Proposed Change to Tiebreaker Cases Due to Timeout

[SpaghettiWeegee]

Hello, Smashboards! I was browsing old tournament footage and the thought occurred to me that, due to the unique nature of Smash Bros. and specifically the diverse roster of Smash 4, the currently-accepted rule for breaking ties in the event of the match clock hitting zero before one player has lost all stocks is insufficient and unfair.

As it is the most recent thoroughly-documented form of generally-accepted Smash 4 rules for competitive play, I make this claim based on the Apex 2015 Rulebook. As it is written:

The winner of a match that goes to time (time out) will be determined by stocks and percentage. When the timer hits 0:00 player with the higher stock count is the winner. If both players are tied in stocks the player with the lower percentage is the winner. In the event of a percentage tie, or a match in which both players lost their last stock simultaneously, a 1 stock tiebreaker will be played with time limit equal to the regular time limit divided by the regular number of stocks, rounded up. Sudden Death is not to be played at all, and will not count. (Emphasis mine)

While this rule makes sense for traditional fighting games, where win/loss conditions are decided based on a set resource pool being drained to zero, the more open-ended way in which victory can be achieved in Smash (with certain characters struggling to KO even at percentages as high as 120% and others having reliable kill options as low as 60% in some matchups) indicates that a similar understanding of fairness cannot be applied to such a different game type. Additionally, some characters, by virtue of being heavy, fall into combos more easily, and as a result repeatedly find themselves at percent deficits faster than other characters despite, in actuality, not actually being any closer to being smashed into blast zones than their opponent. A Bowser at 80% may have a higher percent than a Jigglypuff at 60%, but any Smash player can tell you that Bowser is in no more danger of being knocked out than his opponent in this scenario (some might even argue that Bowser is actually safer than Jigglypuff, although I have no data to support this claim currently).

My proposed solution to this problem is to disregard percent as a factor in determining the victor in these time-over situations and instead use a different, more reliable number-- something that, if I were to propose a dumb name for it off the top of my noggin' would be called something like Smash Susceptibility, although I wholeheartedly encourage suggesting a better name in the thread below.

Essentially, Smash Susceptibility (SS) is the distance which a fighter will be launched if hit by a smash attack. It can be calculated by taking a given character's recorded percent at the moment of timeout and modifying it by their weight value. I'm no dataminer, so I don't have this information on me right now, but I'm going to throw out some dummy numbers to just give a feel for how it might work. Let's say Ganondorf's weight value is 100 (pretty heavy), and he ends a match with 70% left on his final stock. To calculate his SS, we'd simply divide his percent by his weight value (70/100) and we see that Ganondorf concluded the game with a SS rating of .7. Now let's take a look at the G-dawg's opponent, Mr. Game and Watch, who has a weight value of 25 (super lightweight) at the moment of timeout is at 50% on his final stock. G&W's SS calculates out to be 2.0, which is significantly higher than Ganondorf's SS despite G&W's percent damage being lower. In this case, it looks like Ganondorf, with his lower comparative SS, is actually much safer from being KOed by a raw smash attack, and thus will win the tiebreaker.

To summarize, the formula for SS will look something like Final Percent/Weight Value = Smash Susceptibility. Whoever ends the game with the highest SS will be considered the winner. In the event of tied SS, identical proceedings from the Apex 2015 rulebook should be followed.

This model does not take into account a character's vulnerability to gimps or the length of that character's recovery. This could potentially cause it to be considered an unfair rule as well, because it may result in a situation where one character may be launched further than their opponent but be able to recover, while if their opponent were launched by a smash attack of the same power they would be unable to recover. If this is considered terribly problematic, an additional variable taking a character's maximum recovery distance could be taken into account, increasing their SS if their sum total recovery distance falls below a certain threshold (sorry, Little Mac). This, too, encounters trouble when considering some characters have spectacular horizontal recovery but practically nonexistent vertical recovery (e.g. Donkey Kong) or vice versa, and runs into the additional wrinkle of potentially overcomplicating something I'm already probably overcomplicating. Because I do not have a suggestion as to how to proceed in the development of this caveat at this time, I will remark on it no further, but I do encourage discussion on it nevertheless!

Any thoughts? I'm no tournament officiator, so I figured it'd be constructive to take this idea to Smashboards to see what people who know more about the competitive scene than I do think about it. Thanks for reading, and let me know what you think!

 

[Raijinken]

Sounds a ton better than using percent.

Also this was a thing.

I still think it most reasonable to run a Pseudo-Sudden-Death using Handicap. But at this point, I'm just interested in anything but percent.

 

[Ansou]

You already mentioned this, but survivability is not only dependant on character weight. It's dependant on a multitude of factors that are matchup specific. I don't have a proposal for a good tiebreaker rule as I don't think we can craft one that is completely fair, but I don't really think that this is more fair than the standard percent lead while being more complicated. Or maybe I'm just biased because I main a lightweight character with insanely good recovery...

 

[Nintenpro]

The current solution isn't maybe the fairest solution but it's definitely the easiest to apply in a tournament.
Your idea is too complicated, during a match players should be able to easily know if they are winning or losing.

 

[Kaladin]

You already mentioned this, but survivability is not only dependant on character weight. It's dependant on a multitude of factors that are matchup specific. I don't have a proposal for a good tiebreaker rule as I don't think we can craft one that is completely fair, but I don't really think that this is more fair than the standard percent lead while being more complicated. Or maybe I'm just biased because I main a lightweight character with insanely good recovery...

In the status quo, % is used as a gauge of who is closer to death. This is an objective improvement on that system, as it's more accurate.

The current solution isn't maybe the fairest solution but it's definitely the easiest to apply in a tournament.
Your idea is too complicated, during a match players should be able to easily know if they are winning or losing.

Calculators are a thing. As for being able to tell, I think most players will have the intuition as to who gets launched farther at their respective %.

 

[wizrad]

I like the idea of changing the tiebreaker, but what about gimping? Characters like Little Mac and Doc die quite early despite their weights due to their poor recoveries.

 

[LightLV]

This helps to solve the issue of timeouts being biased by weight class, but adds the issue of it just adding another layer onto something that was already pretty simple.

Not opposed to it, but the chance it'll catch on is extraordinarily slim. There are a few arguments that can be made against it as well.

 

[SpaghettiWeegee]

I like the idea of changing the tiebreaker, but what about gimping? Characters like Little Mac and Doc die quite early despite their weights due to their poor recoveries.

I've deliberated on the implications of this and while it's true that many characters can be KOed long before being sent to blast zones if their trajectory is horizontal enough (e.g. not giving them additional height to assist their getting back to the stage) and gimping off the edge is a kill option/preference for many characters (Meta Knight, for example), giving gimp advantage to a defender assumes too much-- just because a character can gimp an opponent doesn't necessarily mean they will. To recontextualize, in the event of overtime, the loser is whomever is closest to being KOed off the top.

One of the major things that makes the percent system in Smash different is that it means different things for different characters. Since it's a record of damage taken, not health remaining, it's difficult to say when two characters have roughly even health in Smash. I would say it's analogous to how different characters have different hit point values in a MOBA, for example-- in Smash, higher weight is similar to higher max HP, since presumably heavier characters can survive for longer. We need to take this into account somehow when determining tiebreakers.

 

[LightLV]

I've deliberated on the implications of this and while it's true that many characters can be KOed long before being sent to blast zones if their trajectory is horizontal enough (e.g. not giving them additional height to assist their getting back to the stage) and gimping off the edge is a kill option/preference for many characters (Meta Knight, for example), giving gimp advantage to a defender assumes too much-- just because a character can gimp an opponent doesn't necessarily mean they will. To recontextualize, in the event of overtime, the loser is whomever is closest to being KOed off the top.

But that has the same assumption that Star KO is a threat to either opponent in the match. Just because they CAN die off the top doesn't mean they will, and for some characters and some matchups, death by gimping is just far more likely.

I see what you're trying to do, but both options are still biased, they're trying to best predict the outcome of the match but the harder you try, the more information you're going to need.


Consider a match with a lightweight at 70% and an above average or heavyweight character at 140%.

By normal rules, the lightweight easily wins. Taking weight into account, heavyweight may still lose. But taking rage, character matchup and stage into consideration, lightweight could be just as close to death as the heavyweight, with either character simply being one throw away from death. It's easy for a heavyweight to say "I was one grab away from winning", and the lightweight could say the same thing and be justified.

 

[SpaghettiWeegee]

I wouldn't argue that my proposed solution is perfect, just that it makes more sense for a game like Smash Bros. If you wanted to be absolutely objective you'd remove the threat of timeout entirely and just let all matches play out until one player loses all stocks, but it's generally accepted that this would lead to matches that drag on for much too long-- bad for tournaments if they can't conclude before the venue use period runs out.

SS is simple and intuitive and also more accurately meshes with the systems at play in Smash. It's not perfect, but perfection was an outcome foregone when the timer hit zero-- now we're reaching for the next best (accessible) thing.

EDIT: Also consider implementing a Rule of Inevitability wherein if one character is obviously unable to recover at the timer hitting zero (e.g. in freefall below the ledge) and the other is safe, regardless of percent, the safe player will be the winner. This better covers for gimping edge-cases while still not dipping into the realm of speculation.

 

[Raijinken]

My suggestion in the above-linked thread was to use handicap to simulate Sudden Death without bob-ombs. To better make it so that players have a reason to still fight while running to time, I'd suggest having a base handicap level (somewhere between 50 and 100%, I use 80% in most examples), which then gets added to their match-end percent, and rounded to the nearest handicappable value (this is easy math, and if that's too hard, a quick glance at a graph containing percent ranges and corresponding percents takes away even that little bit of effort). This can't eliminate flighty play (which is my only problem with stall strategies), but it at least encourages you to rack damage if not outright finish your opponent.

Obviously no system is unbiased, including percent-based or outright sudden death, but here's my reasoning:
Percent-added still acknowledges your success at hitting your opponent. A player ending the match at lower percent will start the tiebreaker with lower percent, unless the difference falls within the same rounding value.
It helps account for weight, which is often (but not always) correlated with a character's ability to run the clock evasively.
It lets players decide between gambling on overtime, or aiming for the finisher. It also gives the other player input to the outcome. i.e. Sonic can run away all day against Bowser, but doing so will put both of them into rage mode, and Bowser needs little to kill Sonic in that setting.

Obviously the drawbacks are that any time out (of which there are few to begin with) will run for a few extra minutes. This also can result in a player who somehow lived to 200 on a flighty character still being able to get a sudden death read (E3?) and win something that some may think they didn't deserve.

Personally, I value the decisive end to the fight, regardless of how it's determined. I think it makes the most sense to do something that leads to that end, rather than making a ruling based on how the fight might have gone with more time.

 

[SpaghettiWeegee]

Personally, I value the decisive end to the fight, regardless of how it's determined. I think it makes the most sense to do something that leads to that end, rather than making a ruling based on how the fight might have gone with more time.

Even though I think the flaws with going into overtime play need to be more carefully sussed out before such a system could be implemented, I actually would like to consider this briefly, but less from a gameplay perspective and more from a spectator one.

If the audience reaction to Hbox at EVO 2015 is any indication, the crowd hates playing lame, which the option of running the clock out facilitates. I don't have a solution for Melee since stock leads are so much more common that it's less likely than in Smash 4 that the winner of overtime comes down to percent difference, but in the case of Smash 4, I think an overtime duel might actually not be a terrible thing to watch. For one, it encourages decisive play-- running the clock down only offsets the inevitable and also kills momentum, making the player in the lead more likely to choose aggressive options instead of camping their opponent out to push their advantage before the reset messes with it. Additionally, the crowd loves seeing a definitive victory, and we all know that only a clean KO will be considered definitive.

The major problem with going overtime like this, however, is that it still lacks push, which was the cause of the overtime in the first place.

Funnily enough, Smash actually deals with this major problem by default. The reason the Bob-ombs start bombarding the stage after a certain number of seconds pass in Sudden Death is to encourage both players to start playing hyperaggressively to avoid the winner being decided by RNG, an outcome that probably few serious players want to happen.

Now I'm not saying that we should use Default Sudden Death, but... ¯\_(ツ)_/¯

 

[Raijinken]

Force the match to Norfair?

I agree with the lack of push, but (and I think we agree on this) that lack of push is present in the default rules anyway. With a stock lead, the winning player never has a reason to go on the offense (this is just gonna happen). With our current rules, though, the player with lower percent has basically no reason to do anything but run, assuming that's an option for their character on the selected stage.

That said, and I obviously lack evidence to back this up because the system isn't run, but I think taking the above system and running it 1s 3m would be fine. Keep adding percent until someone wins. This will either result in one player getting gutsy and finally risking the 50/50 he asked for, or result in both players staring at each other until the TO gets tired of them both deliberately stalling the match.

And as an occasional TO, I think two players deliberately delaying the tournament in the easy case of a decisive blow is pretty easily grounds for disqualification (in a slightly exaggerated sense, though I'd certainly DQ anyone deliberately stalling the tournament).

 

[Megamang]

I don't agree with " With our current rules, though, the player with lower percent has basically no reason to do anything but run, assuming that's an option for their character on the selected stage."

Running away without fighting is how you throw a lead. With our current system, when you are ahead you don't have to approach. Thats just the nature of this (and many) games. You are trying to change who is ahead at any given moment, but the issue you dislike will remain, it will just be more convoluted.

Your system also assumes that a character's likeliness to land a smash is linearly related to his weight which... no.

 

[Raijinken]

I don't agree with " With our current rules, though, the player with lower percent has basically no reason to do anything but run, assuming that's an option for their character on the selected stage."

Running away without fighting is how you throw a lead. With our current system, when you are ahead you don't have to approach. Thats just the nature of this (and many) games. You are trying to change who is ahead at any given moment, but the issue you dislike will remain, it will just be more convoluted.

Your system also assumes that a character's likeliness to land a smash is linearly related to his weight which... no.

Did you watch StaticManny at EVO?

It's true, there is some reason to fight when you're ahead. You can secure a win much faster, especially if you have a distinct skill advantage. But on paper and in practice, there are times when it is just best to keep running the clock. And a lot of those occasions stem from the fact that only one player is under pressure, and in some cases (see again: Sonic), it becomes very easy for the player with the lead to maintain that lead by avoiding the fight. And even worse, in some cases, the other player has virtually nothing they can do but hope for an input error on the flighty player's end.

Basically, flighty play screws with the usual flow, and power dynamic, of the game. The inability to catch an opponent who is trying to evade the fight is the very reason we ban stages that are easy to circle camp. If we're going to enable the issue out of necessity rather than desire, then our resolution methods should be designed as close as possible to discourage the "strategy". A pure percent system lets anyone with a 1% lead force their opponent to approach, regardless of that character's tools. The fact that years into the game's lifespan the percent rule is still not used by the game itself indicates to me that such a dynamic is not intended or desired.

 

[Ghostbone]

Don't the official online tournaments use "damage dealt" to solve ties?

For 1v1, that's the same as the % rule, except it takes into account previous stocks (which is worse since how long it took you to take the first stock shouldn't give you the win, plus it's not easily accounted for by players/the audience)

For real, the % rule is fine, no rule is perfect and % is the simplest. It might not be a measure of how close someone is to death, but it is a measure of how far away they are from the starting point.

 

[Raijinken]

Don't the official online tournaments use "damage dealt" to solve ties?

For 1v1, that's the same as the % rule, except it takes into account previous stocks (which is worse since how long it took you to take the first stock shouldn't give you the win, plus it's not easily accounted for by players/the audience)

For real, the % rule is fine, no rule is perfect and % is the simplest. It might not be a measure of how close someone is to death, but it is a measure of how far away they are from the starting point.

Distance from the starting point is irrelevant when all that matters is who dies last.

Or, if that logic is to be applied, then measuring distance from center-stage is actually a more valid, yet just as simple, ruling.

 

[nebulark]

In my opinion the best tiebraker is one the discourages timing out itself. You also want to have the whole match influecing the tiebraker and not just a small fraktion. Otherwise it can be too dependent on luck or favoring a(n) (uncompetetive) skill, such as dodging bombs in sudden death.

Running away is the best way to time out, so you have to find a way to discourage it. There are many ways, but here is my (utopian solution.
The player that controlled the stage longer wins. A player has stagecontrol as long as he is closer to the center than his opponent. The center is the bottommost point on a plattform/ground and an equal distance away from both side blastzones.

However, this is where reality kicks in. At the moment we have no easy to use way to measure this. So think the question that needs to be answered is the following. What is the best tiebreaker that discurages running away and is easily messureable?
So either we find a thing that can be measured easily or a new measuring method. (Maybe a future mod or something like this http://smashboards.com/threads/hardware-enhanced-melee-proof-of-concept.416055/")

 

[nodle]

If the only problem with SD is the bombs, why not just jump over to special smash for a 300% game. Or turn on handicap, and dial up the damage?


In any case, I think the best solution to the Tiebreaker issue is to permit fewer tie breakers. I propose that tiebreakers only be permitted when a match ends in a tie. Allow individual games to Tie. Basically, A match would be defined as a certain number of games, the player with the most games won at the end of a match wins the match. Using this system would strongly discourage time out play, because in order to do it successfully, you would basically have to win game 1, then force a tie in both game 2 and game 3. Tiebreakers would need to be some kind of single stock match.

 

[Raijinken]

If the only problem with SD is the bombs, why not just jump over to special smash for a 300% game. Or turn on handicap, and dial up the damage?


In any case, I think the best solution to the Tiebreaker issue is to permit fewer tie breakers. I propose that tiebreakers only be permitted when a match ends in a tie. Allow individual games to Tie. Basically, A match would be defined as a certain number of games, the player with the most games won at the end of a match wins the match. Using this system would strongly discourage time out play, because in order to do it successfully, you would basically have to win game 1, then force a tie in both game 2 and game 3. Tiebreakers would need to be some kind of single stock match.

That, largely, was my suggestion. But since simply dialing the damage to 300% doesn't acknowledge the actual progress made during the match, I added the bit where you take a base handicap (near but not at kill percentage, matchup/character dependent) and then add the end-match percent, then round. You get similar results, without entirely invalidating a player's efforts during the match itself. A Sonic can stall all he wants against Bowser, but unless he tries to tack on more percent, he may enter the tiebreaker a ftilt away from death while Bowser takes far more to die. It adds some choice to both player's play while also discouraging pure flighty play.

 

[Touchebag]

That, largely, was my suggestion. But since simply dialing the damage to 300% doesn't acknowledge the actual progress made during the match, I added the bit where you take a base handicap (near but not at kill percentage, matchup/character dependent) and then add the end-match percent, then round. You get similar results, without entirely invalidating a player's efforts during the match itself. A Sonic can stall all he wants against Bowser, but unless he tries to tack on more percent, he may enter the tiebreaker a ftilt away from death while Bowser takes far more to die. It adds some choice to both player's play while also discouraging pure flighty play.

This is pretty much the gist of what we discussed in the previous thread I believe. It would be nice if we could actually test this in tournament settings but even if we get a larger tournament to try it time outs are still far too uncommon overall. It would have to be evaluated over several months before we know if it is as good a system in reality as it sounds.

 

[Raijinken]

This is pretty much the gist of what we discussed in the previous thread I believe. It would be nice if we could actually test this in tournament settings but even if we get a larger tournament to try it time outs are still far too uncommon overall. It would have to be evaluated over several months before we know if it is as good a system in reality as it sounds.

One could get an approximation by running a series of 2 stock 3 minute matches. This puts timeouts well within the realm of feasibility, and while it doesn't necessarily reflect the amount of effort it takes to time out the game to begin with, it can give some practical information about the actual practicability of using handicap, and may give players an idea of how it affects their mindset when a timeout is looming.

 

[nodle]

That, largely, was my suggestion. But since simply dialing the damage to 300% doesn't acknowledge the actual progress made during the match, I added the bit where you take a base handicap (near but not at kill percentage, matchup/character dependent) and then add the end-match percent, then round. You get similar results, without entirely invalidating a player's efforts during the match itself. A Sonic can stall all he wants against Bowser, but unless he tries to tack on more percent, he may enter the tiebreaker a ftilt away from death while Bowser takes far more to die. It adds some choice to both player's play while also discouraging pure flighty play.

The important bit of my suggestion is the change to when tiebreakers occur. Any thoughts on my proposed change to match and game resolution. To summarize my changes...

1. A match is defined as a finite and predeterined number of games.
2. A game can be ended either with Player 1 winning, player 2 winning, or players tieing
3. The winner of a match is the player with the most wins at the end of the match. In the event of a tie, an additional single stock tiebreaker game is played to decide the winner.

...The exact rules of the tie breaker game still needs to be determined. To be clear, a game is considered tied if both players lose there last life simultaneously, or the game times out. This rule set has quite a few advantages over the current ruleset and your proposal...

1. It doesn't set arbitrary timeout rules. the game says the players tied, now we can agree.
2. It disincentives timeout strategies because timeouts are now unwinnable. In order to time out effectively, you first have to gain a game lead by actually winning and manage to time out multiple games.
3. Unlike your proposal, We can basically only have a single Tiebreaker game. With your system, players could potentially play six games in a single 3 game set. This could cause problems with tournament planning.
4. The only thing that needs to be tracked is the results of the games which is tracked anyway. We no longer need to care about damage except as it pertains to killing and getting killed, its true purpose in the game.
5. It seems far less biased than any other system I've seen. This system doesn't care how you play your character. It only cares about your ability to actually win games.

 

[Raijinken]

The important bit of my suggestion is the change to when tiebreakers occur. Any thoughts on my proposed change to match and game resolution. To summarize my changes...

1. A match is defined as a finite and predeterined number of games.
2. A game can be ended either with Player 1 winning, player 2 winning, or players tieing
3. The winner of a match is the player with the most wins at the end of the match. In the event of a tie, an additional single stock tiebreaker game is played to decide the winner.

...The exact rules of the tie breaker game still needs to be determined.

Ah yes, sorry, my reply was addressing the 300% handicap part of your post.

I do rather like the idea of breaking ties at the set level instead of game level. It takes no longer than breaking the tie at the game level, and potentially less time (breaking at game level in a total percent tie, or most other discussed methods, takes full game time plus the tiebreaker, while breaking at set level doesn't take that extra time at all if a player gets 2 wins (3 in a bo5)).

I wonder, though: What are the pros and cons of determining the winner by win count? If a best 2 of 3 has two ties and one win (somehow), who wins? It's almost always a player's conscious choice to run the clock (I say almost because I've had matches in which my opponent and I were both too cautious of an approach, since it was Robin vs Palutena, so while neither of us wanted to stall to time, neither of us could safely approach). Should that choice be discounted and the match treated as a zero-point game? I would agree to that instantly, and I know many others who would, as well.

As for the tiebreaker's rules itself, that match has to have a decisive end, even if it goes to time. I'd be inclined to say 1 stock 3 minutes on a pre-defined stage, since as it's a tie situation, neither player has the "win" to determine who gets to strike or pick or ban stages. We could add handicap levels if we want (50% perhaps), or come up with some other way to make the match come to a stock-taken close. For that matter, we could make it 1 stock, 150% handicap, no time, and require the match go to Norfair just to give players some pressure to fight each other.

I like the idea. I just think it's important to end decisively with the game's support. We can justify dodging Sudden Death since the random nature is undesirable, but anything past that, I'm personally opposed to ruling against the game's screen.

 

[nodle]

I do rather like the idea of breaking ties at the set level instead of game level. It takes no longer than breaking the tie at the game level, and potentially less time (breaking at game level in a total percent tie, or most other discussed methods, takes full game time plus the tiebreaker, while breaking at set level doesn't take that extra time at all if a player gets 2 wins (3 in a bo5)).

I updated my post to include some pros. I can't really think of any cons.

I wonder, though: What are the pros and cons of determining the winner by win count? If a best 2 of 3 has two ties and one win (somehow), who wins? It's almost always a player's conscious choice to run the clock (I say almost because I've had matches in which my opponent and I were both too cautious of an approach, since it was Robin vs Palutena, so while neither of us wanted to stall to time, neither of us could safely approach). Should that choice be discounted and the match treated as a zero-point game? I would agree to that instantly, and I know many others who would, as well.

We wouldn't play best 2 of 3 anymore, we would just play 3 Game Matches. If a player got 2 wins, that player wins immediately, but 2 wins isn't essential for winning a match. The player with the most won games wins the match. In your example, the player with 1 won game wins. Anything else is unfair to that player.

As for the tiebreaker's rules itself, that match has to have a decisive end, even if it goes to time. I'd be inclined to say 1 stock 3 minutes on a pre-defined stage, since as it's a tie situation, neither player has the "win" to determine who gets to strike or pick or ban stages. We could add handicap levels if we want (50% perhaps), or come up with some other way to make the match come to a stock-taken close. For that matter, we could make it 1 stock, 150% handicap, no time, and require the match go to Norfair just to give players some pressure to fight each other.

Handicaps (SDs) aren't really fair though. I agree that we should just predefine a tiebreaker stage. My vote would go to Smashville as a very popular neutral.

I like the idea. I just think it's important to end decisively with the game's support. We can justify dodging Sudden Death since the random nature is undesirable, but anything past that, I'm personally opposed to ruling against the game's screen.

A major advantage of my system is that it agrees with the games screen. The game says no-contest. My system handles it without adding arbitrary tie breakers. It just accepts the tie and moves on.

 

[Raijinken]

I updated my post to include some pros. I can't really think of any cons.


We wouldn't play best 2 of 3 anymore, we would just play 3 Game Matches. If a player got 2 wins, that player wins immediately, but 2 wins isn't essential for winning a match. The player with the most won games wins the match. In your example, the player with 1 won game wins. Anything else is unfair to that player.


Handicaps (SDs) aren't really fair though. I agree that we should just predefine a tiebreaker stage. My vote would go to Smashville as a very popular neutral.

A major advantage of my system is that it agrees with the games screen. The game says no-contest. My system handles it without adding arbitrary tie breakers. It just accepts the tie and moves on.

My vote would go anywhere but Smashville since I don't think we should give Sheik a free counterpick in a tiebreaker.

But that's part of the issue there. Without a winner, they could either Rock Paper Scissors, or we have to pick a stage. Rock Paper Scissors (or seed priority, perhaps) may be the best way to use that.

 

[Touchebag]

I updated my post to include some pros. I can't really think of any cons.

We wouldn't play best 2 of 3 anymore, we would just play 3 Game Matches. If a player got 2 wins, that player wins immediately, but 2 wins isn't essential for winning a match. The player with the most won games wins the match. In your example, the player with 1 won game wins. Anything else is unfair to that player.

I proposed this system in the last thread as well and the general response was pretty similar (positive). This was also how I intended it should work (though I don't think we actually discussed that part). The main benefits of this is that the only situations where a tiebreaker would occur is three ties (very unlikely) or one tie, one win each (slightly less unlikely) in a Bo3.

The downside is that we would have to run another game to decide a tie (as opposed to it being decided straight away from percentage). Add in the fact that a set in which one player wins game 1 and times out game 2 with a percentage lead would also add another game to the set as opposed to the current system we could be dealing with potential time losses here.

Handicaps (SDs) aren't really fair though. I agree that we should just predefine a tiebreaker stage. My vote would go to Smashville as a very popular neutral.

But how do we guarantee this tiebreaker actually ends in the allotted three (or whatever) minutes? There needs to be some incentive to end the match quickly. If both players start at kill percent, timing out that game as well becomes increasingly difficult. And how do we deal with the (however unlikely) situation where that game does time out? Because it will happen.

My vote would go anywhere but Smashville since I don't think we should give Sheik a free counterpick in a tiebreaker.

But that's part of the issue there. Without a winner, they could either Rock Paper Scissors, or we have to pick a stage. Rock Paper Scissors (or seed priority, perhaps) may be the best way to use that.

Also this.

A major advantage of my system is that it agrees with the games screen. The game says no-contest. My system handles it without adding arbitrary tie breakers. It just accepts the tie and moves on.

The game doesn't call it a tie though. It forces the SD to decide the match. But I see your point. I agree that a tie should be a tie and not a win to whoever is best at (ab)using an arbitrary community rule. As a bonus this could also potentially remove the need for suicide clauses.

 

[Raijinken]

The game doesn't call it a tie though. It forces the SD to decide the match. But I see your point. I agree that a tie should be a tie and not a win to whoever is best at (ab)using an arbitrary community rule. As a bonus this could also potentially remove the need for suicide clauses.

Suicide clauses need revision anyway. As sorry as I feel for Bowser players, he was pointedly patched to die first. There's no logical reason to rule against a deliberate change.

Of course, in the case of it going to sudden death, I agree completely.

But yeah. Ruling against the game should be done as little as possible, which is why I think the tiebreaker needs to have a mandatory stock-taken conclusion. I'm a bit extreme, but I'd go so far as to threaten double-disqualification if the tiebreaker is drawn out.

The only way to prevent the tiebreaker from going to time is to remove time, but then we need either extremely strict enforcement/observation (the occasion is rare enough that we could reasonably have a TO watch each tiebreaker to see if a player is deliberately stalling the tiebreaker), or to come up with another way to ensure the game ends. Conveniently, the easiest way to do so is completely and utterly frowned upon.

 

[nodle]

My vote would go anywhere but Smashville since I don't think we should give Sheik a free counterpick in a tiebreaker.

But that's part of the issue there. Without a winner, they could either Rock Paper Scissors, or we have to pick a stage. Rock Paper Scissors (or seed priority, perhaps) may be the best way to use that.

Fair. I guess Rock paper scissors works. It should definitely require going to a neutral. It should probably require going to a neutral the deciding player hasn't won on.

The downside is that we would have to run another game to decide a tie (as opposed to it being decided straight away from percentage). Add in the fact that a set in which one player wins game 1 and times out game 2 with a percentage lead would also add another game to the set as opposed to the current system we could be dealing with potential time losses here.

Yes, but that's also an upside. It strongly disincentives players from timing out the match. You don't gain anything from it anymore, you just play another game.

But how do we guarantee this tiebreaker actually ends in the allotted three (or whatever) minutes? There needs to be some incentive to end the match quickly. If both players start at kill percent, timing out that game as well becomes increasingly difficult. And how do we deal with the (however unlikely) situation where that game does time out? Because it will happen.

Have a quick 30 second foot race....

Honestly, there are many options. My favorite is to just decide the tiebreaker with an actual SD. Players had multiple games to decide a winner, and they ended up not playing (i.e. timing out) so screw it, lets get a winner. My rules would be...

1. A match is defined as a finite and predeterined number of games.
2. A game can be ended either with a player 1 win, player 2 win, or tie if the game goes to sudden death.
3. If a player is ever leading by more games than there are games remaining in the match, the match ends immediately.
   
4. The winner of a match is the player with the most wins at the end of the match.
5. If a match ends in a tie, a 1 stock 3 minute Tiebreaker game is played. (Insert appropriate map rule here). Follow in game rules regarding Sudden Death.

...Honestly, do we care if sudden death is noncompetitive, when players would only enter sudden death after getting at least one time-out/double-death, not have a single player win the other games, then time-out/double-death again in the tiebreaker game. So few games would go to SD, I'm fine using it.

 

[Raijinken]

Fair. I guess Rock paper scissors works. It should definitely require going to a neutral. It should probably require going to a neutral the deciding player hasn't won on.


Yes, but that's also an upside. It strongly disincentives players from timing out the match. You don't gain anything from it anymore, you just play another game.


Have a quick 30 second foot race....

Honestly, there are many options. My favorite is to just decide the tiebreaker with an actual SD. Players had multiple games to decide a winner, and they ended up not playing (i.e. timing out) so screw it, lets get a winner. My rules would be...

1. A match is defined as a finite and predeterined number of games.
2. A game can be ended either with a player 1 win, player 2 win, or tie if the game goes to sudden death.
3. If a player is ever leading by more games than there are games remaining in the set, the match ends immediately.
   
4. The winner of a match is the player with the most wins at the end of the match.
5. If a match ends in a tie, a 1 stock 3 minute Tiebreaker game is played. (Insert appropriate map rule here). Follow in game rules regarding Sudden Death.

...Honestly, do we care if sudden death is noncompetitive, when players would only enter sudden death after getting at least one time-out/double-death and then doing it again in the tiebreaker game. So few games would go to SD, I'm fine using it.

I think you're meaning "set" in a lot of places where you're saying "match," which is making it a bit hard to sort out what you're saying.

We're starting to disagree on the implementation, probably because I don't see why any stage that was fair to play on any other round should be considered unfair for the tiebreaker (or round 1, as the case were). But I do like the idea of breaking the tie at the set level.

 

[nodle]

I think you're meaning "set" in a lot of places where you're saying "match," which is making it a bit hard to sort out what you're saying.

We're starting to disagree on the implementation, probably because I don't see why any stage that was fair to play on any other round should be considered unfair for the tiebreaker (or round 1, as the case were). But I do like the idea of breaking the tie at the set level.

I'm using match because its the word me and my friends use. Here is ruleset in your vocab...

1. A set is defined as a finite and predetermined number of games.
2. A game can be ended either with a player 1 win, player 2 win, or tie if the game goes to sudden death.
3. If a player is ever leading by more games than there are games remaining in the set, the set ends immediately.
4. The winner of a set is the player with the most wins at the end of the set.
5. If a set ends in a tie, a 1 stock 3 minute Tiebreaker game is played. Follow in game rules regarding Sudden Death.

...Regarding stage selection. I don't think the higher-seed/rock-paper-scissors-winner should be allowed to get a free stage counter pick. The point of neutral stages is that they don't bias the match, so the tiebreaker game should go to a neutral stage. A normal rule in Smash is that players can't pick a stage they have won on without mutual agreement of players.

 

[Ansou]

One problem here is that there will almost always be one player that benefits from timing out the game. It's extremely hard to make a ruleset that makes timeouts unfavourable for both players. If we have something similar to sudden death, the losing player would benefit from a timeout. If a match can end in a tie, the player with a match lead would benefit from a timeout. I really don't have a good solution to this problem.

 

[nodle]

One problem here is that there will almost always be one player that benefits from timing out the game. It's extremely hard to make a ruleset that makes timeouts unfavourable for both players. If we have something similar to sudden death, the losing player would benefit from a timeout. If a match can end in a tie, the player with a match lead would benefit from a timeout. I really don't have a good solution to this problem.

There isn't one. The best we can do is make timeouts unimpactful to reduce the incentives for both players. With current rule set. A player is always incentivized to time out if they have a percentage lead. When they succeed, they gain an entire game of advantage on their opponent. With my system, time outs do almost literally nothing. You can't win at all if you only time out. If you manage to timeout every game and then timeout the Tiebreaker, you still would have to win SD. At all times, the losing player (in damage) has a small incentive to time out a game and the winning player (in games) has a small incentive to time out a game. A player is always better off winning a game then timing out a game. Short of double DQs for time outs, I don't think we can incentivize them less.

Once we work out stage selection for Tiebreaker games, I think this system will be the best we can get.

 

[Ansou]

There isn't one. The best we can do is make timeouts unimpactful to reduce the incentives for both players. With current rule set. A player is always incentivized to time out if they have a percentage lead. When they succeed, they gain an entire game of advantage on their opponent. With my system, time outs do almost literally nothing. You can't win at all if you only time out. If you manage to timeout every game and then timeout the Tiebreaker, you still would have to win SD. At all times, the losing player (in damage) has a small incentive to time out a game and the winning player (in games) has a small incentive to time out a game. A player is always better off winning a game then timing out a game. Short of double DQs for time outs, I don't think we can incentivize them less.

Once we work out stage selection for Tiebreaker games, I think this system will be the best we can get.

Well yeah, your system makes a lot more sense than character weight being a part of it or looking at how close you are from centerstage. I guess this also makes more sense than percent lead determining tiebreakers because it's more okay to win one game and time the other two games out than to get some percent on the opponent and time out the rest of the game. But still, anyone that is significantly more damaged than the opponent will want to time out the game since it is better to tie a match than to lose a match.

 

[nodle]

Another interesting feature of my proposed rule set is that Even-Game Sets are well defined and well behaved. We could, for example switch to 3-stock 8-min 2-game sets while maintaining about the same round time for tournaments.

EDIT: Saying anyone with significant damage would want to time out the game is a bit misleading. I don't think its a problem because it wouldn't be the players strategy. It would just be a tactic that they used for that specific game. In order for it to be effective, the player both has to recognize that they are too far behind to go for the win, be ahead or tied in game wins, and actually be capable of stalling out the game even though they are losing by a lot. It seems like its not a real concern to me.

 

[Big O]

What if instead of having ties count as a draw, ties counted as double losses. Sets could then be determined by whoever has lost 2/3 times or 3/5 times. In the event that both players reach 2 L's in a Bo3 or 3 L's in a Bo5 simultaneously, both players are eliminated.

This wouldn't stop all timeout draws, but it does minimize the impact of timeouts on tourney time. Timeouts on the last game eliminate both players, timeouts during any "match point" games would eliminate the losing player. Only draws in the last match after a bracket reset in grand finals would require a tiebreaker. It leaves timeouts as a viable option to close out a set when ahead in games, but containing timeouts to just these games and having an sufficiently long timer doesn't make it practical to abuse.

 

[Ghostbone]

What if instead of having ties count as a draw, ties counted as double losses. Sets could then be determined by whoever has lost 2/3 times or 3/5 times. In the event that both players reach 2 L's in a Bo3 or 3 L's in a Bo5 simultaneously, both players are eliminated.

This is impossible in double elimination, you can't have both players from winners bracket going to losers bracket.

And again, this just promotes the losing player to go for a timeout, because having both players lose is far better than you losing and your opponent winning.

 

[Big O]

This is impossible in double elimination, you can't have both players from winners bracket going to losers bracket.

And again, this just promotes the losing player to go for a timeout, because having both players lose is far better than you losing and your opponent winning.

What if both players are no shows in a winner's bracket match? That doesn't sound like an impossible scenario either, so I doubt that hasn't happened and hasn't been resolved before. Set draws by timeout could be handled similarly.

I don't really see how it would favor the losing player over the winning player with the proposed system. If they are losing in games, timeouts are the same as a draw until match point, which would then lead to elimination. For example, if you are down 1 game in a Bo3 you lose immediately following a timeout, while in a Bo5 you would need to reverse 3-0 them afterwards (and never get timed out too). In both of those situations the losing player during the match gets the same result after a timeout (elimination for Bo3 and reverse 3-0 to win for Bo5).

If they are even in games, it is the same as a draw with the added danger of being eliminated should you continue to keep having timeouts. If game 1 goes to time in a Bo3, game 2 would effectively match point. If game 1 goes to time in a Bo5, it would effectively transition to a Bo3 set. A timeout at game 3 of an even set would make game 4 match point. At this point, you could say that my proposed system would give the losing player an advantage compared to the % tiebreaker, but if you gave counterpicking rights to the player with the % lead after a timeout I think the overall result would be the same. The "loser" according to the % lead would have to beat the winner on his cp to win the set. This just makes that happen game 2 instead of game 3.

All that said, there is no reason to switch from % based tiebreakers if people feel it is a good enough way to determine who is winning. For those that don't feel it is as fair as it could be or feel timeouts should be discouraged, it is interesting to brainstorm solutions that are to our liking. I just thought it would be interesting to have a system that potentially makes timeouts actually cut down tourney length and discourages them at the same time.

 

[nodle]

@ Big O

Our systems are virtually identical. Here is your ruleset in my language...

1. A set is defined as a finite and predetermined number of games.
2. A game can be ended either with a player 1 win, player 2 win, or tie if the game goes to sudden death.
3. If a player is ever leading by more games than there are games remaining in the set, the set ends immediately.
4. The winner of a set is the player with the most wins at the end of the set.
5. If players have an equal number of wins, both players lose.

All you've done is trade a tiebreaker match for weird 'both players lose' tech. This would lead to weird tournament brackets where players are randomly given buys because both of their possible challengers lost. Do you really think it would be fair for two players to both lose because they couldn't quite finish their last game? If they both have 2 wins and both players have been playing aggressively all game, but the kill has remained elusive, you really think both players deserve to lose? Remember, the alternative is as simple as a 3 minute tiebreaker game.

EDIT: The system outlined above isn't quite identical to your system. There is no easy way to translate your system from a system that cares about losses to a system that cares about wins. The question I pose still stands. Here is your system...

1. A set is defined as an odd, finite, and predetermined K number of games.
2. A game can be ended either with a player 1 loss, player 2 loss, or mutual loss if the game goes to sudden death.
3. If a player ever has floor(K/2)+1 losses, the set ends immediately
4. The winner of a set is the player with the fewest losses at the end of the set.
5. If players have an equal number of losses, both players lose.

 

[Charey]

What if both players are no shows in a winner's bracket match? That doesn't sound like an impossible scenario either, so I doubt that hasn't happened and hasn't been resolved before. Set draws by timeout could be handled similarly.

I don't really see how it would favor the losing player over the winning player with the proposed system. If they are losing in games, timeouts are the same as a draw until match point, which would then lead to elimination. For example, if you are down 1 game in a Bo3 you lose immediately following a timeout, while in a Bo5 you would need to reverse 3-0 them afterwards (and never get timed out too). In both of those situations the losing player during the match gets the same result after a timeout (elimination for Bo3 and reverse 3-0 to win for Bo5).

If they are even in games, it is the same as a draw with the added danger of being eliminated should you continue to keep having timeouts. If game 1 goes to time in a Bo3, game 2 would effectively match point. If game 1 goes to time in a Bo5, it would effectively transition to a Bo3 set. A timeout at game 3 of an even set would make game 4 match point. At this point, you could say that my proposed system would give the losing player an advantage compared to the % tiebreaker, but if you gave counterpicking rights to the player with the % lead after a timeout I think the overall result would be the same. The "loser" according to the % lead would have to beat the winner on his cp to win the set. This just makes that happen game 2 instead of game 3.

All that said, there is no reason to switch from % based tiebreakers if people feel it is a good enough way to determine who is winning. For those that don't feel it is as fair as it could be or feel timeouts should be discouraged, it is interesting to brainstorm solutions that are to our liking. I just thought it would be interesting to have a system that potentially makes timeouts actually cut down tourney length and discourages them at the same time.

So what if we are in grands with a reset bracket and that ends in a timeout making both players lose grands?