At position R16,pass,O17,Q16,Q15,R17,P16, @aesalon started the conversation:
I’m pretty uncertain of this one. It was played somewhat frequently(10-20% of the time) until late 2018. Mistake might be too harsh but the pro sequences don’t seem to merit inclusion.
( FYI @mark5000 )
The explanation for “why is this a mistake” seems sound though: white could have been killed, but instead gets the corner in sente?
( FYI @mark5000 )
@aesalon Should low-probability pro play (rarely last year, and never this year) ever trump a source talking about the position? I think the answer must be no, because the source presumably knows what he or she is talking about, otherwise the source would be invalid.
However, pro game examples contradict this claim. White is not taking the opportunity to live in the corner and is fighting instead. Doesn’t seem like a very sound explanation.
( FYI @mark5000 )
And if it was being played in a not insignificant amount of pro games after the authors publication, should that be considered when judging the move?
( FYI @mark5000 )
That is the question then. Invalid is a strong term but you have also admitted that this particular source is not among the strongest. It is another issue of contradictory sources and fact-checking.
I tagged the move Good pending further discussion. It comes down to what information we consider more trustworthy:
I think the best synthesis of this information is pretty close to what it is now. The Good label better reflects that some professionals played it, that none has called it joseki, and that one nearly-pro-level player says it’s not.
Oh I messaged you, going to paste it here:
Just saw the change of https://online-go.com/joseki/25364. It is a discussion of the contradicting sources and how we might fact-check source statements(I guess I wonder if you believe we should at all?)
Since our criteria for where the line of Good -> Mistake is still unclear, I think it’s more likely that it was correctly labeled a mistake. I want to find out how we can come to an understanding[and work on documenting it for the sake of consistency] on why that is our conclusion.
The evidence that the source gives for claiming it a mistake is also a bit of a red-flag. Other sources(there is actually a Chinese article on the position, which also calls it a mistake but article and book Mistakes definitely don’t always equal Joseki Explorer mistakes) seem to point out that the refutation is not “taking the corner in sente” but instead “cut and fight”
edit: Obviously I should share the other source. Please ignore the top left stone, the search space was 10x10. I made it smaller than 13x13 because I wanted to cover various situations (since we aren’t even considering the possibility it is an IDEAL move to be played in all situations) Here is the SGF game tree
If this is the case, why say that “Mistake might be too harsh” and that “it was being played in a not insignificant amount of pro games after the authors publication”? I took for granted that your latter statement was correct, that pros had played it a not insignificant number of times, and so it must be at least situational to those games.
I’d like to move the ball forward, so I’d appreciate more clarity on what changes you’re advocating for if not a change in labeling.
Because those are both true statements. It is not clearly a mistake… since we do not have criteria for it maybe… under a lot of criteria that we have each considered before, it could be Good or a Mistake.
To start, change the mistake(if we ultimately decide it is a mistake) variation and description. Your variation of Play Go at online-go.com! | OGS includes O18 which woudl be the worst move in the sequence. How can that be a proper refutation?