thrive on stages and against characters that are polar opposites to Bowser's ideal locations and opponents.
Say I'm against an opponent that plays Fox and Ivysaur. I'm strongly incentivized to play Bowser against his Fox, and Toon Link against his Ivysaur.
First problem: the blind pick. Because this happens before the stage is selected, the matchup played could wildly favor either of us, dependent partially on the RPS, and partially on blind luck.
Basically, roughly speaking in terms of this microcosm meta,
Fox >> Toon Link
Bowser > Fox
Ivysaur >>>>>> Bowser
Toon Link = Ivysaur
Because I can't risk an auto-loss to Ivysaur, I'm forced to blind pick Toon Link. My opponent knows this, and blind picks Fox.
My must-ban stages are FD and Dreamland - I can't afford to get camped out by lasers on FD, and Toon Link loses to Fox period -- and Fox's recovery/gimp game is way stronger, so he benefits much more from Dreamland's huge blastzones.
Because he wants room to run around, My opponent's must-ban stages are FoD and Yoshi's Story.
Basically, we're bound to either play on Battlefield (if I'm lucky), or PS2 (if he is). I'm at a disadvantage either way, but it's a significantly more pronounced one if we play on PS2.
If I win game 1: I ban Final Destination, PS1, and Dreamland. He probably picks PS2 or Skyloft - once again, I can't afford to get counterpicked, so I have to suck it up and play Toon Link against Fox on an ideal Fox stage with lots of running space and a low ceiling. I am incredibly unlikely to win game 2, but I'll at least get an even game on game 3. I only have a 50/50% chance of winning the set, despite winning game 1.
If he wins game 1: he bans Yoshi's, FoD, and WW for obvious reasons. He goes Ivysaur, I counterpick Toon Link on battlefield- hardly ideal but at least it's an even game. I've pretty much already lost the set, though - because I'll need to play Toon Link on a Fox stage game 3 even if I win the even game 2.
**********************
All this changes pretty dramatically if we're playing Allstar. Maybe I can mix in a character that I'm not confident enough to main, but with which I may be able to take some stocks off the opponent before he's adapted - such as Roy, whose 'blitz factor' can at times outweigh that he's a relatively weak character overall - and he has a great matchup against Fox. Regardless, though, the most important thing IMO is that:
The ability to incorporate multiple counterpicks in a single game must necessarily reduce variance. Because you can hedge your bets against any character (or set thereof), a game becomes more about adaptation and knowledge of your opponent's habits, than centralized around any particular matchup. This property is huuuge for tournament play, because consistent, safe play is almost always what separates great players from good players.
There are other bonuses, too.
1. Spectator excitement should be much more palpable if a match features a rotating cast of characters from both players (or even just one player plays multiple characters), because there isn't time for any one matchup to get stale - a common problem in PM due to the prevalence of campy characters.
2. Because there are so many fun characters in PM, it's often tough to resist the temptation to play a bunch at the same time. Using Allstar in tournament would allow players to dip their toes in the water of playing a new character in a high pressure situation, without necessarily fully committing.
3. Related to (2), this would allow players to more comprehensively capitalize on the ability to play a bunch of characters, and thereby force their opponents to adapt. At the same time, dedicated single-character players are not severely penalized - their more focused expertise will necessarily still provide them an advantage over players who spread themselves too thin. If they could deal with every matchup before, they can still do the same -- if anything, Allstar would make them less likely to have to deal with long, grindy matches in bad matchups due to opponents' hedging.
4. Increased style points. Even if a player doesn't necessarily think he'll get an advantage by playing multiple characters in a match, it can still be another means to express his skill or dominance of the game - a key motivator for many players, including myself.
****************************
Anyway, as for rulesets, I think picking 4 characters per player per game would be way too excessive. Here's an example alternative ruleset. I'm not totally bent on it, but I've thought it out plenty, and think this is probably a good starting point for discussion. Keep in mind that in this example, I'm trying to give an advantage to the loser of the RPS game (to compensate for the RPS loss), but to make that advantage as small as possible, for obvious reasons. :
For the FIRST game,
1. Each player blind-picks a 'team' of three characters. This character pool is known BEFORE stage selection. These characters will later be organized into positions, such that the four stocks for each player are filled like this: PMMA, where P = Point, M = Middle, A = Anchor. I don't know much about Marvel, but I love their terminology here, and think it fits. Right now, the players' character lists look like this:
P1: Stocks: ???? ||| Pool: A, B, C
P2: Stocks: ???? ||| Pool: D, E, F
2. Normal Rock-Paper-Scissors takes place, and a stage is selected through bans as normal.
3. The winner of the RPS game (and therefore the final selector of the stage) must pick his 'anchor', or the last character that will see play. This is a relatively low-commitment decision, as this character is the least likely to see play.
Result:
P1(W): Stocks: ???A ||| Pool: B, C
P2(L): Stocks: ???? ||| Pool: D, E, F
Now, P1 has 1 decision left to make, and P2 has 2 decisions left to make.
4. P2 (the RPS loser) now chooses his Middle character- this is the highest-commitment decision he can make, because the Middle character is probably the most important due to consuming 2 stocks and almost certainly seeing play. The reason we have P2 making this high-commitment decision is so that P1 has something to work with -- because P1 only has one decision left - so it should be a relatively informed one. I hope that makes sense.
The result:
P1(W): Stocks: ???A ||| Pool: B, C
P2(L): Stocks: ?DD? ||| Pool: E, F
5. Now, P1 makes his only remaining decision before the match: how does he order B and C?
P1(W): Stocks: BCCA ||| Pool:
P2(L): Stocks: ?DD? ||| Pool: E, F
6. P2 finishes up.
P1(W): Stocks: BCCA ||| Pool:
P2(L): Stocks: EDDF ||| Pool:
7. Game is played!
8. FOR THE SECOND GAME:
9. The winner of G1 is restricted to his current team.
10. The loser of G1 can either pick a new team pool, or stick with his current one.
11. The players' teams are re-ordered via the same process as occurred before game 1 - as though the loser of the first game had lost Rock Paper Scissors.
12. The second game is played.
*** NOTE: I think it would probably be best to restrict players to AT MOST 2 teams per set, to keep things running smoothly, and reduce complications. While teams may be re-ordered for subsequent games, players are restricted to their first two pools.
Wonderful alternative suggested by AustinP: Same process, except at step 10, the loser of G1 can swap out at most one character on his team before the re-ordering on step 11. This optional swap can occur after every game of the set.
--------------
Now, obviously, there's probably room for refinement here. Maybe some portion of the process matters too much in terms of the game's results. Maybe the whole thing is too complicated - or too simple. But the simple fact here is that by giving each player multiple decision points to minimize his disadvantages, the probable variance of the match is smoothed out. There is literally no way for this system to produce a larger advantage for one player or the other than would be necessary if they played under a current ruleset.
I anticipate that the biggest objection people have will be that of time. TOs are under a lot of pressure as it is, and tournaments tend to go overtime already. Further, tournament attendees may find themselves confused by the relatively complicated rules.
I have two answers to this objection:
1. We don't know yet that a system like the ruleset outlined above would actually take more time than the current system! It is, in fact, entirely possible that it may take less time than current rulesets, and here's why: players tend to spend much more time on big decisions than small ones. I may spend several minutes waffling between several characters and stages if I have a lot of complex variables to consider. But because this proposed ruleset splits these big decisions into much smaller, easier to make ones, it is entirely possible that the process will be streamlined by making each decision much lower-commitment and more straightforward.
2. If an extra minute or two in character select produces higher-quality and more fun tournament sets, then it's a price worth paying. Full stop.
-----------------
I'm sure there will be more objections. I'd actually like to raise some questions in regards to my proposed example ruleset, in addition to those posed in the OP. I'll post my gut-instinct answers to these, but please add your own if you think you have something worth contributing.
Q1. After selecting a character pool of 2 or more characters, should a player be obligated to use all those characters?
A1. I think it would actually probably be best if players always selected 3 different characters, and then, if they didn't want to use all of them, simply place the same character in multiple positions. Eg, I could have Bowser playing my Point to try to surprise an opponent, and then have Toon Link as both Middle and Anchor, as he's a much safer bet in most matchups.
Q2. What to do after the first game? How many teams should each player be allowed to play in a regular set? What about a Finals set? What if finals are reset?
A2. I'm not really sure. Those details will probably take a lot of work.
Q3. What makes 3 an ideal maximum number of characters? Why arrange them into the 3 positions, as PMMA?
A3. First, it's asymmetric. Having the same character stick around for at least 2 stocks should stabilize the midgame somewhat, reducing the chaos that's inevitable with multiple characters from each player. This should help players (and spectators and commentary) to frame matches within a coherent narrative. And Odds's Middle Peach is finally down! Will he be able to bring it back with his Bowser? ... seems a lot more intrinsically sensible and exciting to me than Odds's first 3 characters are gone. Maybe the fourth time's the charm?. In addition, there's a few other things.
- First, there's a precedent in Marvel vs Capcom (and also King of Fighters!). Using familiar terminology and number of characters will allow veterans of that game to engage with the new mode much more easily. Also, MvC/KoF's success would appear to demonstrate that there's maybe something to it!
- Second, it's convenient from an organizational perspective - holding 3 characters in mind takes much less brainpower than 4, and requires more actual choices to be made (particularly if those characters' arrangement is necessarily asymmetric).
- Third, 3 characters allows much more matchup coverage (and thus, smoothing of matchup-related bell-curves) than two.
Q4. If you can just play a single character, why bother with Allstar?
A4. If people play the same character across the whole thing: WHO CARES? if the worst case is literally the same as the status quo, then it's pretty much inarguably a strict improvement.
------------------------
I'd love all feedback on this. I'll probably produce a much more polished, well-reasoned, and fleshed-out ruleset proposal once enough problems and solutions are found with this one. Many thanks to everyone who actually read this monster of a post, and more thanks yet to anyone who provides meaningful feedback.
IMO, given that the interface is decently polished: absolutely.Do you think Allstar could form the basis of a viable tournament ruleset?
Reasoning: Current rulesets are already designed to minimize variance, and allow players to compensate for bad maps and matchups, via the counterpicking/banning systems for stages and characters. Unfortunately, these systems are unable to compensate for variance within matches, and can place an undue amount of importance on playing the matchup game, rather than the other player.
Here's an example of such a situation [in 3.02]. Because I'm locally known as a Bowser player, everyone always bans the illuminati stages against me (Yoshi's Island, FoD, WarioWare), because Bowser is disproportionately strong on those stages, particularly against characters that are dependent on camping. Unfortunately for the other player, Bowser isn't the only character I can play - I also have a (relative to the local scene) strong Toon Link, who tends to thrive on stages and against characters that are polar opposites to Bowser's ideal locations and opponents.
Say I'm against an opponent that plays Fox and Ivysaur. I'm strongly incentivized to play Bowser against his Fox, and Toon Link against his Ivysaur.
First problem: the blind pick. Because this happens before the stage is selected, the matchup played could wildly favor either of us, dependent partially on the RPS, and partially on blind luck.
Basically, roughly speaking in terms of this microcosm meta,
Fox >> Toon Link
Bowser > Fox
Ivysaur >>>>>> Bowser
Toon Link = Ivysaur
Because I can't risk an auto-loss to Ivysaur, I'm forced to blind pick Toon Link. My opponent knows this, and blind picks Fox.
My must-ban stages are FD and Dreamland - I can't afford to get camped out by lasers on FD, and Toon Link loses to Fox period -- and Fox's recovery/gimp game is way stronger, so he benefits much more from Dreamland's huge blastzones.
Because he wants room to run around, My opponent's must-ban stages are FoD and Yoshi's Story.
Basically, we're bound to either play on Battlefield (if I'm lucky), or PS2 (if he is). I'm at a disadvantage either way, but it's a significantly more pronounced one if we play on PS2.
If I win game 1: I ban Final Destination, PS1, and Dreamland. He probably picks PS2 or Skyloft - once again, I can't afford to get counterpicked, so I have to suck it up and play Toon Link against Fox on an ideal Fox stage with lots of running space and a low ceiling. I am incredibly unlikely to win game 2, but I'll at least get an even game on game 3. I only have a 50/50% chance of winning the set, despite winning game 1.
If he wins game 1: he bans Yoshi's, FoD, and WW for obvious reasons. He goes Ivysaur, I counterpick Toon Link on battlefield- hardly ideal but at least it's an even game. I've pretty much already lost the set, though - because I'll need to play Toon Link on a Fox stage game 3 even if I win the even game 2.
**********************
All this changes pretty dramatically if we're playing Allstar. Maybe I can mix in a character that I'm not confident enough to main, but with which I may be able to take some stocks off the opponent before he's adapted - such as Roy, whose 'blitz factor' can at times outweigh that he's a relatively weak character overall - and he has a great matchup against Fox. Regardless, though, the most important thing IMO is that:
The ability to incorporate multiple counterpicks in a single game must necessarily reduce variance. Because you can hedge your bets against any character (or set thereof), a game becomes more about adaptation and knowledge of your opponent's habits, than centralized around any particular matchup. This property is huuuge for tournament play, because consistent, safe play is almost always what separates great players from good players.
There are other bonuses, too.
1. Spectator excitement should be much more palpable if a match features a rotating cast of characters from both players (or even just one player plays multiple characters), because there isn't time for any one matchup to get stale - a common problem in PM due to the prevalence of campy characters.
2. Because there are so many fun characters in PM, it's often tough to resist the temptation to play a bunch at the same time. Using Allstar in tournament would allow players to dip their toes in the water of playing a new character in a high pressure situation, without necessarily fully committing.
3. Related to (2), this would allow players to more comprehensively capitalize on the ability to play a bunch of characters, and thereby force their opponents to adapt. At the same time, dedicated single-character players are not severely penalized - their more focused expertise will necessarily still provide them an advantage over players who spread themselves too thin. If they could deal with every matchup before, they can still do the same -- if anything, Allstar would make them less likely to have to deal with long, grindy matches in bad matchups due to opponents' hedging.
4. Increased style points. Even if a player doesn't necessarily think he'll get an advantage by playing multiple characters in a match, it can still be another means to express his skill or dominance of the game - a key motivator for many players, including myself.
****************************
Anyway, as for rulesets, I think picking 4 characters per player per game would be way too excessive. Here's an example alternative ruleset. I'm not totally bent on it, but I've thought it out plenty, and think this is probably a good starting point for discussion. Keep in mind that in this example, I'm trying to give an advantage to the loser of the RPS game (to compensate for the RPS loss), but to make that advantage as small as possible, for obvious reasons. :
For the FIRST game,
1. Each player blind-picks a 'team' of three characters. This character pool is known BEFORE stage selection. These characters will later be organized into positions, such that the four stocks for each player are filled like this: PMMA, where P = Point, M = Middle, A = Anchor. I don't know much about Marvel, but I love their terminology here, and think it fits. Right now, the players' character lists look like this:
P1: Stocks: ???? ||| Pool: A, B, C
P2: Stocks: ???? ||| Pool: D, E, F
2. Normal Rock-Paper-Scissors takes place, and a stage is selected through bans as normal.
3. The winner of the RPS game (and therefore the final selector of the stage) must pick his 'anchor', or the last character that will see play. This is a relatively low-commitment decision, as this character is the least likely to see play.
Result:
P1(W): Stocks: ???A ||| Pool: B, C
P2(L): Stocks: ???? ||| Pool: D, E, F
Now, P1 has 1 decision left to make, and P2 has 2 decisions left to make.
4. P2 (the RPS loser) now chooses his Middle character- this is the highest-commitment decision he can make, because the Middle character is probably the most important due to consuming 2 stocks and almost certainly seeing play. The reason we have P2 making this high-commitment decision is so that P1 has something to work with -- because P1 only has one decision left - so it should be a relatively informed one. I hope that makes sense.
The result:
P1(W): Stocks: ???A ||| Pool: B, C
P2(L): Stocks: ?DD? ||| Pool: E, F
5. Now, P1 makes his only remaining decision before the match: how does he order B and C?
P1(W): Stocks: BCCA ||| Pool:
P2(L): Stocks: ?DD? ||| Pool: E, F
6. P2 finishes up.
P1(W): Stocks: BCCA ||| Pool:
P2(L): Stocks: EDDF ||| Pool:
7. Game is played!
8. FOR THE SECOND GAME:
9. The winner of G1 is restricted to his current team.
10. The loser of G1 can either pick a new team pool, or stick with his current one.
11. The players' teams are re-ordered via the same process as occurred before game 1 - as though the loser of the first game had lost Rock Paper Scissors.
12. The second game is played.
*** NOTE: I think it would probably be best to restrict players to AT MOST 2 teams per set, to keep things running smoothly, and reduce complications. While teams may be re-ordered for subsequent games, players are restricted to their first two pools.
Wonderful alternative suggested by AustinP: Same process, except at step 10, the loser of G1 can swap out at most one character on his team before the re-ordering on step 11. This optional swap can occur after every game of the set.
--------------
Now, obviously, there's probably room for refinement here. Maybe some portion of the process matters too much in terms of the game's results. Maybe the whole thing is too complicated - or too simple. But the simple fact here is that by giving each player multiple decision points to minimize his disadvantages, the probable variance of the match is smoothed out. There is literally no way for this system to produce a larger advantage for one player or the other than would be necessary if they played under a current ruleset.
I anticipate that the biggest objection people have will be that of time. TOs are under a lot of pressure as it is, and tournaments tend to go overtime already. Further, tournament attendees may find themselves confused by the relatively complicated rules.
I have two answers to this objection:
1. We don't know yet that a system like the ruleset outlined above would actually take more time than the current system! It is, in fact, entirely possible that it may take less time than current rulesets, and here's why: players tend to spend much more time on big decisions than small ones. I may spend several minutes waffling between several characters and stages if I have a lot of complex variables to consider. But because this proposed ruleset splits these big decisions into much smaller, easier to make ones, it is entirely possible that the process will be streamlined by making each decision much lower-commitment and more straightforward.
2. If an extra minute or two in character select produces higher-quality and more fun tournament sets, then it's a price worth paying. Full stop.
-----------------
I'm sure there will be more objections. I'd actually like to raise some questions in regards to my proposed example ruleset, in addition to those posed in the OP. I'll post my gut-instinct answers to these, but please add your own if you think you have something worth contributing.
Q1. After selecting a character pool of 2 or more characters, should a player be obligated to use all those characters?
A1. I think it would actually probably be best if players always selected 3 different characters, and then, if they didn't want to use all of them, simply place the same character in multiple positions. Eg, I could have Bowser playing my Point to try to surprise an opponent, and then have Toon Link as both Middle and Anchor, as he's a much safer bet in most matchups.
Q2. What to do after the first game? How many teams should each player be allowed to play in a regular set? What about a Finals set? What if finals are reset?
A2. I'm not really sure. Those details will probably take a lot of work.
Q3. What makes 3 an ideal maximum number of characters? Why arrange them into the 3 positions, as PMMA?
A3. First, it's asymmetric. Having the same character stick around for at least 2 stocks should stabilize the midgame somewhat, reducing the chaos that's inevitable with multiple characters from each player. This should help players (and spectators and commentary) to frame matches within a coherent narrative. And Odds's Middle Peach is finally down! Will he be able to bring it back with his Bowser? ... seems a lot more intrinsically sensible and exciting to me than Odds's first 3 characters are gone. Maybe the fourth time's the charm?. In addition, there's a few other things.
- First, there's a precedent in Marvel vs Capcom (and also King of Fighters!). Using familiar terminology and number of characters will allow veterans of that game to engage with the new mode much more easily. Also, MvC/KoF's success would appear to demonstrate that there's maybe something to it!
- Second, it's convenient from an organizational perspective - holding 3 characters in mind takes much less brainpower than 4, and requires more actual choices to be made (particularly if those characters' arrangement is necessarily asymmetric).
- Third, 3 characters allows much more matchup coverage (and thus, smoothing of matchup-related bell-curves) than two.
Q4. If you can just play a single character, why bother with Allstar?
A4. If people play the same character across the whole thing: WHO CARES? if the worst case is literally the same as the status quo, then it's pretty much inarguably a strict improvement.
------------------------
I'd love all feedback on this. I'll probably produce a much more polished, well-reasoned, and fleshed-out ruleset proposal once enough problems and solutions are found with this one. Many thanks to everyone who actually read this monster of a post, and more thanks yet to anyone who provides meaningful feedback.
Last edited: