Pending Implementing a Turn Limit for BSS/VGC (pseudo-Game Time)

[DragonWhale]

So the main cartridge formats, VGC and BSS, has a 20 minute game timer that is currently not implemented on PS.

The primary reason for the lack of timer was that it's currently impossible to accurately implement it as everything (turn ending operations, animations, a ton of stuff) have different lengths of time that eat into the timer. So currently the solution is to just have no tiebreaker at all.

However this creates a significant shift in the metagame, as timer win conditions are very important elements in these formats (a little less so in VGC compared to BSS, but still important). This creates a environment where these metagames on PS will develop differently to the actual formats on cartridge that they are supposed to be simulating. As such, I am suggesting a solution that, while not precise to the second, simulates the cartridge timer tiebreakers more accurately than the current implementation: adding a turn limit.

The actual turn number to limit games to can be up for discussion, but I am suggesting a maximum of 30, and realistically 25. In gen 6 BSS where the timer is 30 minutes we can just raise the turn limit to 1.5x (gen 7 has no game timer). VGC might be lower due to the doubles nature, but again, this can be up for discussion.

 

[DaWoblefet]

I'm opposed to a turn limit for 3 main reasons:

• Hypothetical turn limits assume metagame-specific elements. Although nobody would actually do so, it is easily possible to get way more than 25-30 turns in a 20-minute singles cartridge game. If both players are just mashing inputs for moves that play no animation, you're going to get significantly higher turn counts. Of course, two players mashing moves like Sleep Talk and Snore isn't metagame-relevant, but when you get into what is metagame-relevant, you introduce variance.
  • If metagame relevance is considered, the value chosen will be arbitrary until a metagame has enough time to reasonably develop to see if the number of turns should be tweaked.
  • If metagame relevance is considered, the turn count could vary from format to format, making it difficult for players to keep track of their prospective win condition.
  • If metagame relevance is not considered, you'd have to pick the limit based on some metagame where more turns are likely to account for all cases, which means the turn limit is going to on average be "too high" for most metagames.
• We don't have significant data on the turn counts of timeout games. In VGC, the number of recorded games that go to time is low relative to the total number of recorded games. Of those, I guarantee nobody has been going through and tracking turn counts, so we're basing our data off of too small a sample size. In NPA, the VGC equivalent of SPL, it was common for a time for turn limits to be imposed to limit Chansey stall, for example. One year this was 40 turns, and in another this was 30. I find it surprising the VGC community would pick this number when the turn count is, on average, going to be much lower than the proposed 30 turns for BSS. Perhaps BSS has more recorded timeouts than VGC that can be used as a reference, but I would find that surprising.
  • Above all else, we don't want games that would have otherwise completed on cart to terminate early on Showdown.
• Turn-based limits offer different and unique win conditions compared to time-based limits. To reference NPA again, most players felt that these turn limits actually buffed defensive strategies like Chansey stall. With a time-based win condition, it's often quite difficult to predict what turn a match will go to time, which leads to different play. Do I need to conserve my healing move now, or save it? How many Bulk Up can I get set up before I need to start attacking? Total time also turns Your Time into a resource, needing to know when it's wise to start taking the full selection time vs when to save thinking time. Timeouts are also not always set in stone - perhaps a board state that would favor one player suddenly gets crit, and now the opponent can reverse timer stall. These factors don't exist with a turn limit, where you can plan your endgame much more precisely.

I do think there is likely some merit to the point of metagame divergence being a problem (I've shared usage stats below comparing cart and PS). You would expect most of the top 15 Pokemon to be identical, but unlike VGC's relative uniformity, BSS only shares 12/15 of the top 15 Pokemon. For others reading this thread, that might not sound like a lot, but it is unusual when two identical formats diverge. Unlike VGC, you can't attribute the BSS changes to say, OTS availability, and it is unlikely that playerbase distribution, i.e. cartridge tending to have more Japanese players and Showdown tending to have more Western players, explains the difference either.

Data for VGC Regulation D

Showdown usage stats (1760 August PS usage)

1. Flutter Mane
2. Urshifu Rapid Strike
3. Iron Hands
4. Amoonguss
5. Landorus-T
6. Tornadus
7. Chien-Pao
8. Chi-Yu
9. Rillaboom
10. Heatran
11. Gholdengo
12. Urshifu
13. Cresselia
14. Ursaluna
15. Gyarados

Cartridge usage stats (Home Season 9; can't use Season 10 bc of DLC1 lol)

1. Flutter Mane
2. Urshifu Rapid Strike
3. Tornadus
4. Chien-Pao
5. Iron Hands
6. Amoonguss
7. Landorus-T
8. Heatran
9. Rillaboom
10. Cresselia
11. Dragonite
12. Gholdengo
13. Ursaluna
14. Chi-Yu
15. Urshifu

Data for BSS Regulation D

Showdown usage stats (1760 August PS usage)

1. Dragonite
2. Flutter Mane
3. Urshifu-Rapid-Strike
4. Chien-Pao
5. Gholdengo
6. Ting-Lu
7. Dondozo
8. Mimikyu
9. Iron Bundle
10. Landorus-T
11. Ursaluna
12. Heatran
13. Grimmsnarl
14. Toxapex
15. Gothitelle

Cartridge usage stats (Home Season 9; can't use Season 10 bc of DLC1 lol)

1. Dragonite
2. Flutter Mane
3. Chien-Pao
4. Gholdengo
5. Urshifu-Rapid-Strike
6. Iron Bundle
7. Ting-Lu
8. Ursaluna
9. Landorus-T
10. Heatran
11. Scizor
12. Mimikyu
13. Chi-Yu
14. Goodra-Hisui
15. Dondozo

The ideal case would of course be that someone documents all possible animation durations in various circumstances, then we implement those on Showdown. The animation durations themselves would not change on PS, but time would be reduced from the 20-minute timer as if they were the appropriate duration. This is a ridiculously difficult research project, but I do not think a turn limit is a good substitute for it.

 

[DragonWhale]

The hypothetical turn limit I am trying to introduce can never be as accurate as implementing the animation duration timer countdown system, which like you said is a ridiculously difficult research project, but I believe having at least something to simulate a tiebreak when the game drags on is better than the current policy of doing absolutely nothing at all. Turn limit is just the best band-aid solution until we are able to get something more accurate like the animation duration research completed.

You make a great point in that the worst case of this implementation is that games are ended earlier than they would do on cartridge because of the turn limit. This is why my suggestion is to place the turn limit to the highest end of what can reasonably be seen as the number of turns for the time to run out on cartridge (aka not snore spamming to get the record). However, I don't think we can really take metagame relevance into account with this proposal. It's never going to be 100% accurate and it can potentially take the limit down below the reasonably possible, which we both agree is the worst case scenario. Having the turn limit be "too high" is a better alternative, while also being better than having no turn limit at all.

30 turns is something that BSS games, and especially VGC games, will not reasonably reach given that it is 40 seconds a turn. Out of personal experience, each move/switch takes on average 7-8 seconds, that's around 15 for both pokemon, longer if the move has stat altering effects or the pokemon switched in has ability activation or switches into hazards. Communication lag adds 5 more seconds, and end of turn effects is commonplace. This isn't even counting the long terastalize animation either. That's way less than 20 seconds of move selection time per turn, if any at all, and given that most players will willingly play for time with their 45 seconds if given the board position to do so, reaching 30 turns is pretty much impossible. I'd be surprised if a BSS game even goes to 20 turns honestly, and the turns last even longer in VGC. To give another perspective, you run out of your time if you average 14 seconds per turn at turn 30, which is hard to think you even come close.

Given that BSS is probably the biggest competitive Pokemon content on Youtube, we can probably get more data if needed be (although a lot of games that goes to time is probably cut down). Ideally I don't have to spend time scouring through videos looking for them and counting turns like a madman, but if that's what's needed then I guess I can bite the bullet.

I have to disagree with turn-based and time-based limits significantly altering win conditions and therefore play. On cartridge you have the 3 minute alert that pops up (which they finally fixed), which by both format standards will be the last 3 turns of the game given that there's always going to be a player that benefits from the tiebreak and as such will use their 45 seconds that turn to reach it. At that point the turn and time limit plays the same: both players will play the turn like there's 2 more turns after. Now there may be a difference before that point to be the case, but most of the time in those situations the players are aware that one of the players are playing to time and therefore somewhat aware of the amount of remaining turns left, provided that they kept track of the time that the game started and perhaps not as precisely. Timeout board positions not being set in stone is something that applies identically to both turn and time-limit based scenarios; a board position can get flipped to favor another player for the timer win and a turn win pretty much the exact same way, since they are usually aware of roughly how many turns are left in a timer scenario anyway. In VGC, the current implementation has this issue for games that drag on, which I believe affects plays a lot more erratically than a turn limit since it becomes a speedclicking game, while if a turn limit were in place the tiebreak would come into play before the your time becomes a problematic issue instead of a skill issue.

I appreciate you bringing up the usage stats disparity, because it really highlights how the lack of a tiebreaker really impacted certain Pokemon's viability. The primary culprit here is Gothitelle, which on HOME it is placed 117th. It's not too surprising given how every team has a Dark type, and it's possible that the usage is skewed on PS because a few players spam it and because of the playerbase size it can have an impact. However I want to point out the issue by looking at the moveset usage stats. The set in question has 95% usage: Max Def Choice Scarf Trick/CM/Rest/Stored Power. I think it's fairly reasonable to assume that using this set as effectively on cart is almost impossible due to the 20 minute timer, since you don't have an infinite amount of turns to PP stall and reach +6 to sweep. It is a lot more representative of the format to place the same shackles on this set as the one it has on cartridge, and I believe a turn limit is the best solution that we can currently introduce until we find something better. One of the main reasons why I even posted this in the first place is because Reg E is going to introduce probably the most infamous Pokemon in terms of games going to tiebreak: Gliscor. This Pokemon is going to play completely differently simply based on whether there's a tiebreak or not, and having at least something to better represent the cartridge metagame is something that will improve PS.