Continuing his epic voyage to explain why various old fuseki patterns have disappeared in the AI age, Meijin Shibano Toramaru has turned his attention to the dangerously endangered species - the pincer. I think Bill was the first to alert us to this side effect of digital warming, but I don't think he - or Shibano or anyone else - has satisfactorily explained why pincers have gone to the brink of extinction.
Oddly enough, I think I may have discovered the reason, though quite accidentally. I will leave the details to another post, but (if I'm right) it's connected with another AI feature that Bill is currently forcing us to re-examine - the press. My own clue, however, was the concept of the barmkin.
Shibano, at least for the time being, dodges the reason for pincer endangerment and focuses instead on the survival of one particular pincer.
Shibano notes that for a long time the teiban (定番) joseki against the high approach was A. A teiban (short for 定番商品) is a plain standard product that sells consistently well. We might call it the plain vanilla cornet that mature adults prefer. Kiddies may be tempted by extra hundreds and thousands and monkey's blood, and (judging by adverts I've seen on tv under enforced lockdown) ice-cream vendors may try to seduce the grown-ups with flavours such as ruby cocoa beans - I think they mean chocolate but organolepsis may prove otherwise - but plain vanilla never fails the true aficionado.
But the go board is a cruel place, and it seems that tastes have changed. Plain vanilla still exists but comes in fancy new packaging - a safe tub instead of a wobbly, melting cornet!
The new variety here is the low two-space pincer.
This pincer itself is not new - in fact this year is the 200th anniversary of its appearance. It has, however, been especially popular in recent years, though it cannot claim to be a child of the AI revolution. It is rather a child of the naughty noughties. A millennial. What is special about it is that it has apparently survived the AI onslaught.
AI killed off the Magic Sword, which had previously been assessed by humans as good for the sword holder (Black here. The reason was that Black typically got magnificent thickness in return for White profit in the corner. But in gote. It is now well known that AI has shown that thickness may appear as terrifying as a chimera but it is often just as much a phantasm. In the case of the Magic Sword, according to Shibano, AI rates sente as more than enough compensation for the awe-invoking chimera.
The low pincer apparently does not suffer from this defect. Shibano only gives a hint as to why, but one aspect appears to be that the low stone is easy to sacrifice - to treat lightly - and as a result Black is the one that can usually get sente. He does not then get an advantage - just parity. But that's enough to keep it in the armoury, especially for formations such as the one on the right, which is the one Shibano shows.
It appears that White's favoured response now is the odd-looking move at B (which I think I now understand as a result of my accidental discovery). I only have nine examples of that in the GoGoD database, It is an AlphaGo innovation, but FineArt has also played it. In fact, the three AI examples are all against other bots. But Shibano gives a couple of examples that don't appear in the database, so it would seem that most activity in this area is in trial internet games.
I hope Shibano will come back to this topic next week, but in the interim it would be interesting to hear opinions as to why this low pincer may have found a way to survive against the odds.
|