(A) 42. h6!? is a stronger winning try for White than we thought at the time.
After 42...Bxh6 43. g7 Bxg7 44. Rxg7 our first intent was 44...b2, but this
is once again premature: 45. Rg1 Na5 (else Black loses the b-pawn) 46. Re1!
cuts off Black's King, and after 46...Nc4 47. Bb4 e5 48. Kg2 e4 49. Rb1
Black will lose the b-pawn. This and the line A1122) below with 49. Bc3+!
was found by Ross Amann in after-the-fact analysis with Regan. Also 44...Na5
allows White to organize a blockade: 45. Bb4 Nc4 (45...b2 46. Rg1 Nc4 47. Re1!
transposes to the last line) 46. Rg3 b2 47. Rb3 Ke4 48. Kg1 and Black can do
nothing more while White's King enters the scene. Black must re-centralize:
44...Ke4!
[DIAGRAM?]
F11) 45. Rg3 and now:
F111) 45...Kd5? illustrates the delicacy of the position: 46. Ba3! and now:
F1111) 46...b4 47. Bb2 and now:
F11111) 47...Na5? 48. Rg5+ +/-
F11112) 47...Nd4 48. Kg2 Kc4 49. Kf1 and I think Black cannot hold it.
F11113) 47...Kc4 48. Kg2 and again, it seems Black's N cannot find good
squares in time, e.g. 48...Na5 49. Kf1 Kb5 50. Rg5+ [this is
actually still pretty complicated but seems to win].
F1112) 46...Na5 47. Bb2! (not 47. Bb4? b2!) and now:
F11121) 47...Nc4 48. Rxb3 Nxb2 49. Rxb5+ is the fatal flaw of 45...Kd5.
F11122) 47...Kc4 48. Kg1! Kb4 (else what is Black doing?) 49. Bc3+! Ka4
50. Rg8! and now:
F111221) 50...b4 51. Rg4! Nc6 52. Bb2 Kb5 53. Kf1 Na5 54. Rg5+
transposes into F11113).
F111222) 50...e5 51. Rb8! b2 (51...Ka3 52. Ra8! b2 53. Rxa5+ Kb3
54. Bxb2 +/-) 52. Bxb2 Nc4 53. Bc1 and Black cannot support
the other b-pawn well.
F1113) 46...Kc4 47. Bb2! seems to leave Black little more than
transpositions after 47...Na5 48. Kg1 or 47...Nd4 48. Kg2.
F112) 45...Na5! is the correct timing. Now:
F1121) 46. Bb4 b2! gains a key tempo: 47. Rg1 Kd3! 48. Bxa5 Kc2=.
F1122) 46. Ba3 Nc4! and now:
F11221) 47. Rxb3 Nxa3 48. Rxa3 Kd4/d5! EGTB= (8/8/4p3/1p6/4k3/R7/8/7K+b)
is the single most important reason Black is holding line A);
note that if White's King were on g1 White would win
(8/8/4p3/1p6/4k3/R7/8/6K1+b #24).
F11222) 47. Bb4! White's tempo loss makes a strange impression, but
it prevented 46...b2, and since now 47...b2 48. Rb3 is a
blockade that allows White's King to enter, Black's next is
forced: 47...Ne3!
[DIAGRAM?] Kamikaze time for Black's Knight.
F112221) 48. Bc5 b2! 49. Rxe3+ Kd5 50. Rb3 Kxc5 51. Rxb2 b4/Kc4 EGTB=
(8/8/4p3/1pk5/8/8/1R6/7K+b). Again Black would be lost with
White's King on g1 (8/8/4p3/1pk5/8/8/1R6/6K1+b #25).
F112222) 48. Bd2 Nf1 49. Rg4+ Kd3 lets Black penetrate.
F112223) 48. Rg1 Kd3 ditto. Notice that 48. Rh3, 48. Kg1, and 48. Kh2
are all "illegal". Since 48. Ba3 Nc4 repeats, White has only
F112224) 48. Bc3. White seems to be doing well, with control of the
long diagonal. However, such is the basic strength of Black's
centralized King and Knight that Black seems to have no fewer
than three separate ways to draw this:
F1122241) 48...b4!? and:
F112224121) 49. Bxb4? b2 50. Rg1 Kd3=.
F112224122) 49. Bd2 Nf1 draws as in F112222).
F112224123) 49. Bf6 (49. Bg7 Nf5=) and now:
F1122241231) 49...e5!? 50. Bg5 (50. Rg1? Nc4 =+; 50. Bxe5 Kxe5
51. Rxe3+ Kd4 52. Rxb3 Kc4=) Nd5 (50...Kd3
51. Rxd3+ Kc2 52. Re2+ Kb1 53. Bf6! +/-) 51. Rxb3 Kd4
may be AOK for Black. For example, after 52. Kg1 Kc4:
F11222412311) 53. Rb1 b3 54. Rc1+ Kd3 55. Rd1+ Kc4 and:
F112224123111) 56. Rd2 e4 57. Kf2 e3+! 58. Bxe3 Nxe3 59. Kxe3
Kc3! is EGTB= (8/8/8/8/2k5/1p2K3/3R4/8+b)
F112224123112) 56. Bc1 Nb4! 57. Rd8 Nd3 58. Rc8+ Kb4! 59. Bd2+
Ka3! 60. Ra8+ Kb2 61. Be3 (61. Kf1 Kc2 and
62...b2=) Kc2 62. Rc8+ Kd1 and White seems to
have nothing better than repetition.
F11222412312) 53. Rg3 b3 54. Bc1 Nf4! leads to an even better
penetration than the last line.
F1122241232) 49...Kd3! and now:
F11222412321) 50. Kg1 (50. Bg5 b2=) Kd2! White's King is boxed
out owing to the check on d1. Now
F112224123211) 51. Bg5 b2! is still =.
F112224123212) 51. Rg7/g8 Nc4 52. Rb7/b8 b2 53. Rxb4 Kc2
54. Rxc4+ Kb3 is bang-bang =.
F112224123213) 51. Bh8/a1 e5! (MacChess does find this, quickly)
52. Bxe5 Nc4 and White must give back a piece.
F112224123214) 51. Bd4 Kd3!! (this is when the analysis is most
fun!) and since 52. Bxe3 b2 53. Bc1+ Kc2 54. Bxb2
Kxb2 is EGTB= (8/8/3p4/8/1p6/6R1/1k6/6K1+w), White
has nothing better than repetition.
F11222412322) 50. Rh3/Bh8 Kd2! is no better for White.
F112224124) 49. Bb2/h8 Kd3! also seems no different from the last lines.
F1122242) 48...Kd3!? An accelerated version of the 48...b4 49. B-moves
Kd3 idea. Since 49. Ba1/b2/e5/f6/g7/h8 Kd2! 50. Bd4 Kd3!
51. Bxe3 b2 52. Bc1+ Kc2 53. Bxb2 Kxb2 is still EGTB=
(8/8/4p3/1p6/8/6R1/1k6/7K+w; note 8/8/4p3/1p6/8/6R1/1k6/6K1+w
is +/- with WK on g1), play seems no worse for Black than in
F1122241). But Black need not grovel---Black can play as if
for the win with
F1122243) 48...Nd1! and now:
F11222431) 49. Bf6?! Nf2+! Black's Knight will reach the ideal square
d3, and White's pieces step on each other's toes.
F112224311) 50. Kg1 Nd3 and now:
F112224311) 51. Kf1 b2 =+ as White's King blocks the Rook.
F112224312) 51. Rg2 e5! and White had better start looking
for a draw.
F112224313) 51. Rg5 e5! Black may be winning here.
F112224314) 51. Kh2 (ugh!) e5 52. Rg1 is best and just =.
F112224312) 50. Kg2 Nd3 seems no better than 50. Kg1.
F112224313) 50. Kh2 e5! (White's B on f6 is loose) 51. Rg1 b2
52. Kg3 (51. Kg2 Ng4 and 52...Kd3 =+) Kd3 53. Bxe5 Nd1!=,
another fun move. But here there is also nothing wrong
with 52...Nd3 and 53...Ke3 or 53...Kd4 next.
F11222432) 49. Bg7 Nf2+ 50. Kg1/g2/h2 Nd3. White played to improve on
previous lines, as 50. Kh2 e5 was not on and now 51. Rg4+
is possible. But 51. Rg4+ Kf5!? leaves White with little
to show as 52. Rd4? b2 is a loss, and 51...Ke3 is also
fine here. And 51. Bc3 b4! shows how dominated White's
Bishop is.
F11222433) 49. Bd2 Nf2+ 50. K-any Nd3 51. Bc3 b4 transposes into the
last note.
F11222434) 49. Be1! b2 50. Rb3 Kd4 51. Kg2 Kc4 52. Rb4+ Kd3 also
lets Black in.
So this major line is not only a draw, but shows Black equalizing without
much groveling. White's alternatives to 45. Rg3 either transpose or
cause Black less trouble:
F12) 45. Ba3 and now:
F121) 45...Kd5? 46. Rg3 transposes into F111)
F122) 45...Kd3!? may be OK for Black, e.g. 46. Rg3+ Kc2 47. Bb2 Na5
48. Bf6/g7 Nc4 49. Rg3+ Ne3! transposes into F1122242) down a tempo,
so 50. Kg1! Kd2 51. Bd4 Kd3? is no longer on, but 51...Nc2!
52. B-moves Ne3 may not give White anything better than repetitions.
F123) 45...Na5! is the thematic reply. Now
F1231) 46. Rg3 Nc4! transposes into F1122).
F1232) 46. Rg4+ Kd3 and:
F12321) 47. Rb4 Kc2 48. Rxb5 Nc4 49. Rc5 Kd3 50. Rb5 Kc2=.
F12322) 47. Bb4 b2!=
F12323) 47. Rg3+ Kc2 48. Rg2+ (anything else?) Kd3 and now:
F123231) 49. Bb2 Nc4 50. Ba1 b2! 51. Bxb2 Nxb2 52. Rxb2 Kc4!
EGTB= (8/8/4p3/1p6/8/3k4/1R6/7K+b).
F123232) 49. Bb4 Nc4 50. Rg3+, and both 50...Kc2 and 50...Ne3
are fine for Black.
F1233) 46. Rg5 Nc4 47. Be7 b2 (47...e5 48. Rg3!?) 48. Rxb5 Kd3 49. Rb4
Kc2! 50. Rxc4+ Kb3 51. Rb4+ Kc2=.
F1234) 46. Rd7 Nc4 and 47...b2=.
F13) 45. Rb7 Kd5 46. Ba3 (46. Bf8 Kc4 and White has to play 47. Ba3 anyway
as 47. Rc7 b2! loses) Kc4, and now:
F131) 47. Rb6 b4 and now:
F1311) 48. Bb2 Kc5 49. Rb7 Na5. Black has won many tempi and will
equalize with ...Nc4 soon.
F1312) 48. Bc1 Ne5! 49. Rxe6 Nd3=
F132) 47. Rc7 b4 and now:
F1321) 48. Bb2 Kb5 49. Rh7 (to oppose 49...Na5) e5! and now:
F13211) 50. Kg2 Na5! 51. Bxe5 (51. Kf1/f3 Nc4 52. Rh2 Nxb2 53. Rxb2 Kc4
is too late) Nc4 52. Bd4 b2 53. Rh1 Ka4=.
F13212) 50. Rh5 Ka4! 51. Kg1/g2 Na5 52. Rh4 Kb5! essentially transposes
into F13211).
F1322) 48. Bc1 Kb5 seems no better for White.
F133) 47. Kg1/g2 Na5! 48. Rb6 Kc3! 49. Rxb5 Nc4=.
F134) 47. Bb2 b4 48. Kg1/g2 Na5 49. Rb6/b8 Kc5 and White cannot hold off
...Nc5 indefinitely.
F135) 47. Bc1 Ne5! is completely fine for Black.
Maybe a little more needs to be filled in, but that is the basis of Black's draw.