Review the PvP usage % for Season Changes

[Imperial]

Background

Every 3 months the season cycle changes, and that also means Pokemon raising or dropping from a tier. At the moment 4.36% is the cut-off point at the third of every month, where the main changes happen. Usually there are 1-2 Pokemon, occasionally more, which change from each tier. However, this season will be different.

 

Issue

The introduction of 'Randoms' matchmaking has proven very popular within the community - however that has come with consequences, with UU/NU matchmaking being completely screwed over as many of the players find it very difficult to find games unless in certain timezones, with the main way to play being through official or community tournaments.

 

I believe there has been similar statistics on another thread, but just to indicate as evidence:

 

At the end of Season 5 where Randoms was introduced in the last month

 

 

Season 6 when Randoms was fully embedded into the game

 

 

 

 

I have used Rotom-Heat as an example as this has been the most frequently used Pokemon in UU at the end of Season 5 and throughout Season 6 - NU would be similar to these examples so I will avoid posting these sceenshots.

 

Basically the total number of games have dropped significantly during the last three months, with the numbers at their lowest this month (again I appreciate it's too early to judge) - this leads onto my next point.

 

Month 3

Based on the current system, it is easier for players to spam certain Pokemon for it to raise or drop into a certain tier. An example was Lucario at the end of Season 4, where players managed to use it in enough games to keep it in OU, before inevitably dropping and becoming banned from UU.

 

This one is really difficult to predict as it's only been 8 days and lots can change, however looking at early statistics Mantine has incredibly high usage when it has hardly been considered in the UU meta, whilst Aggron (considered as untiered and has barely been used by experienced players in UU), has a higher usage than Lanturn and Gligar, both very effective walls which have been around in the tier for a long time until recently.

 

 

The main point is that due to the lack of NU/UU games and the 4.36% staying how it is, this season will see the most changes happen. It also enables players to use Pokemon which are usually completely unviable to rise into UU, or even OU (e.g. Aerodactyl constantly bouncing between OU and UU)

 

Proposal

• Calculate the usage % over the course of three months to obtain a general average - this reduces the chance of players spamming Pokemon in the final month in an attempt to abuse the system
• Slightly increase the usage % at the end of the third month (e.g. From 4.36% to 5%, or if that's too drastic between 4.6-4.8%) and review the situation after a few seasons
• Take into account tournament usage % for these Pokemon, as the majority of experienced PvP players participate in official tournaments, rather than having the time to spam matchmaking games. This will give you a more clearer indication of what the PvP community uses in each tier.

 

To emphasise, I'm happy with the introduction of 'Randoms', as it's proven very popular, but the UU/NU meta needs revival and more stabilisation than before - not only increasing the incentives to increase the number of players for these tiers, but also making sure that the lack of games is not taken advantage of with the introduction and spamming of completely unviable Pokemon.

Edited December 8, 2021 by Imperial

[Quinn010]

29 minutes ago, Imperial said:

Calculate the usage % over the course of three months to obtain a general average - this reduces the chance of players spamming Pokemon in the final month in an attempt to abuse the system

best solution imo 

also crazy to see how much those numbers dropped 

[danhred]

1 hour ago, Imperial said:

 

Very sad those data

[BaliAds]

Something Orange suggested is give separate data for above 600 ELO players that way we could better judge the actual win rates and usage of things. 

[Meltdown]

4 hours ago, BaliAds said:

Something Orange suggested is give separate data for above 600 ELO players that way we could better judge the actual win rates and usage of things. 

600 ELO is achieved with very few games. I could be mistaken, but this suggestion was made for higher ELOs. In a way it makes sense, since you can't compare (with exceptions, of course) a 500 game with a top25+ ladder game, in terms of optimizing the use of a particular pokemon. However, I believe that if this statistic were made available, it would further narrow the use of certain things and diversification (which almost doesn't exist, since most players already use the most "successful" teams) would decrease a lot, so I believe this problem must be considered.

[caioxlive13]

On 12/9/2021 at 12:51 AM, Meltdown said:

600 ELO is achieved with very few games. I could be mistaken, but this suggestion was made for higher ELOs. In a way it makes sense, since you can't compare (with exceptions, of course) a 500 game with a top25+ ladder game, in terms of optimizing the use of a particular pokemon. However, I believe that if this statistic were made available, it would further narrow the use of certain things and diversification (which almost doesn't exist, since most players already use the most "successful" teams) would decrease a lot, so I believe this problem must be considered.

In this point, meltdown is right. High elo of all Tiers, except Randoms, always have the same teams, and you rarely see a High ladder using something different of a Garchomp(OU), Rotom-heat(UU) or Blaziken(NU), for example. So, to have diversification in teams, data of low ladder teams need to be considered

But i really don't know why they're mattering with UU and NU. With metagaming too boring with wall abusers, simillar to OU, and a Poor Balancing in matches(350 Elo against 700 elo, before add randoms.), it's expect that tier died with randoms coming, with people wanting FUN in matchs.

[ThinkNicer]

I think you should even consider if usage is still the best way to go about defining the lower tiers. The strength of usage based tiering is that you have so much objective data to form tiers with, that it gets difficult to argue for more subjective approaches. Now though, we don't have the benefit of plenty of data anymore. Making it easier for stats to get skewed in the lower tiers. IMO TC should look at other options to form the lower tiers.