Suggestions for Pokémon Showdown

Status
Not open for further replies.
please add an opt out button for the thing where the IVs are set automatically. Setting Ditto's HP type is a nightmare

EDIT: this applies even moreso now Photon Geyser is a thing
 
Last edited:
You shouldn't be able to get Ash Greninja in Hackmons Cup, just like how you can't get Mega Evolutions, Zen Darmanitan, or Blade Aegislash.
 
Maybe a bit of a stretch, but would it be possible to add a Battle Tree/Maison/Tower format on the server? Just like competitive player vs player battling, this would serve as a substitution for what one can expect to find and do in-game, but here with the focus on the competitive features that you can do by yourself (as in not playing against another player, but against a CPU/bot, and aim to get high win streaks). It would be nice to play those modes on Showdown with its animations and format, which are quite fast (and still good-looking) compared to what you find in-game (which is very slow and time consuming).
 
When we play or watch battles in real time on showdown we have total freedom when it comes to regulating volume. Music, effect (mostly cries) and notifications can be adjusted for our liking seperately. I'd suggest implementing the same for replays (bar notifications since they don't apply there). Currently when watching replays we can only choose between music+cries at max volume and no sound at all. At the very least having an option to mute music and cries seperately would be more than welcome.

PS. It's not my original idea, it was originally suggested by some user (forgive me for not remembering his name) in RoA who doesn't have an account on smogon.
 
Entering Ultra Necrozma into the teambuilder fails all team validation tests. Personally, I think that Ultra Necrozma should just be labelled as "Illegal," just like Darmanitan Zen and Meloetta Pirouette is, since it can't be placed directly onto a team.
 
I'm going to Antar for more visibility, apologies if it was unwarranted.

Elo should not be the main ladder rating. In a contest of rating systems to predict the winner of 7,800 chess games, Elo came out second to last when compared to Glicko, Glicko 2, PCA, and Chessmetrics.

Here is the blog post for the halfway mark of the contest which explains some more: http://blog.kaggle.com/2010/09/21/elo-vs-the-rest-of-the-world-at-the-halfway-mark/
http://blog.kaggle.com/2010/09/21/elo-vs-the-rest-of-the-world-at-the-halfway-mark/
and here are the final results: https://www.kaggle.com/c/chess/leaderboard
https://www.kaggle.com/c/chess/leaderboard
Chessmetrics got 37th place with a Root Mean Squared Error of 0.70866

Glicko 2 got 67th place with a RMSE of 0.71747

Glicko got 68th place with a RMSE of 0.71760

Elo got 137th place with a RMSE of 0.74377

PCA got 152nd place with a RMSE of 0.76607

I don't know the feasibility of Chessmetrics being incorporated to showdown but Glicko 2 should be made the main ladder rank due to its higher predictive power when compared to Elo.

If Chessmetrics could be incorporated then that would be even better, their website explaining their formula is here: http://www.chessmetrics.com/cm/CM2/Formulas.asp
 

pants

*dies inconveniently*
is a Battle Simulator Moderator Alumnus
Can drivers be allowed to set modchat to autoconfirmed and turn autoconfirmed off? Staffing can be really annoying when there are no mods on/gmods in help and ac is off.
 
I'm going to Antar for more visibility, apologies if it was unwarranted.

Elo should not be the main ladder rating. In a contest of rating systems to predict the winner of 7,800 chess games, Elo came out second to last when compared to Glicko, Glicko 2, PCA, and Chessmetrics.

Here is the blog post for the halfway mark of the contest which explains some more: http://blog.kaggle.com/2010/09/21/elo-vs-the-rest-of-the-world-at-the-halfway-mark/
and here are the final results: https://www.kaggle.com/c/chess/leaderboard
Chessmetrics got 37th place with a Root Mean Squared Error of 0.70866

Glicko 2 got 67th place with a RMSE of 0.71747

Glicko got 68th place with a RMSE of 0.71760

Elo got 137th place with a RMSE of 0.74377

PCA got 152nd place with a RMSE of 0.76607

I don't know the feasibility of Chessmetrics being incorporated to showdown but Glicko 2 should be made the main ladder rank due to its higher predictive power when compared to Elo.

If Chessmetrics could be incorporated then that would be even better, their website explaining their formula is here: http://www.chessmetrics.com/cm/CM2/Formulas.asp
Firstly I'd question why you would advocate the #37 in a test rather than the #1. Surely the entire purpose of the event was to establish the best rating system so not to use it would raise questions about why use the findings at all. (Although in those case the winning predictive scoring system dubbed "ELO++" did share certain similarities with the Chessmetrics system.)

I'd also question the extent which the findings are a) reliable and b) relevant. On the point of reliability, while 8000 sample games with 1600 as training data may seem like a lot, when compared to the literally millions of battles on Showdown this pales into insignificance as a sample size. Additionally most of the systems performed comparably well so there is no huge advantage even if the results are taken as 100% accurate and reliable. On relevance, Chess clearly has several factors that differentiate it from Pokémon, first and foremost being the innate advantage of playing white which is taken into account by a lot of the systems that were tested.

Also tbh I don't think it's that big of a problem although I would definitely like to see an option to sort the ladder by GXE or other ratings for example.
 
Firstly I'd question why you would advocate the #37 in a test rather than the #1. Surely the entire purpose of the event was to establish the best rating system so not to use it would raise questions about why use the findings at all. (Although in those case the winning predictive scoring system dubbed "ELO++" did share certain similarities with the Chessmetrics system.)

I'd also question the extent which the findings are a) reliable and b) relevant. On the point of reliability, while 8000 sample games with 1600 as training data may seem like a lot, when compared to the literally millions of battles on Showdown this pales into insignificance as a sample size. Additionally most of the systems performed comparably well so there is no huge advantage even if the results are taken as 100% accurate and reliable. On relevance, Chess clearly has several factors that differentiate it from Pokémon, first and foremost being the innate advantage of playing white which is taken into account by a lot of the systems that were tested.

Also tbh I don't think it's that big of a problem although I would definitely like to see an option to sort the ladder by GXE or other ratings for example.
Because iirc #1 did not give out their formula. The only systems that have their formulas available to us are the benchmark systems I posted.

As for reliability, we have no other test to base our outlook of each system's predictive power on. 7,800 sample games is far, far, far, far, better than 0. Until a better test comes around, we should stick with the only data we have. There is no reason to stick with Elo other than ease of calculating, which is not important since showdown calcs for us. Predictive power >>> ease of calculation. The systems did perform comparatively well but might as well go with the best one. Why take third place (Elo) rather than first place (Glicko 2)? The ladder is there to show is who is the best out of the players that played, why not use the best system availabale to us.

As for relevance, White going first does not affect any of the benchmark systems, the systems purely take past performance and work from there. that concept applies to showdown, which is why we have Elo and Glicko 2 there in the first place.
 
We need a feature where we can search for another battle in the background while the current one is searching. This would enable us to fight current gen battles while searching for previous gen matches.

Years ago this was brought up and was declared to be on the to-do list.

I'm wondering where is this feature where we can use find battle multiple times while the others search in the background?

This isn't some rinky dink suggestion - this could literally revitalize previous-gen battling.

It baffles me at all the fixes and features that have been pushed in the past few years, but not this??

Can we please make this an immediate dev priority? It has been 3 years of waiting, it's not going to get done unless somebody hunkers down today, this week, this month and gets it done.

I can promise you will see battles increase by thousands per month if you do this, it will revive so many gens that people don't bother with at the moment. If you can search for 3 battles at once then you won't mind waiting 1 hour to find a rby match.

Soon after this feature is implemented that might decrease to 30 minutes, then 15, maybe even 5! Let's get this done guys.

Smogon took the mantle of online Pokemon away from PO because they're supposed to be better. Well, let's prove it.

Can we do this so trying to find previous gen battles isn't something like this:
 
I'd like to ask that we make practice battles for people to test out their teams in a specific tier (practice OU, practice UU, Practice Gen 3 Ubers, etc). In this practice area, people can practice out new team compositions without putting their tier score at any risk. This may lead to more room for people to experiment, and for people to learn how to use Pokemon they may have been scared to use before; out of the fear of losing their ladder spot. Thank you for reading :)
 
I'd like to ask that we make practice battles for people to test out their teams in a specific tier (practice OU, practice UU, Practice Gen 3 Ubers, etc). In this practice area, people can practice out new team compositions without putting their tier score at any risk. This may lead to more room for people to experiment, and for people to learn how to use Pokemon they may have been scared to use before; out of the fear of losing their ladder spot. Thank you for reading :)
Well you could always use alts
Alts = alternate accounts basically meaning just make another account on Pokemon Showdown and practice using that account
 
Last edited:
Smogon constantly bans pokemon like Pheramosa, Aegislash, Mega Metagross, and most recently, Naganadel from ou for the sake of balance. But these many of these pokemon not viable in ubers and AG meaning that there isn't a 6v6 singles format that these pokemon can be used viably which is sad because many of these pokemon are really fun to use. What I propose is that smogon add a format to pokemon showdown that has the exact same rules as battle spot singles, but instead of 3v3, it's 6v6. This wouldn't replace ou but would be an alternative, maybe an OM. I would just like to see pokemon like aegislash be viable again.
 
I do have one suggestion, a way to initiate a battle and have the other side controlled by an AI. Could be used for team testing and also for custom matches used in tournaments (I'm running one and wanting to have Team Rocket battle against members depending on when it is during as an example). Could have another player control them but would want to be able to have everyone involved in the event if the creator is part of the battling too.
 

Ryota Mitarai

Shrektimus Prime
is a Tiering Contributoris a Contributor to Smogonis a Top Smogon Media Contributor
Smogon constantly bans pokemon like Pheramosa, Aegislash, Mega Metagross, and most recently, Naganadel from ou for the sake of balance. But these many of these pokemon not viable in ubers and AG meaning that there isn't a 6v6 singles format that these pokemon can be used viably which is sad because many of these pokemon are really fun to use. What I propose is that smogon add a format to pokemon showdown that has the exact same rules as battle spot singles, but instead of 3v3, it's 6v6. This wouldn't replace ou but would be an alternative, maybe an OM. I would just like to see pokemon like aegislash be viable again.
This is not the place to suggest format ideas.
 
Firstly I'd question why you would advocate the #37 in a test rather than the #1. Surely the entire purpose of the event was to establish the best rating system so not to use it would raise questions about why use the findings at all. (Although in those case the winning predictive scoring system dubbed "ELO++" did share certain similarities with the Chessmetrics system.)

I'd also question the extent which the findings are a) reliable and b) relevant. On the point of reliability, while 8000 sample games with 1600 as training data may seem like a lot, when compared to the literally millions of battles on Showdown this pales into insignificance as a sample size. Additionally most of the systems performed comparably well so there is no huge advantage even if the results are taken as 100% accurate and reliable. On relevance, Chess clearly has several factors that differentiate it from Pokémon, first and foremost being the innate advantage of playing white which is taken into account by a lot of the systems that were tested.

Also tbh I don't think it's that big of a problem although I would definitely like to see an option to sort the ladder by GXE or other ratings for example.
Ok if nothing is going to happen then we should at least be able to sort by gxe and glicko like you said
 
It would be super helpful if the damage calculator showed a Pokemon's stats after boosts/drops/items/abilities are applied (maybe to the right of the boost/drop panels; there's a ton of free space there).
 
Status
Not open for further replies.

Users Who Are Viewing This Thread (Users: 1, Guests: 1)

Top