Endless Stall Clause

[Zarel]

This thread is to discuss what an Endless Stall Clause, a set of rules for forcing a stall-v-stall game to tie, might look like.

This is a new thread ONLY for force-tie. If your suggestion involves a player making any direct choice about whether or not to end the game in a draw, (or if it requires an optional draw to be usable), it belongs in the other thread: https://www.smogon.com/forums/threads/hammering-out-a-fifty-turn-rule-proposal.3610169

I'm serious about this, and will be deleting any post suggesting a tie (this isn't censorship because you can still post in the other thread). My hope is that we'll be able to discuss ideas without that distraction.

 

[Zarel]

On the other thread, I suggested forcing a tie in the following situations:

- turn 1000
- 200 turns without a KO
- 10 turns of no PP use or net damage dealt

A lot of people didn't like "200 turns without a KO", so let me suggest something more moderate: 400 turns without a KO.

Some people liked "10 turns of no PP use" and some people didn't; I'm still leaning towards that being left in.

My fear is still... 400 turns is around 3 hours. That's better, but not ideal. 10 turns/PP with players actively trying to drag it out is still not going to end in 1000 turns.

Ideally, we should have some way to detect measurable progress. KO is a good idea, but KOs are too infrequent.

 

[UltiMario]

At 10 turns of no pp usage in a stall mirror you could easily go for plays where you force your opponent to attack in ways that create major disadvanges for them. That rule would become a "win" condition once sufficiently behind, potentially forcing your opponent into situations between choosing a tie or to make losing plays.

There's only two places where an X turns without PP usage can lie: low enough to where its a "win" condition or high enough to where it's completely and totally irrelivant. It's a complete waste of time to try and implement into the ruleset.

 

[ABR]

I addressed this in the other thread but UltiMario is entirely right about this style of rule and it's faults. Similarly, going by KO's either infringes on regular gameplay or goes far beyond a raw 1000. Many endless situations can be created with only 3 mons on each side.

It seems that 1000 turns is a good turn limit, just that real life time is still a concern, and that is a very valid concern. As such, maybe PS' timer can also be adjusted in some form. I'm not too familiar with the historical versions of our timer or what we can and can't practically do, but ideally we can reform it in some way.

One possibility is to just incrementally decrease timer every x turns or so, where x is 100 or greater. It can also initially kick in later than turn 100 if you want. I'm also not sure on the specific amounts we could cut it by but I'm sure there's a solution out there.

Another option is to have the vgc-style total player timer after a certain point. So, let's say PS timer remains normal for the first x turns (500? can be as low as 250). At turn x, it switches to "each player has a total of y minutes left." Again I'm not sure on a specific amount but maybe somewhere in the 30-60 minute range is good.

Thinking about turns themselves will never fall in a proper middle ground so we should really be looking to timer solutions on top of something like a 1000 turn rule, if not instead of.

The incremental timer decrease needs to be alongside a 1000 turn rule but the total player timer can be standalone.

 

[Jellicent]

10 turns of no PP use or net damage dealt:
Full paralysis and sleep turns are times where no damage is dealt or PP used that are out of the player's control. Not that both players are anywhere near likely to full para 10 times in a row, but if they're in a chain of double-switches, one player attempting to attack and fping might set off another round of switches before the opportunity to make that attack again presents itself. I think 10 is too close to reasonable playing, especially when luck can be a factor with it. Going much higher makes it worse at preventing lengthy stalling, though (just switch x times, attack once, and go back to switching). This isn't one I'd want at all, tbh.

X turns without a KO:
Chansey PP stalling each other out is a legitimate play in RBY that pops up every so often. They might have 112 PP, and will likely be paralyzed (so 1/4 chance of using no PP in a turn), making just that scenario take potentially 150 or so turns before they start Struggling. This could be due to both players having an offensive strategy to execute that requires the opposing Chansey neutralized (or at least out of TWave PP) first. I'd say 200 turns cuts way too closely to that (and other strategies like Wrap or even just holding on to basically KOed mons as death fodder or to keep sleep / freeze clause active), though 400 is pretty generous if the goal is to prevent multi-hour stallfests. The grey between kinda runs into both problems, where this could spoil a good lengthy game while doing little to prevent the actual problem. You can't really measure progress in KOs alone, so I don't think this one should be enforced either.

Turn 1000:
I'd rather see this drop to whatever the max possible on-cart is, if anyone ever figured out what that'd be. Either way, this seems like a totally acceptable time to just force the game to end. Is there a reason we would favor tying over awarding a winner after all that time, though? In VGC, a winner's determined at the end of timeouts in a way that I think would make sense for us as well:

• If both players run out of “Your Time” on the same turn, the game will automatically determine the winner using the following conditions in hierarchical order. This should also be applied by the tournament staff for any game that is not finished when match time is called and the final turn has resolved:
o Whoever has the most Pokémon remaining wins.
o If both players have the same number of Pokémon remaining, the player with the highest ratio of total HP remaining / total maximum HP for all Pokémon in their party wins.
▪ This is found using the following equation: (Current HP of remaining Pokémon) / (Maximum HP of all Pokémon brought to battle)
• If there is no clear winner from conditions 1 or 2, then the player with the highest total HP remaining wins.
• If there is still no clear winner, the game will be a tie.

With forced ties, the best move for a player in a situation where they can't win is to start stalling to the turn limit, turning it into a pseudo-victory condition that losing players will shoot for. Wasting hundreds of turns becomes a much bigger gamble if you need to maintain an advantage over your opponent during that time. Calling the game at the turn limit is just another (needlessly time-consuming and risky) victory condition for the player in a winning position.

 

[Zarel]

One possibility is to just incrementally decrease timer every x turns or so, where x is 100 or greater. It can also initially kick in later than turn 100 if you want. I'm also not sure on the specific amounts we could cut it by but I'm sure there's a solution out there.

Another option is to have the vgc-style total player timer after a certain point. So, let's say PS timer remains normal for the first x turns (500? can be as low as 250). At turn x, it switches to "each player has a total of y minutes left." Again I'm not sure on a specific amount but maybe somewhere in the 30-60 minute range is good.

These work, but I hope you understand that this gives basically no time to execute a strategy. You won't be able to use one attack every 100 doubleswitches in this situation.

This will make stall-v-stall end in timeouts rather than playing to a win.

I'd rather see this drop to whatever the max possible on-cart is, if anyone ever figured out what that'd be. Either way, this seems like a totally acceptable time to just force the game to end. Is there a reason we would favor tying over awarding a winner after all that time, though? In VGC, a winner's determined at the end of timeouts in a way that I think would make sense for us as well:

• If both players run out of “Your Time” on the same turn, the game will automatically determine the winner using the following conditions in hierarchical order. This should also be applied by the tournament staff for any game that is not finished when match time is called and the final turn has resolved:
o Whoever has the most Pokémon remaining wins.
o If both players have the same number of Pokémon remaining, the player with the highest ratio of total HP remaining / total maximum HP for all Pokémon in their party wins.
▪ This is found using the following equation: (Current HP of remaining Pokémon) / (Maximum HP of all Pokémon brought to battle)
• If there is no clear winner from conditions 1 or 2, then the player with the highest total HP remaining wins.
• If there is still no clear winner, the game will be a tie.

With forced ties, the best move for a player in a situation where they can't win is to start stalling to the turn limit, turning it into a pseudo-victory condition that losing players will shoot for. Wasting hundreds of turns becomes a much bigger gamble if you need to maintain an advantage over your opponent during that time. Calling the game at the turn limit is just another (needlessly time-consuming and risky) victory condition for the player in a winning position.

The maximum possible turns on-cart in a singles wifi battle in optimal conditions is around 300 turns (1 hour), assuming you take less than 1 second to make each decision. Other conditions are even lower (stuff like entry hazards). I'd support this, but I suspect a lot of people would consider this too low.

 

[Jellicent]

The maximum possible turns on-cart in a singles wifi battle in optimal conditions is around 300 turns (1 hour), assuming you take less than 1 second to make each decision. Other conditions are even lower (stuff like entry hazards). I'd support this, but I suspect a lot of people would consider this too low.

Singles Wi-Fi looks like just 3v3 at Level 50, so I guess Quick Link would be the more appropriate one to shoot for. Is a similar limit in place there? If so, then yeah, at least for gens that have such a limit, it should be implemented for accuracy. If not, then no, it shouldn't just cap at the same limit as 3v3. I actually think doubling that and going 600 is pretty reasonable, though. Very few games really last that long, and it would give stall vs stall an end-game to plan out that won't take eight hours or more to execute (at least, if going with decided winners instead of forced ties).

 

[Josh]

this 4444zsdsďf #×@##"'cdwf3

 

[DragonWhale]

Singles Wi-Fi looks like just 3v3 at Level 50, so I guess Quick Link would be the more appropriate one to shoot for. Is a similar limit in place there? If so, then yeah, at least for gens that have such a limit, it should be implemented for accuracy. If not, then no, it shouldn't just cap at the same limit as 3v3. I actually think doubling that and going 600 is pretty reasonable, though. Very few games really last that long, and it would give stall vs stall an end-game to plan out that won't take eight hours or more to execute (at least, if going with decided winners instead of forced ties).

Quick Link (and any online battle that's not rated) has a 60 minute hard timer from XY onwards. Not sure what this would mean for older formats for us though.

 

[Eo Ut Mortus]

The goalposts seem to have shifted a bit. The original issue was the existence of games that, if played optimally, would never end because each player's best strategy was to switch ad-infinitum. A 1000 turn limit would have resolved this. But Zarel, your most recent posts in the last thread suggest this solution is insufficient because people being stuck in a single ladder match for hours is still an issue. Fine if it is, but then the issue is no longer "endless stall" as this thread's title would suggest, but literally, people being stuck in ladder matches for hours. These cases are not exclusive to stall vs. stall, and as such, factors such as PP usage and the absence of KOes, which correlate with stall, should no longer be considered as draw criteria when we can instead confront the problem in a more direct manner: constraining the length of a game via time, rather than turns.

A timer-based solution is a step in the right direction, but there's no good reason it should be based on turn count. Instead, lower the timer or prevent it from refilling after a set period of time (and if possible, make this "blitz mode" player-specific to prevent abuse). This has several advantages: it handles all cases of malicious time stalling (moving slow on purpose), it doesn't affect stall vs. stall in which both players play quickly, and we can directly control the amount of time we want games to last rather than estimating based on turn count.

If actual endlessness in stall vs. stall is still deemed an issue, I think lowering the proposed turn limit further from 1000 turns is the simplest, least intrusive solution. I don't think this is necessary, though, if we gate game time properly.

 

[Zarel]

Singles Wi-Fi looks like just 3v3 at Level 50, so I guess Quick Link would be the more appropriate one to shoot for. Is a similar limit in place there? If so, then yeah, at least for gens that have such a limit, it should be implemented for accuracy. If not, then no, it shouldn't just cap at the same limit as 3v3. I actually think doubling that and going 600 is pretty reasonable, though. Very few games really last that long, and it would give stall vs stall an end-game to plan out that won't take eight hours or more to execute (at least, if going with decided winners instead of forced ties).

Quick Link is also 1 hour (300 turns), and keep in mind that 300 turns is an upper bound for a very unrealistic game with no items, no abilities, no hazards, inhuman button-pressing speeds, etc. In a realistic stall v stall game, 1 hour is enough for around 100-200 turns if you play quickly.

The goalposts seem to have shifted a bit. The original issue was the existence of games that, if played optimally, would never end because each player's best strategy was to switch ad-infinitum. A 1000 turn limit would have resolved this. But Zarel, your most recent posts in the last thread suggest this solution is insufficient because people being stuck in a single ladder match for hours is still an issue. Fine if it is, but then the issue is no longer "endless stall" as this thread's title would suggest, but literally, people being stuck in ladder matches for hours.

Are you really going to nitpick the clause name? "Endless Battle Clause" also makes sure battles end in significantly fewer than 1000 turns. If Endless Battle Clause had been a 1000-turn limit, most people would not think that it had done its job.

I think "Endless Stall Clause" is a good name for this clause, but feel free to suggest a different name.

These cases are not exclusive to stall vs. stall, and as such, factors such as PP usage and the absence of KOes, which correlate with stall, should no longer be considered as draw criteria when we can instead confront the problem in a more direct manner: constraining the length of a game via time, rather than turns.

A timer-based solution is a step in the right direction, but there's no good reason it should be based on turn count. Instead, lower the timer or prevent it from refilling after a set period of time (and if possible, make this "blitz mode" player-specific to prevent abuse). This has several advantages: it handles all cases of malicious time stalling (moving slow on purpose), it doesn't affect stall vs. stall in which both players play quickly, and we can directly control the amount of time we want games to last rather than estimating based on turn count.

If actual endlessness in stall vs. stall is still deemed an issue, I think lowering the proposed turn limit further from 1000 turns is the simplest, least intrusive solution. I don't think this is necessary, though, if we gate game time properly.

I have no problem with a strict timer, but once again: This will make most stall-v-stall games end in a timeout. A 1 hour time limit is not enough time for people to play out the sorts of games that people have been complaining about my proposals force-tying.

Look, if all of you agree, I will be very happy to implement a 1 hour time limit. I'm just afraid you're going to start complaining about all your stall games ending in a timeout. A 1 hour time limit is going to give you around 150 turns if you play very quickly.

 

[Zarel]

Heck, I can even tune the animation engine to give you enough time to play 300 turns if you rush. It's just, earlier, you guys were talking about my rules not letting you PP stall 1 PP every hour, and now suddenly you're fine with a 1 hour time limit? I'm quite happy with it, but I'm confused by the inconsistency.

 

[DragonWhale]

If a hard end solution is going to be implemented I wouldn't recommend using the time as the cutting point. People are going to flip out when people start using timer stalling as a tactic to reach the hour mark the last 10 minutes while in a winning position. If anything it should definitely be turn number-based.

 

[Zarel]

If a hard end solution is going to be implemented I wouldn't recommend using the time as the cutting point. People are going to flip out when people start using timer stalling as a tactic to reach the hour mark the last 10 minutes while in a winning position. If anything it should definitely be turn number-based.

Presumably, it would be a separate timer for each player. So running out the timer just means you lose.

Although a global timer that leads to a tiebreaker does exist in pokemon, and sometimes VGC players timerstall to get that that. Although that's because the VGC global time limit's like 15 minutes.

 

[Bughouse]

So long as there continues to be some amount of increment each turn and you don't literally have a fixed time bank, then I see nothing wrong with decreasing the amount of time that is incremented into each player's allotment after some high amount of turns.

But I don't want to make timer the primary way that long games are sped up because the PS timer is not remotely transparent compared to timers for just about any other game I've ever played.

I think it's better to force more draws rather than force unintentional timer losses.

 

[Eo Ut Mortus]

I'm nitpicking because people being stuck on the ladder for a long time in games that should go on forever if played properly are different from people being stuck on the ladder for a long time in games that will inevitably end. And yes, I think the goalposts have shifted given the overarching context, because a literal endless match was what prompted undisputed's thread to be posted, and a literal endless match is what was referenced in the post of Ciele's that you and others quoted in the previous thread. Just for clarification, I went and asked him myself, and he was indeed referring to literal endless matches, not just long ones. Like I said, if battle length is an issue people are concerned with, fine, and I agree it's worthy of address, but my point was to directly confront this problem by modifying the timer rather than continuing to develop roundabout solutions that kill passive games but don't directly prevent timer stalling.

As for the implementation of the timer, I personally never said anything about a strict one hour limit, so I hope I'm not being included in "you guys." To be clear, what I had in mind was (assuming current timer is base time + x refill for each player) to lower the base time and/or lower/remove the refill after y minutes, with these y minutes unique to each player to prevent purposefully timestalling to modify the timer mechanics. I do not necessarily favor the static approach (completely removing the refill) over the other ones--I was just throwing it out there--and I do not have any values in mind. It is my hope, however, that there exist some values that will allow for legitimate long matches to be played while still (in conjunction with a hard turn limit) limiting literal endless matches to a more reasonable timeframe than what we have now, although I admit that a compromise will probably have to be made on either end.

 

[Zarel]

I'm nitpicking because people being stuck on the ladder for a long time in games that should go on forever if played properly are different from people being stuck on the ladder for a long time in games that will inevitably end. And yes, I think the goalposts have shifted given the overarching context, because a literal endless match was what prompted undisputed's thread to be posted, and a literal endless match is what was referenced in the post of Ciele's that you and others quoted in the previous thread. Just for clarification, I went and asked him myself, and he was indeed referring to literal endless matches, not just long ones.

I'm curious how you phrased your question. Personally, I started this thread intending to tackle 7-hour battles (since everyone already agreed on a 1000-turn limit to handle literally endless battles), and I did it because Ciele's post explicitly referred to ending ABR vs Tele, which was not an endless match.

Anyway, I don't think this is a big deal. We agreed on a 1000-turn limit like two threads ago, everything since then (including the argument on "offer draw") has been work on long games, not endless games.

Like I said, if battle length is an issue people are concerned with, fine, and I agree it's worthy of address, but my point was to directly confront this problem by modifying the timer rather than continuing to develop roundabout solutions that kill passive games but don't directly prevent timer stalling.

As for the implementation of the timer, I personally never said anything about a strict one hour limit, so I hope I'm not being included in "you guys." To be clear, what I had in mind was (assuming current timer is base time + x refill for each player) to lower the base time and/or lower/remove the refill after y minutes, with these y minutes unique to each player to prevent purposefully timestalling to modify the timer mechanics. I do not necessarily favor the static approach (completely removing the refill) over the other ones--I was just throwing it out there--and I do not have any values in mind. It is my hope, however, that there exist some time setup that will allow for legitimate long matches to be played while still (in conjunction with a hard turn limit) limiting literal endless matches to a more reasonable timeframe than what we have now, although I admit that a compromise will probably have to be made on either end.

You say you don't have values in mind, but can you at least throw out suggestions for x and y? It kind of sucks that I'm the only one suggesting values (I don't even play SM OU, which seems the main affected tier) and everyone's just criticizing mine instead of suggesting values of their own.

But I don't want to make timer the primary way that long games are sped up because the PS timer is not remotely transparent compared to timers for just about any other game I've ever played.

Yeah, this is mostly my fault for designing a sophisticated timer system rather than something simple.

Let me describe how the timer currently works in ladder games:

Both players start with 210 seconds in the bank. Every turn, their bank is rounded up to the nearest 10, and then 10 seconds are added to their bank (and another 10 is added if they have below 160 seconds left).

At most 150 seconds can be used every turn.

That's it. Most of the weirdness comes from when your bank is above 150 seconds, because then it just says "150 seconds left". If your bank is under 150 seconds, it'll say how much time you have left, and that time will reliably go up by 20 every turn.

The 150 second per-turn limit was mostly chosen to be enough time to run and pee if you rush. The 210 starting bank was chosen because 3 minutes seemed reasonable.​

Honestly, I feel like displaying 2 numbers instead of 1 is probably all we need to make it un-confusing.

The "+10, +10 more if you're below 160 seconds" is a really weird provision that can probably just be switched to a flat "+10 per turn" and a cap on bank size. Maybe "+20 per turn" for doubles games?

 

[Jellicent]

Quick Link (and any online battle that's not rated) has a 60 minute hard timer from XY onwards. Not sure what this would mean for older formats for us though.

Unless they also have a hard timer programmed into them somewhere, just leave them alone. The last time a discussion about dealing with long battles in GSC (the stereotypical "stallish" gen) came up, it was for those rare battles that hit the 600+ turn mark. I'm not sure why SM is occasionally having issues with two players making their best possible moves and still taking over 1000 turns to win, but it looks like GF preemptively solved it with mandatory time limits built into PvP battles anyway.

Quick Link is also 1 hour (300 turns), and keep in mind that 300 turns is an upper bound for a very unrealistic game with no items, no abilities, no hazards, inhuman button-pressing speeds, etc. In a realistic stall v stall game, 1 hour is enough for around 100-200 turns if you play quickly.

I have no problem with a strict timer, but once again: This will make most stall-v-stall games end in a timeout. A 1 hour time limit is not enough time for people to play out the sorts of games that people have been complaining about my proposals force-tying.

Look, if all of you agree, I will be very happy to implement a 1 hour time limit. I'm just afraid you're going to start complaining about all your stall games ending in a timeout. A 1 hour time limit is going to give you around 150 turns if you play very quickly.

I mean, I honestly have less of a stake in this as an old gens player (and think it kinda sucks), but if there's literally no way for two peeps to battle on-cart for over an hour in gens 6+7, then that's how it should be simulated. This is most akin to Team Preview becoming a mandatory condition of PvP in BW. If we're going to break cartridge mechanics, we at least need to create a clause that explains why battles in those gens might last for several hours.

So we'd have both a clause to extend the length of battles beyond in-game capabilities and one that somehow tries to restrict them from being too long by whatever standards we're able to agree on...

 

[Jellicent]

The game being called at the one hour mark is a built-in win condition, so it's a fundamental part of the game starting with XY. We do play on a simulator, so yeah, we can create a clause to break from cartridge mechanics. It's something we should only really use as a last resort to prevent game-breaking situations, such as with Sleep Clause. Creating a clause to break cartridge mechanics and another to fix the issues resulting from doing that is odd to say the least.

Why are we ignoring an integral part of the game to begin with?

 

[Eo Ut Mortus]

You say you don't have values in mind, but can you at least throw out suggestions for x and y? It kind of sucks that I'm the only one suggesting values (I don't even play SM OU, which seems the main affected tier) and everyone's just criticizing mine instead of suggesting values of their own.

Without knowing the intent behind each function of the current timer, this is what I would suggest:

- Remove the +10 bonus time for <160 seconds. This means players get 10 seconds + however many seconds are rounded up, which I'll treat as the average (5s) for a total of 15s worth of bonus time each turn. Effectively, this means a player who takes the maximum amount of time each turn to move can only draw out a 1000 turn game for 4 hours, compared to 7 hours using the same calculations with the current timer.

- In line with what I proposed earlier, we can also decrease the bonus time from 10 to 5s after a player's individual timer has decreased by a cumulative 30 minutes. Assuming a player takes 15s to move each turn, this will mean that the bonus time will decrease after at least 120 turns have passed. The remaining 880 turns of a 1000 turn game will take around 2.5 hours at most to complete, for a total time of 3 hours.

- I'm assuming 5 seconds to be the amount rounded up to the timer each turn, but of course, there could exist cases in which the average time rounded up ends up being greater than this. If these are a concern, then I would suggest removing the rounding, assuming it's unneeded, and changing the bonus time to a flat 15 and 10 seconds, respectively.

- The bank and/or per-turn limits can be increased if taking breaks / using the bathroom is a concern, since these values don't really affect the numbers for longer matches. As an aside, is there a reason the 150s per-turn limit and the 210 bank total have to be different? It would make more sense to me if the values were radically different from each other, like in chess, but it seems to me that there's more worth in clarity than in these specific values.

- If this is still too long for a match, then I'd suggest decreasing the bonus time values to 10/5, which would cap the maximum length of a 1000 turn stall game at slightly more than 1.5 hours.

 

[Zarel]

The game being called at the one hour mark is a built-in win condition, so it's a fundamental part of the game starting with XY. We do play on a simulator, so yeah, we can create a clause to break from cartridge mechanics. It's something we should only really use as a last resort to prevent game-breaking situations, such as with Sleep Clause. Creating a clause to break cartridge mechanics and another to fix the issues resulting from doing that is odd to say the least.

Why are we ignoring an integral part of the game to begin with?

There's some contention in Smogtours, so to be clear, this is what I'm going to do. Anyone who disagrees should propose a clause-mod, like Sleep Clause Mod or HP Percentage Mod.

(Closing; direct further discussion to the new thread: https://www.smogon.com/forums/threads/1-hour-battle-time-limit.3617340/ )