White 28
 

Here's a consolidation of our analysis (for after [B]27 bxc3[/B]) from the Move 27 thread.

- -

If [B]27 ... Nd5[/B], then [B]28 Re1[/B] would deter 27 ... e5 and threaten both Bxd5 and Nxe6. Black can't guard against both threats with a single move, so we'll win at least another pawn and be two connected passed pawns up.

- -

If [B]27 ... Nxd4[/B] we can't afford 28 cxd4 yet, but
[B]28 Re1[/B] (not 28 Rc1 Ne2+) _does_ threaten 29 cxd4, thus finishing our solution to the d4-fork problem.
Now the N/d4 _must_ retreat or be captured.
(Black has two other tries, 28 ... Nc2 and 28 ... Nb3, but both would be to our advantage:
If 28 ... Nc2 29 Rc1 and the knight has nowhere safe to go.
If 28 ... Nb3 29 Rb1 pins the N/b3 against the R/c8. 29 ... Nxd2 30 Rxb8 Nxc4 does not let Black escape with two pieces for the rook:
After 29 ... Nxd2 we must instead play
30 Nd6+ to avoid losing that knight to 30 ... Nxc4.
Black now has two king moves, but both lead to material loss for Black. (This is a payoff for our having built a strong queen-side position.)
If 30 ... Kd8 31 Rxb8 threatens 32 Nxc8
If 31 ... Kc7 32 Rxc8+ Nxc8
(not 32 ... Kxd6 33 Nb7+
If 33 ... Ke5 34 Rc5+ Kf6 35 Rxa5 and we're up the Exchange plus a pawn,
If 33 ... Kd7 34 Rd8+ Kc7 35 Rxd2 we're a rook up. If now 35 ... Bxc3 36 Rc2 Nd5 37 Bxd5 exd5 38 Nxa5 d4 39 Nb3 and we're still up a rook)
33 Ne8+ Kd8 (or Kb6 34 Nd7+ and 35 Nxg7) 34 Nxg7 e5 and we're a bishop up.
or
If 30 ... Kf8 31 Rxb8, then
If 31 ... Kg8 32 Nxc8 and we're a rook up.
If 31 ... Bxc3 32 Nxc8.
If 31 ... Nc6 or ... Nd5 or ... Nf5 32 Rxc8+)

The only two safe retreats are
[B]28 ... Nc6[/B] or [B]28 ... Nf5[/B], whereupon (either case) we capture Black's e-pawn with
[B]29 Nxe6[/B] and we're a pawn up.

- -

[B]27 ... e5[/B] would be met by
[B]28 d5[/B] (not 28 dxe5 mundane pawn swap favoring Black) which forces [B]28 ... Na7[/B] or [B]28 ... Nd8[/B].

Now
[B]29 Nd6+[/B] forces either
[B]29 ... Kd8[/B] (if 28 ... Na7) or
[B] 29 ... Kf8[/B] where [B]30 Rf1+[/B] is uncomfortable for Black.

or we could make different threats with

[B]29 Re1[/B] (instead of the 29 Nxa5 pawn grab) which aims at both the e-pawn and, behind it, Black's N/e7 and K/e8, threatening [B]30 Nxe5[/B] and [B]31 Nxg6[/B]. If [B]30 ... Bxe5[/B] [B]31 Rxe5[/B] threatening 32 d6 if Black doesn't move his king. (And his a5-pawn is still undefended.)

- -

[B]27 ... Nb4[/B] -- where the knight is, yes, temporarily immune from capture because of the d4-fork (28 cxb4 Bxd4+ and 29 ... Bxa1). Why might Black prefer this to 27 ... Nxd4 (besides foolishly hoping we hadn't yet seen the d4-fork) ? Maybe for a followup move by that knight (Na2, Nc2, Nd3, Nbd5 or Na6) -- so we'll need to look at each of those in our following analysis to see whether there's some real threat.

Anyway, just as with Nxd4 we must refrain from immediate capture.

Possible responses for us probably should include taking the main sting (king and rook fork) out of the d4-fork. A king move or a rook move would suffice for that, as long as we retain the possibility of meeting ... Bxd4+ with Be3 so as to guard our N/c5 which still remains vulnerable to the (king and knight fork) portion of the d4-fork.

[B]28 Rb1[/B] threatens 29 cxb4. However, that does NOT mean that we could meet a subsequent 29 ... axb4 with 30 Rxb4 or Bxb4, because our B/d2 is still tasked with responsibility of meeting ... Bxd4+ with Be3 to guard our N/c5. But that b4-pawn wouldn't be going anywhere soon, and we'd pick it up in the near future.
What about [B]28 ... Na2, Nc2, Nd3, or Nbd5[/B] follow-up? Each could be answered with [B]29 Rxb8[/B].
If [B]28 ... Na6 29 Rxb8 Nxb8 30 Nd6+ K-moves 31 Nxc8 and 32 Nxe6[/B] puts us up two (connected, passed) pawns.

[B]28 Rc1[/B] threatens 29 cxb4.
If [B]28 ... Na2[/B] annoys, but [B]29 Ra1[/B] forces either [B]29 ... Nb4[/B] where we can do [B]30 Rb1[/B] (so 28 Rb1 might have been better), or [B]29 ... Nxc3[/B] where [B]30 Bxc3[/B] stops all remaining d4-fork threat.
If [B]28 ... Nc2 29 Rxc2[/B] and [B]29 ... Rb1+[/B] is met by [B]30 Rc1[/B].
If [B]28 ... Nd3 29 Nxd3[/B]
If [B]28 ... Nbd5 29 Nd6+ K-move 30 Nxc8 and 31 Nxe6[/B]
If [B]28 ... Na6 29 Nd6+ K-move 30 Nxc8[/B] Here,
if 30 ... Kxc8 31 Nxa6 Black can't threaten much with his rook and our N/a6 can hop back safely to c5 ... and we're a bishop up.
if 30 ... Nxc8 31 Nxe6+ Ke8 or Ke7 32 Nxg7 and Black can't trap our knight: e.g. 32 ... Kf7 33 Bxh6 g5 34 Be4 Rb6 35 Rf1+ Kg8 36 Nf5 Kh7 37 Bxg5. We stay at least a piece and pawn up.
if 30 ... Nxc5 31 Nxe7 Nd3 32 Nc6+

[B]28 Rd1[/B] threatens 29 cxb4.
If [B]28 ... Na2 [/B][B]29 Nxa5[/B] and we're two pawns up.
If [B]28 ... Nc2 29 Rc1[/B] and if[B] 29 ... Nxd4 30 cxd4 Bxd4+[/B] either [B]31 Be3[/B] or [B]31 Ne3[/B] guards our N/c5.
The next two are the same as after 28 Rc1:
If [B]28 ... Nd3 29 Nxd3[/B]
If [B]28 ... Nbd5[/B][B] 29 Nd6+ K-move 30 Nxc8 and 31 Nxe6[/B]
If [B]28 ... Na6[/B][B] 29 Nd6+ K-move 30 Nxc8[/B] Here,
if 30 ... Kxc8 31 Nxa6 Black can't threaten much with his rook and our N/a6 can hop back safely to c5 ... and we're a bishop up.
if 30 ... Nxc8 31 Nxe6+ Ke8 or Ke7 32 Nxg7 and Black can't trap our knight: e.g. 32 ... Kf7 33 Bxh6 g5 34 Be4 Rb6 35 Rf1+ Kg8 36 Nf5 Kh7 37 Bxg5. We stay at least a piece and pawn up.
if 30 ... Nxc5 31 Nxe7
if 31 Nxa4 32 Nc6+ forking K and R.
if 31 Kxe7 dxc5 we have doubled isolated c-pawns, but we're up a bishop.

[B]28 Re1[/B] threatens not only 29 cxb4 but also Nxe6.
If [B]28 ... Na2[/B][B]29 Nxa5[/B] and we're two pawns up, with the same Nxe6 threat.
If [B]28 ... Nc2[/B] [B]29 Rc1[/B] and if[B] 29 ... Nxd4 30 cxd4 Bxd4+[/B] either [B]31 Be3[/B] or [B]31 Ne3[/B] guards our N/c5.
If [B]28 ... Nd3[/B][B] 29 Nxd3[/B]
If [B]28 ... Nbd5[/B][B] 29 Nxe6 Bxe6 30 Rxe6 Nc7 (or Rb1+) 31 Re1 [/B]we're two pawns up.
If [B]28 ... Na6[/B][B] 29 Nd6+ K-move 30 Nxc8[/B] and it's pretty much the same as with our rook on d1, except here we're bearing on e6 with our rook.

[B]28 Rf1[/B] - the analysis is pretty much the same as for [B]28 Rd1[/B][B].
If 28 ... Na6 29 Nxa6 Bxa6[/B] skewers our N/c4 and R/f1, but that's not a real threat: [B]30 Nd6+[/B] saves us:
If 30 ... Kd7 31 Rf7 Bxd4+ 32 cxd4 and if 32 ... Kxd6 33 Bf4+[B]. [/B]If now 33 ... e5 34 Bxe5+ Ke6 35 Rf6+ Kd7 (or Kd5) 36 Bxb8.

[url=http://mersenneforum.org/showpost.php?p=373163&postcount=143]Black's move: [b]27 ... Nxd4[/b][/url]
Timestamp: "10 May 14, 20:52

Maybe they didn't want to have to think about it over Mother's Day (U.S.) tomorrow, and wanted us to have to worry about it instead of honoring our mothers. :-) At least it's the move we've been expecting for a while.

Our deadline (with a one-hour cut to allow for slight oversteppage, if you guys don't mind) is Saturday, May 17, 19:52 GMT = Sunday, May 18, 02:52 Thai time = Saturday, May 17, 15:52 EDT = Saturday, May 17, 14:52 CDT.

We could be mean/boastful, and shoot back [b]28 Re1[/b] immediately, to show that ... Nxd4 was no surprise to us. Or ... maybe we could think a while about whether we've overlooked anything.

The all-legal-moves thing again.

At top are the possibly viable moves, which each have at least one good reason and no immediate refutation. [I](=> They're in order of board position, [U]not merit[/U]![/I] <=) Leading "*" denotes a leading candidate move.

Leading "?" denotes dubious move, listed after non-dubious moves. Moves for which I've seen no apparent good reason (but please point out those for which you see good reason!) are listed, struck-out, below the dubious moves. Concretely-refuted (usually by loss of material) moves are listed, struck-out, at the very bottom.

[B]* 28 Re1[/B] (Gets rook off the d4-fork diagonal, threatens cxd4, deters ... Ne2+, attacks e6 pawn)

-

? [B]28 Nd6+[/B] (Forces ... Kd8 or ... Kf8, but does nothing about the d4-fork, and may be stronger after Re1 or Rf1)

? [B]28 Ne4[/B] (Deters ... Ne2+ but does not threaten cxd4, is _not_ refuted by ... Nb3 because of similarities to 28 Re1 Nb3)

? [B]28 Kf1[/B] (Gets king off the d4-fork diagonal, prevents ... Ne2+, but must look out for ... Ba6 skewer)

? [B]28 Kf2[/B] (Prevents ... Ne2+, but may interfere with rook operation on open f-file)

? [B]28 Kh1[/B] (Gets king off the d4-fork diagonal, but takes it farther from center)

-

[strike]28 Ra2[/strike] refuted by ... Ne2+ or ... Rb1+
[strike]28 Ra3[/strike] refuted by ... Rb1+
[strike]28 Rb1[/strike] refuted by ... Rxb1+
[strike]28 Rc1[/strike] refuted by ... Ne2+
[strike]28 Rd1[/strike] refuted by ... Ne2+
[strike]28 Rf1[/strike] refuted by ... Ne2+
[strike]28 cxd4[/strike] refuted by ... Bxd4+
[strike]28 Na3[/strike] refuted by ... Ne2+
[strike]28 Nxa5[/strike] refuted by ... Ne2+
[strike]28 Nb6[/strike] refuted by ... Rxb6
[strike]28 Ne5[/strike] refuted by ... Bxe5
[strike]28 Ne3[/strike] refuted by ... Ne2+
[strike]28 Nb2[/strike] refuted by ... Ne2+ or ... Rxb2
[strike]28 Na6[/strike] refuted by ... Ne2+ or ... Bxa6
[strike]28 Nb7[/strike] refuted by ... Ne2+ or ... Bxb7
[strike]28 Nd7[/strike] refuted by ... Bxd7 or ... Kxd7
[strike]28 Nxe6[/strike] refuted by ... Nxe6
[strike]28 Nd3[/strike] refuted by ... Ne2+
[strike]28 Nb3[/strike] refuted by ... Nxb3
[strike]28 Bc1[/strike] refuted by ... Ne2+
[strike]28 Be3[/strike] refuted by ... Ne2+
[strike]28 Bf4[/strike] refuted by ... Ne2+
[strike]28 Bg5[/strike] refuted by ... Ne2+ or ... hxg5
[strike]28 Bxh6[/strike] refuted by ... Bxh6
[strike]28 Ba8[/strike] refuted by ... Ne2+ or ... Rxa8
[strike]28 Bb7[/strike] refuted by ... Ne2+ or ... Bxb7
[strike]28 Bc6+[/strike] refuted by ... Ndxc6
[strike]28 Bd5[/strike] refuted by ... exd5
[strike]28 Be4[/strike] refuted by ... Ne2+
[strike]28 Bf1[/strike] refuted by ... Nf3+
[strike]28 Bf3[/strike] refuted by ... Nxf3+
[strike]28 Bh1[/strike] refuted by ... Ne2+
[strike]28 Bh3[/strike] refuted by ... Ne2+
[strike]28 g4[/strike] refuted by ... Ne2+
[strike]28 h3[/strike] refuted by ... Ne2+
[strike]28 h4[/strike] refuted by ... Ne2+

- - -

Well, that doesn't leave much doubt ... unless I've overlooked something.

If you guys don't see anything I missed, maybe we _can_ be mean/boastful, and shoot back [B]28 Re1[/B] immediately.

Vote:

28 Re1 - 5
else - 0

Let's look closer at the two knight-retreat possibilities for Black after [B]28 Re1[/B]:

[QUOTE=cheesehead;373006]< snip >
[B]28 Re1[/B] < snip >

The only two safe retreats are
[B]28 ... Nc6[/B] or [B]28 ... Nf5[/B][/QUOTE]

[B]28 ... Nc6[/B] would return this knight to where it was before the pawn-capture. What might Black threaten from there that it couldn't do before, because now our d4 pawn is gone?

If [B]29 ... e5[/B], now we don't have the d5 (or dxe5) response we had before. This advance would free Black's light-square bishop along the c8-h3 diagonal, unless we immediately play 30 Nd6+ and 31 Nxc8.
[B][B]
-

28 ... Nf5[/B][/B] would position this piece at a new outpost, from which it has different threats than it had from c6. Also, here it frees the N/e7 from guard duty against our former Bxc6+ possibility. But it leaves a5 unguarded.
Chasing the N/f5 with g4 would just provoke ... Nh4[B].

[/B][B]29 Nxa5[/B]. That, in turn, leaves b6 and b2 unguarded, allowing, for instance,
29 ... Rb2 30 Nc4 Ra2 or just
29 ... Rb6 guarding e6 and threatening to transfer to the king-side after ... e5 -- but here 30 Nc4 chases the rook back to b8.

I retract my vote for [b]28 Re1[/b] because a longer look at [b]28 Nd6+[/b] reveals some merits I overlooked before.

- - -

[QUOTE=cheesehead;373196]
[B][B]28 ... Nf5[/B][/B] would < snip >[/QUOTE]Also, it prevents our Nd6+ possibility.

[quote]Chasing the N/f5 with g4 would just provoke ... Nh4[B].[/B][/quote]However, chasing the N/f5 with g4 would prevent that knight from guarding the d6 square, so we could do Nd6+ afterwards.

BTW, after 29 g4 the _only_ safe move for the N/f5 is 29 ... Nh4. We could avoid a knight-for-bishop trade with 30 Be4, but that would obstruct our rook's power along the e-file. Other moves to avoid such a trade are 30 Bf1 or 30 Bh1 (30 Bh3 isn't a good idea because of 30 ... g5).

We could postpone the g4 chase until after capturing the e-pawn (when our rook might wind up on e6.)

-

Those thoughts about [B][B]28 Re1 Nf5[/B][/B] have led me to re-evaluate [b]28 Nd6+[/b]. Pro: getting that in before a potential ... Nf5. Maybe I dismissed it too readily above.

[QUOTE=cheesehead;373188]

? [B]28 Nd6+[/B] (Forces ... Kd8 or ... Kf8, but does nothing about the d4-fork, and may be stronger after Re1 or Rf1)[/QUOTE]

Wow! It gets so complicated (and [U]good[/U]!) after [B]28 Nd6+[/B] that I ran out of edit time for post #5.

[SIZE=4][I][U][B]LaurV, DO NOT POST 28 Re1 as our move yet !![/B][/U][/I][/SIZE]

I will soon post a major revision of the last part of post #5. [B][U]Don't post a move for us until I've completed it![/U][/B]

Hm... I don't like Nd6.
- it moves the king out of link for free (the strength of Re1 move is the fact that is linked to the black king, too. If we move Re1 then black must lose a king move later).
- it loses control of b2 and b6 squares (if 28 Re1 Nec6 then 29 Bf4 and black loses a figure, and the game therefore black has to retreat the horse)
- the horse in d6 blocks our Bf4 move (see the point above).
- it does not get rid of the fork, so we still need to play the rook. In this light, the next point:
- after Nd6 Kf8 we have no play. Except of course, Re1, which is now much weaker, as b2, b6 are not guarded, etc.
[edit: so Nd6 Kf8 Re1 Rb2, then what? Nc4 back? Ra2 or?]

[QUOTE=LaurV;373244] In this light, the next point:
- after Nd6 Kf8 we have no play. Except of course, Re1, which is now much weaker, as b2, b6 are not guarded, etc.
[edit: so Nd6 Kf8 Re1 Rb2, then what? Nc4 back? Ra2 or?][/QUOTE]
Okay, once I completed looking at [b]28 Nd6+ Kf8[/b] I see that we can't do what I thought of yesterday.

[quote]if 28 Re1 Nec6[/quote]simply 29 cxd4 (since we can meet 29 ... Bxd4+ with 30 Be3)
or
[quote]then 29 Bf4 and black loses a figure,[/quote]Let's show that explicitly, move by move.
[b]28 Re1 Nec6 29 Bf4[/b] threatens 30 Bxb8 and the rook's only "safe" move is 29 ... Ra8, but that still leaves the rook on our B/g2's long diagonal.
If [b]29 ... Ra8 30 Nb6 Ra7 31 Nxc8[/b]
If [b]31 ... Ra8 32 Nd6+[/b]
If [b]32 ... Kd8 or Kf8 33 cxd4 Bxd4+ 34 Be3[/b]
If [b]32 ... Ke7 33 cxd4 Bxd4+ 34 Be3 Bxe3+ 35 Rxe3 Kxd6 36 Rxe6+ Kxc5 37 Rxc6+ Kb4 38 Rb6+[/b] (or [b]37 ... Kd4 38 Rd6+[/b]). Now, no matter where the king moves, 39 Bxa8 leaves us at least a bishop up in a winning endgame.
If [b]31 ... Rf7 32 Nd6+ and 33 Nxf7[/b]
If [b]29 ... e5[/b] the e-pawn is pinned, and thus can't immediately capture with 30 ... exf4, but it does block the B/f4's attack on the R/b8. Then
[b]30 cxd4[/b] puts us up a piece and threatens 31 Bxc6+, but our b4 square is no longer guarded against Rb4.
If [b]30 ... Nxd4 31 Nxe5[/b] threatens 32 Nc6+ and 33 Nxc8.
If [b]31 ... Kf8 (not ... Kd8 32 Nf7++ or Nc6++) 32 Ned7+ Bxd7 33 Nxd7+ K-moves 33 Bxc8 or Nxc8[/b].

Okay, I'm convinced. Back to:

Vote

28 Re1 -- 5
else -- 0

Hmm ... I thought there was something else, but it didn't pan out.

I don't see anthing better than Re1. I also looked at Nd6+ and found it lacking in forcing a favorable outcome.
Re1 5
anthing else 0

Oops again. Another busted analysis where I thought we would win a piece, but ...

(Nothing wrong with 28 Re1, though)