I ran into an interesting position having just won the exchange where
material balance against positional factors meant that the right thing
to do was to immediately swap my rook back for a bishop and pawn. I
didn't play this and ended up in a blocked pawn ending RRB vs BBR and
a draw. Analysing it on various engines afterwards shows the extent to
which different engines factor in material vs positional factors. The
position is:

2r2rk1/1p3ppp/p3p3/PqBpP3/1Pb2P2/5B2/2PQ2PP/3R3K b - - 0 23

I actually played Rfe8 which at the time I thought was the best
continuation. I wasn't brave enough to go for RxB which I thought
might win, but traded what I believed was a bigger advantage for a
mere pawn.

I think the best play for white is to move the bishop to safety and
then block the position. After Bd4 it is quite hard for black to break
through with an attack. Crafty seems quite lost in this position, and
Fritz8 struggled a bit at first but would have played the best move at
tournament rates.

However the various engines on overnight analysis see things
differently:

Shredder at ply 25 sees

-1.82 1. ... Rxc5 2. bxc5 Qxc5 3. Rb1 f5 4. Qc7 Re1
-1.56 1. ... Rfe8 2. Qe3 Qa4 3. Qc3 Rxc5 4. bxc5
-1.33 1. ... Rfd8 2. Bb6 Re8 3. Ra1 Qd7

Rybka at ply 21

-1.28 1. ... Rxc5 2. bxc5 Qxc5 3. Rb1 Bb5 4. Rb3 Rc8
-0.76 1. ... Rfe8 2. Bd4 Qd7 3. Re1 h6 4. Bh5 Rc7
-0.76 1. ... Rfd8 2. Bb6 Re8 3. Bd4 Qd7 4. Re1 h6

Fritz at ply 20

-1.12 1. ... Rxc5 2. bxc5 Qxc5 3. Rb1 Bb5 4. f5 Qc7
-0.84 1. ... Rfe8 2. Bd4 Qd7 3. Bc5 Qa4 4. Rc1 Qb5
-0.81 1. ... Rfd8 2. Bd4 Qd7 3. Bc5 Re8 4. Qf2 Qa4

Crafty at ply 15 (finally sees RxB as best)

-1.31 1. ... Rxc5 2. bxc5 Qxc5 3. Rb1 Bb5 4. f5 Rc8
-1.25 1. ... Rfe8 2. Qf2 Rcd8 3. Bg4 Qa4 4. Kg1 Qa3
-1.23 1. ... Rfd8 2. Bg4 Rd8 3. Qf2 g6 4. Kg1 Qa4

I suspect that Shredder overemphasises structural factors and that
Rybka is about the most realistic.

Regards,
Martin Brown

i thought that was interesting how the different programs played that
out..