Kevin Pacey wrote on Mon, Jul 22, 2019 02:52 AM UTC:
Hi H.G.:
Chess Assistant16 doesn't provide the engine that it used for evaluations, regrettably. However I experimented in another way (though not as you would prefer), using Houdini3 (released 2012) as an engine instead. In the exact position in the line culminating with 9.Qc2-c3 that I gave, after some considerable time to look, Houdini3 evaluates the position as 0.15 (or "+=/="), somewhat better than the 0.04 of CA16's engine (as an aside, a disbelieving master friend dared me to play this line as Black against him in any tournament, back in the 1980s, and gallingly I lost the one time I obliged - in spite of previously winning a number of nice master-level games with it, at least sometimes attacking with queens on the board).
I tried replacing the B on c1 with a White knight (a fairly lousy square for a Kt), and this time Houdini3 gave the position as -0.12 (or "=") after considerable time spent. I then tried looking at the same position but with the White knight on d2 (not a great square completely for a knight, either, but it allows White to develop at least a move faster, probably); in this case Houdini3 rated the position as exactly 0.00 (or "="). Note: replacing the Kt on d2 with a dark-squared White bishop gives a position Houdini3 evaluated as 0.17 (or "+=/="); White can formally develop all his pieces a move faster (probably), but he'll want to eventually improve the lot of that piece (White B) now on d2 - though that seems true if a White Kt were there, too.
It may be worth noting that CA16's engine rates the Nimzo-Indian Defence (position after 3...Bf8-b4) as worth 0.07 (or "+=/=") for White, while Houdini3 gives it as 0.16 (or "+=/=") for White. CA16 considers the Nimzo-Indian complex of defences as numerically Black's best defence to 1.Pd2-d4 (i.e. with 'best' play).
Hi H.G.:
Chess Assistant16 doesn't provide the engine that it used for evaluations, regrettably. However I experimented in another way (though not as you would prefer), using Houdini3 (released 2012) as an engine instead. In the exact position in the line culminating with 9.Qc2-c3 that I gave, after some considerable time to look, Houdini3 evaluates the position as 0.15 (or "+=/="), somewhat better than the 0.04 of CA16's engine (as an aside, a disbelieving master friend dared me to play this line as Black against him in any tournament, back in the 1980s, and gallingly I lost the one time I obliged - in spite of previously winning a number of nice master-level games with it, at least sometimes attacking with queens on the board).
I tried replacing the B on c1 with a White knight (a fairly lousy square for a Kt), and this time Houdini3 gave the position as -0.12 (or "=") after considerable time spent. I then tried looking at the same position but with the White knight on d2 (not a great square completely for a knight, either, but it allows White to develop at least a move faster, probably); in this case Houdini3 rated the position as exactly 0.00 (or "="). Note: replacing the Kt on d2 with a dark-squared White bishop gives a position Houdini3 evaluated as 0.17 (or "+=/="); White can formally develop all his pieces a move faster (probably), but he'll want to eventually improve the lot of that piece (White B) now on d2 - though that seems true if a White Kt were there, too.
It may be worth noting that CA16's engine rates the Nimzo-Indian Defence (position after 3...Bf8-b4) as worth 0.07 (or "+=/=") for White, while Houdini3 gives it as 0.16 (or "+=/=") for White. CA16 considers the Nimzo-Indian complex of defences as numerically Black's best defence to 1.Pd2-d4 (i.e. with 'best' play).