The 8p wasn’t mocking the guy who was stuck at 1k for 10 years, she meant that improving by many stones in one year is unlikely for someone who stayed at 1k for 10 years.
7 Likes
The trick is to have more qualified parties looking for them >.<
Edit: I wonder what a qualified third-party even looks like. I don’t imagine Go has the resources to fund a global anti-AI team. Just implement the camera rule!
2 Likes
I read it as clear mocking. 
I hadn’t seen that, thanks for pointing it out.
1 Like
Idk I watch Yoonyoung’s streams occasionally, and she really doesn’t seem like the type to outright mock anyone.
The fact that a player had a stable rank for 10 years is certainly relevant information.
5 Likes
It is, but the way to write it varies.
idk, I never warmed up to her videos.
1 Like
Absolutely not. You assume the person tried to improve in the past 10 years. If the guy is 50, 40-50 is a very busy family time period. Now he might be empty nested and tried to avoid spending time with his wife in the house. 
Also the last year or so was tailor-made for people who want to sit inside and study hard. 
2 Likes
I don’t believe a 50 year old guy can improve that fast, but an unmarried guy below 30 who doesn’t have a demanding job, maybe. In that case, as the 8p puts it, he is talented enough to become the next US pro.
4 Likes
Indeed. Besides that 1 mistake, the graph indicates that for the rest of the game, black was playing at a level of 10d EGF (think World Champion level, like Go Seigen or Shin Jinseo at their best). That level is not as strong as AI at high playouts (superhuman level of about 12d EGF), but still consistent with AI at lower playouts.
Also, this black player has been stuck at 2k-1k AGA (~4k EGF) level for 10 years already. How likely is it that they became 13 stones stronger overnight to reach World Champion level?
The unlikelyhood of this is not proof without doubt that he was cheating, but I’m more than 99.9% certain that either they used AI, or Shin Jinseo was sitting next to them (although I doubt that Shin Jinseo could play at his peak level when playing this fast).
10 Likes
Just to be clear since you mentioned the OGS AI and I don’t know which AI you were using to check I am pretty sure the OGS AI is a 20 block model so it will show very different results if we assume that he used a 40block model.
(I have not done any checking of my own btw just making it clear that the AI he probably used would have likely “NEVER” been a 20 block model)
1 Like
Sure, you quote the 8p, nice.
Just the AI on this site. Is 40 block AI better than 20?
1 Like
Yes 40 block AI is better then 20 block AI
1 Like
I was hoping you say the other way around. Then the 40 block AI would still say move 105 is horrible and he would not make that move, unless the tool also says if you do that, I can still make you win the game. 
1 Like
As I said, I don’t think 105 is really relevant for the allegation, nor is the distinction between 20 or 40 blocks. Fact of the matter is that in the majority of the exchanges Black made better ones than White did. It’s the way Black can outplay White consistently throughout the game that is suspicious, and 105 being a bad move does not change that fact.
As I also said before, it appears like the roles of 6d and 1k are switched. But since we can assume that the 6d is not cheating, White not only plays as good as a 6d, but plays doubly as good, hence more like a professional player.
6 Likes
My 15-block Katago says that after move 105, Black is only 4 points behind. If Black cheats with an AI, then it’s easy to catch up 4 points in the endgame against a 6d.
5 Likes
what is the negative % drop of move 105 on your ai?
1 Like
Before move 105, winning percentage was 95% and Black was leading by 24 points.
After move 105, winning percentage became 19% and Black was behind by 4 points. That’s certainly a huge drop, but someone who wants to make a deliberate mistake can certainly win with confidence with an AI at hand.
I am not saying there is any proof that move 105 was a deliberate mistake, just that move 105 is compatible with cheating with AI.
6 Likes
That I don’t agree with you. If I use AI to cheat, I only play winrate. I do understand 95% winrate does not guarantee much on the win, cause it is based on the ideal moves and the point difference may be very low.
To recap, if we debate the likelihood he cheated, I would always agree it is very high.
If I were the 6d,
would I feel frustrated? 100%
would I complain to AGA? 19%
would I whine publicly? 1%
would I regret I whined publicly a week later? 95%
Do I still feel the 1kyu not cheated? 1%
Do I dislike team montreal’s actions? 99%
2 Likes