Leon_Trotsky wrote on 09/12/19 at 22:21:03:
In my experience, in openings like the Pirc and French Winawer where Black deliberately cedes space and assumes a "turtle" approach, it gives outrageously high White evaluations, like +1,30, when Komodo would give around 0,00 to 0,10.
The question is: Who is correct?
Have you followed the given lines by Leela, figured out where she diverges from SF/Komodo/etc, and gotten the impression that she just misses all avenues of counterplay, or why she so grossly 'overestimates' the chances?
https://lichess.org/analysis/standard/2r1r3/1p3nkp/pP1pn1p1/P2N1p2/2P2P2/R3P2P/6... in this sample position out of a Dutch Leningrad, SF @depth32 spits out +0.5 with a variety of moves (Rc3, Rc1, Ra4, Bf1). White Advantage, but nothing dramatic.
https://www.chess.com/computer-chess-championship#game=44 Leelenstein did go for one of the SF suggestions (35.Rc3), but thinks White is completely winning (+2.6), and btw has been thinking so.. for pretty much the entire game!
The initial position after the opening book is evaluated by SF as +0.29, and by LS as a whopping +1.59
As late as move 54, SF believes to be in a tenable position (+0.76), while Leelenstein is long confused why there's no resignation coming (+4.26)..
After 55.Bf1, SFs eval suddenly explodes in White's favour, and never recovers.
Did Leelenstein overestimate White's chances here..?