After a discussion with the other leaderboard moderators, we came to the conclusion of allowing the use of the Korean PSN version for speedruns, which makes it the fastest known version. Here's the story:

It was previously thought that dubs didn't matter for Sly 2 (as opposed to Sly 3), but this was proven to only be the case for dubs available on the PAL version. Suappi realized that both Korean and Japanese have a faster dub than English. See his language comparison between English, Japanese and Korean here (PS2):

The Japanese version was never released to PSN, but the Korean version was. Oggy therefore tried his luck and ended up getting his hands on said version. After some initial testing, it was clear that Korean could save over 10 minutes! Here is a 5:16:40 run I did on the Korean dub, which is the current WR:

Here's the catch. In order to get your hands on the Korean PSN version, one needs a Korean PSN account. In order to make a Korean PSN account, one needs a so-called "I-Pin", only acquirable through the use of a Korean phone no. or something comparable to a Korean social security no. In other words: We can't simply make an account and purchase the game. We need to use an existing Korean account, or have a Korean person buy and download the game for us.

This is where the gray zone kicks in, and why this wasn't immediately allowed. The easiest way to get your hands on the version is to buy an account online, then add funds through 'Live Cards' and buy the game that way. It is specifically noted in the PSN Terms of Service ( https://www.playstation.com/en-us/network/legal/terms-of-service ) that selling/buying an account is not allowed. However, the ToS appears targeted to online play and also states that "cheat, exploit or use any bugs, glitches, vulnerabilities or unintentional game mechanics in PSN Services or any of its products or services to obtain an unfair advantage" is also not allowed. In other words, they indirectly claim speedrunning is against the ToS. The ToS is likely extremely wide so that Sony can strike whoever they want. ((Personal opinion))

We would like to make it abundantly clear that should runners get the Korean version of the game, they run the risk of those ToS violations.

Despite this, all four of the leaderboard moderators voted for allowing it, pointing at previous examples within the speedrunning community. PC speedrunners allow the use of torrenting games to obtain pre-patched versions of games to speedrun on, and communities have allowed niche, rare and/or expensive consoles in the past (as long as they were "official" consoles) to save time in loads and/or lag.

There was also talk about splitting the leaderboards to maintain the current leaderboards; that isn't happening. There was also some arguments for adding a 'language' variable, but that was shut down in favor of simply using the already existing 'Platform' variable.

Good luck to those who want to take the plunge into the Korean version for the time save. 😃

well put post

As much as I'm excited for time save, we did have a poll where majority of us voted to split the leaderboards. I know it was a poll by oggy who isn't a moderator, so it probably wasn't in your consideration, but the community did vote. I remember in some communities the difficulties of having certain things like this, (especially ones like this where it's pretty shady) like the iQue from OOT speedrunning. Just wanted to point that out as an example, I respect your decisions as moderators to do whatever.

Awesome! These times can be even lower because of this. 🙂

Unfortunately, a poll carried out in a twitch stream doesnt have any influence on how these boards are run. We are a team of moderators that have all been in the Sly series for longer than most. Keep in mind that viewers voting on a poll is entirely arbitrary as the viewers of people's streams really shouldn't affect how we as a speedrunning community runs. All mods agreed on this. If any RUNNER has any issues with this, they can bring it up with the mod team. We were all relatively vocal and we all agreed that adding another leaderboard was a complete waste of effort and simply too arbitrary to be considered.

Speedrunning is a hobby. And as with any other hobby, you'll need to spend money if you want the best equipment. This case is no different. Making a separate leaderboard for the Korean version would mean we would have to make a separate leaderboard for every single different version. We will not be flooding the boards with pointless leaderboards. Thank you.

As I said, I respect the moderators' decision. Overall, these types of decisions have gone over well in most communities, with exceptions being exceptions. With a smaller size community, I feel like you guys probably made the right decision.

Here are presented the excact time differences between English and Korean in order to make all future runs in Korean more comparable to runs done in English. This analysis does not include cut scenes that runners traditionally attempt to skip in order to save time. Although the skippable cut scene in Disco Demolitions (saves 2.44 seconds in Korean) and the 2nd skippable cut scene in Blimp HQ (saves 5.44 seconds in Korean) are included in this comparison. This comparison was done between Korean PSN and PAL PSN versions for PS3.

For the entire any% speedrun Korean on its own saves 9 minutes and 32 seconds when compared to English. Because of moderately cycle based Episode 6 and heavily cycle based Episode 8 and Episode 7 Operation the most accurate estimation with current strats is about 9:10, +/- 15 seconds, so still likely 9+ minutes.

Cairo is 7 seconds faster in Korean for any% and 29 seconds faster for 100%.

Episode 1 on its own is 64 seconds faster in Korean, 62 seconds faster after excluding the skippable cut scene in Disco Demolitions and the fact that it dosen't save time anymore even with SSD.

Episode 2 is 55 seconds faster in Korean.

Episode 3 is 64 seconds faster in Korean. (Rip-off the Ruby CSS saves only 0.9 seconds less in Korean so still definately worth it.)

Episode 4 is 81 seconds faster in Korean. (Note that the skippable opening cut scene for Code Capture is 9.5 seconds faster so going for CC FF CSS can save only 16-17 seconds on its own in Korean. This analysis did not take the possibility of CC FF CSS into consideration.)

Episode 5 on its own is 89.7 seconds faster in Korean, 83 seconds after excluding the opening cut scenes for Know Your Enemy (-4.97 seconds) and Tank Showdown (-1.97 seconds) due to tank cycles. If we assume that the photo of the blimp can be taken on your way to the tank then an estimation of 85 seconds is the most accurate.

Episode 6 on its own is 57.5 seconds faster in Korean, unfortunately train cycles in several missions exclude the time save in Korean completely or diminish it. The most accurate estimation is 51 seconds, +/- 5 seconds.

Episode 7 on its own is 96 seconds faster in Korean. The only mission where cycles matter in this Episode is the Operation where Korean possibly makes an early cycle possible for Bentley during his Eagle Egg run. Cut scenes leading up to this situation are 7.57 seconds faster in Korean but making that cycle saves a total of 15 seconds from what was ever possible in English. Missing that cycle on the other hand practically saves no time over English. So by not making that cycle Korean can be considered 88.5 seconds faster but by making it 103 seconds faster.

Episode 8 on its own saves 42 seconds when only the 1st cut scenes of Blimp HQ and Charged TNT Run (and the super long Blimp HQ cut scene of course) are excluded, 36 seconds when the 2nd skippable cut scene in Blimp HQ is excluded as well. Fan cycles in this episode make practical estimation very hard and we still haven't yet looked into properly that what we can do with shorter cutscene lenghts for this Episode. After also excluding the final cut scene in Blimp HQ that shows the inducers slowing down (-1.58) and opening cut scenes for Sly/Bentley (-3.45) and Bentley/Murray (-1.82) the solid time save in Korean still stands at 29.5 seconds if basically nothing cycle based is taken into consideration.

This sheet below lists lenghts of cut scenes in frames (59.94 fps) for splits and seconds for speedrunning categories.
https://docs.google.com/spreadsheets/d/1tqvKlH4md30owAiR4ym8c6FvXf49U7ZVrtx7GUL8pts/edit?usp=sharing

Just if people weren't sure if dialog skip was a thing in korean, Mikkel got it in a run
https://clips.twitch.tv/OriginalAnnoyingBubbleteaGrammarKing