Published using Google Docs
VGC Cheating 2023
Updated automatically every 5 minutes

====

Twitter Chain

====

VGC Cheating 2023: with the Pokémon World Championships concluding last weekend, we’ve gathered >850 recent rental teams and analyzed them with our usual *free hack checks*. A season finale of @4Chris_Brown’s nightmare! Overall stats below:

856 teams is a hefty dataset, more than triple the number we collected last year for SW/SH’s VGC series. Like before, we’ve done a bit of extra data analysis of competitive trends, so check out the doc if you’re into that. But first, let’s recap!

Link to this published write-up – You are here!

Regulation D was the most buzzing in the number of rental teams that were uploaded, both for regular ladder play as well as from Worlds. Looking back at this year and years prior, we still see a roughly 50/50 split of teams being illegally modified… including prizes & awards.

This year at Worlds, TPC upped their game in a simple way, adding in a “new” hack check. They were able to disqualify quite a few competitors for transfer Pokémon missing their HOME tracker, but their implementation only verifies locally. What was missed?

https://twitter.com/DaWoblefet/status/1690103118803972096

Shohei Kimura’s first place team was not without sin – a pivotal Pokémon, Amoonguss, was modified to have a 0 ATK IV (originally 18). TPC does not check for raid RNG correlations, even though it takes ~1 millisecond to calculate. Hopefully no Amoonguss Mystery Gift in the future!

Another fun example is Tang Shiliang’s team, which bested Wolfe Glick’s in the deciding final round of Day 1. The two Max Lair genies are missing their RNG correlation, 100% hacked. In Day 2, Tang Shiliang went 5-3, losing to Shohei (#1), Michael (#2), and Hirofumi (#22).

In closing, nothing has changed, and I’m not optimistic for things changing going forward. At Worlds, GameFreak could have revealed new features with massive changes to accessibility, but they did not. Scarlet & Violet continue to have the same hurdles, and cheaters gonna cheat.

(this ^ tweet is the same as my tweet from last year, where I assumed S/V would be exactly the same):

Previous Analysis & Writeups

S/V Finale - Worlds

SW/SH Finale - Worlds

S/V Series 2 - Regulation A & B 

S/V Series 1 - ChatGPT prompts

Previous Tweet Chains:

S/V Korean Disqualifications: https://twitter.com/Kaphotics/status/1665006560706453504 

S/V Regulation C: https://twitter.com/Kaphotics/status/1664466465985282049 

S/V Regulation B: https://twitter.com/Kaphotics/status/1642004136970567680 

S/V Regulation A: https://twitter.com/Kaphotics/status/1619766752841437184

SW/SH Finale: https://twitter.com/Kaphotics/status/1563979743094906880

SW/SH Regionals: https://twitter.com/Kaphotics/status/1525318121757650944

SW/SH Players Cup: https://twitter.com/Kaphotics/status/1455009546972262402

2021 Aug Tournament: Arash https://twitter.com/Kaphotics/status/1426277865734033408

Wolfe’s Moltres: https://twitter.com/Kaphotics/status/1525667071471652865 

2023/08/18 Rental Team Dataset (bad rental team binaries only, OK teams removed for privacy):

https://mega.nz/file/rFMWTLRR#gndhqkbONLjaJNZfsDW_hdqL7ytKKfJDdOuRn6hsxvk

2023/03/31 Rental Team Dataset (bad rental team binaries only, OK teams removed for privacy):

https://mega.nz/file/qE81yQiJ#KDzNdKe3cAypNO9ghd3lxlIsF3a-vpAxwWM18jzfcNk

===

Scarlet/Violet 2023 Analysis Write-up

===

Thanks for your interest in these write-ups! When we collect data for the entire season, we like to see if there’s any underlying trends in how players acquire teams, spend resources, and how cheaters influence these statistics. You’ll find similar analysis for Sword/Shield and the first two regulations of Scarlet/Violet in the prior ones linked above, so expect to see similarly shaped charts below.

Downloading >800 rental teams was no easy task, but fear not, we have automations where we just feed rental code IDs and we get the teams to look at & judge quickly. Unlike Sword/Shield, the rental code data format does not include a 32-bit rental code upload ID, so we can’t tell you the rate of rental code uploads and the rate per each regulation. I expect it to be a similar curve to SW/SH; a high initial rate that tapers off, followed by a small uptick when Regulation D became the current meta. Regulation D had a ton of players sharing their rentals, likely because obtaining the meta-centric Pokémon species required playing & transferring from games other than Scarlet & Violet.

===

Worlds Uploads

===

After Worlds concluded, I definitely noticed the usual trend of non-Japanese players being reluctant to share their rental codes. So, here’s a breakdown of rental team uploader languages for WCS 2023:

As we can see, about two thirds of the players that shared their rental codes after the tournament were playing with the Japanese language, and 90% were either Japanese or English. Based on the attendance, I expected a more diverse distribution, but this disparity is likely because:

  1. Regulation D is not finished – in September, there will be two tournaments (one in Pittsburgh, USA and another in Barcelona, Spain).
  2. Players are fatigued/still traveling (has been about a week since Worlds started).
  3. Players fear the “unofficial hack checks” 🙂

I always find it funny when players share pokepast.es links (just EVs/Movesets), or a screenshot of the rental code screen (without the code!); these don’t benefit the community as they’re unusable inside the game. People want to playtest your teams! (and I want to check them too!)

“New” Hack Check

About a week before Worlds started, players received a “strong recommendation” to be wary about how they acquire their team. As players soon found out, their “new” hack check was able to find things that were not previously checked for, and quite a few players received team sheet errors for each (hacked) disqualified Pokémon. If enough Pokémon were disqualified (<4 remaining), they were disqualified from the tournament.

As linked in the Twitter chain, the community immediately suspected the “HOME Tracker”. As I’ve previously covered in Sword/Shield, the HOME Tracker is a 64-bit number that HOME uses to track the upload & movement of Pokémon data as it enters and leaves the cloud servers. Pokémon that are not correctly migrated through HOME (for example, from Legends: Arceus to Scarlet) will not have a valid HOME tracker. The community eventually started pointing the finger to the HOME Tracker, and players who were disqualified sent their rental codes to the community for verification; yep.

Trade Bots were never intended to provide tournament-ready Pokémon. For a bit of background, the vast majority of trade bots are running SysBot.NET, an automated platform using PKHeX with users being able to request trades from Discord or Twitch. These trade bots can create Pokémon, and they can also copy/modify Pokémon. The cheaters using trade bot Pokémon did not know that these trade bots can wipe HOME Tracker data even if they were transferred legitimately, as again, the trade bots are not for use in competitions. So they got caught by the “new” hack check. The HOME Tracker check is an incredibly simple check. They just check if the HOME Tracker value is zero!

TPC was also able to catch others with flaws I’ve discussed in previous tweets. These team(s) were not publicly shared. I enjoyed watching the Worlds' pairings pages, with players suddenly dropping from the tournament :)

Trends

I’ll include a bunch of charts summarizing a bunch of trends. The same 508 EVs, 3/3 PP Up trends continue to hold up.

One of the favorite memes in calling out generated teams is noticing the number of PP Ups that a player has. A usual hallmark is the overkill – most moves do NOT need 3/3 PP Ups used and most users aren’t going to waste the time applying all 72 of them. As always, 3/3 PP Ups does not indicate the team is hacked; these are just observations after a team has been labeled.

3/3 is sus indeed. Do you really need 64 PP for Haze?

Same for the 508 EVs – most players are not extremely precise (with the EV Wings) and end up hitting 510 on at least one of their team:

Players creating/modifying Pokémon import a 508 EV spread from Showdown.

Do players obtain their teams themselves, and how much help do others provide?

The more that were from the OT, the less likely they were to be hacked.

How many mints are used by players?

Lack of Synchronize (ability) means players need to use Mints to get the right nature.

How many times did players change the Tera Type of their team? (Different from initial):

Quite a few Tera Shards are consumed per team!

How many Pokémon were bred from an egg on a team?

Breeding sucks in Scarlet/Violet.

From what encounters (in Scarlet/Violet) do players obtain their teams?

It’s easier to capture a Wild Pokémon and share egg moves via Mirror Herb.

How many TMs are used per team? (about 10/11)

855 teams, 9107 used (10.65 avg, 2.66)

470 legal teams, 5022 used (10.69 avg, 2.67)

385 bad teams, 4085 used (10.61 avg, 2.65)

From which games did players obtain their teams for the World Championships?

S/V, SW/SH, and Legends: Arceus as expected, with some help from past games.

From which games did players obtain their teams for Regulation D?

 Looks quite similar to the above chart for Worlds.

Thanks

Closing

Was there more cheating in Scarlet & Violet? Definitely. Even though we see just under 50% of teams being noticeably hacked this time, we acknowledge that players (outside of Japan) are reluctant to share teams. Players have been emboldened to openly discuss genning, as evidenced by the plethora of discourse on Twitter/Reddit/YouTube/Facebook/Discord/etc.

I’ll repeat what I said at the end of the Sword/Shield writeup:

Will cheating be more common on the next console? Nope. Excluding another monumental screw-up in console security, the community will likely never have widespread cheats on the next console. Best hope the game developer makes strides in reducing all barriers to obtaining competitive PvP teams. GameFreak is renowned for their innovation pace, so we’re all optimistic! /s