White 29

[WMHalsdorf]

If 29 ... e5 30 Nxe5 Nxc5 31 Nxg6 if Rook moves to a square where it can't be captured then 32 Rxe7+ Kd8 33 Bc7 mate.

[cheesehead]

Here's my current analysis tree, including the [B]29 Bf4 e5 30 Nxe5 Nxc5 31 Nxg6[/B] branch:

[QUOTE=cheesehead;374118]< snip >

[B]29 Bf4[/B] (threatens 30 Bxb8 and the rook has _no_ "safe" move).

< snip >

If [B]29 ... e5[/B][/QUOTE]

If [B]30 Bxe5[/B] ????

If [B]30 Nxe5[/B]

If [B]30 ... Nxc5 31 Nxg6[/B]

Here, let's do an all-legal-moves table for Black:

If [strike]31 ... Ra8[/strike] 32 Bxa8
If [B]31 ... Rb1 32 Rxb1 Nxg6[/B] [speculation] 33 Rb8 Ne7 34 Bh3 Kd8 35 Bd6 Nd7 36 Bxe7+ Kxe7 37 Bxd7 Kxd7 38 c4[/b] we have a rook and two pawns for two bishops
If [strike]31 ... Rb2[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Rb3[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Rb4[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Rb5[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Rb6[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Rb7[/strike] 32 Bxb7 Bxb7 33 Rxe7+ Kd8 34 Rxg7 we're a rook and two pawns up.
If [strike]31 ... Nxa4[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Nb3[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Nd3[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Ne4[/strike] 32 Bxb8
If [strike]31 ... Ne6[/strike] 32 Bxb8
If [strike]31 ... Nd7[/strike] 32 Bxb8 Nxb8 33 Rxe7+ Kd8 34 Rxg7
If [strike]31 ... Nb7[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Na6[/strike] 32 Bxb8 Nxb8 33 Rxe7+ Kd8 34 Rxg7
If [strike]31 ... Ba6[/strike] 32 Rxe7+ Kd8 33 Bxb8
If [strike]31 ... Bb7[/strike] 32 Rxe7+ Kd8 33 Bxb8
If [strike]31 ... Bd7[/strike] 32 Rxe7+ Kd8 33 Bxb8
If [strike]31 ... Be6[/strike] 32 Bxb8
If [strike]31 ... Bf5[/strike] 32 Rxe7+ Kd8 33 Bxb8 Bxg6 34 Rxg7
If [strike]31 ... Bg4[/strike] 32 Rxe7+ Kd8 33 Bxb8
If [strike]31 ... Bh3[/strike] 32 Rxe7+ Kd8 33 Bxb8
If [strike]31 ... Kd7[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Kd8[/strike] 32 Bxb8
If [strike]31 ... Kf7[/strike] 32 Rxe7+
If 32 ... Kf6 33 Bxb8
If 32 ... Kxg6 33 Bxb8
If 32 ... Kg8 33 Bxb8
If [strike]31 ... Bxc3[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... Bd4+[/strike] 32 cxd4
If 31 ... Be5
If 32 Rxe5 Kf7
If 33 Nxe7 ????
If 33 Rxc5
If 33 ... B-moves 34 Bxb8
If 33 ????
If 32 Bxe5 Rb6 33 ????
If [strike]31 ... Bf6[/strike] 32 Bxb8
If [strike]31 ... Bf8[/strike] 32 Bxb8
If [strike]31 ... Bh8[/strike] 32 Rxe7+ Kd8 33 Bc7++
If [strike]31 ... h5[/strike] 32 Rxe7+ Kd8 33 Bc7++


If [strike]30 Rxe5[/strike] Bxe5 31 Bxe5 Nxc5 32 Bxb8 we're only a pawn up.

If [B]30 Be3[/B] ????

As you see, most Black moves after [B]29 Bf4 e5 30 Nxe5 Nxc5 31 Nxg6[/B] are indeed refuted, because they lead to mate or win of a rook ... except ...

the not-yet-resolved [B]31 ... Be5[/B]

If [B]31 ... Be5[/B]
If 32 Rxe5 Kf7
If 33 Nxe7 ????
If 33 Rxc5
If 33 ... B-moves 34 Bxb8
If 33 ????
If 32 Bxe5 Rb6 33 ????

and

If [B]31 ... Rb1 32 Rxb1 Nxg6[/B] [speculation] 33 Rb8 Ne7 34 Bh3 Kd8 35 Bd6 Nd7 36 Bxe7+ Kxe7 37 Bxd7 Kxd7 38 c4[/b] we have a rook and two pawns for two bishops

- -

Can anyone find a win for us after [B]29 Bf4 e5 30 Nxe5 Nxc5 31 Nxg6[/B][B] Rb1[/B] or at least an outcome better than (rook and two pawns for two bishops) ?

How about finishing the analysis after [B][B]29 Bf4 e5 30 Nxe5 Nxc5 31 Nxg6[/B] Be5[/B] ? The Be5 sacrifice (whose value is that it makes our R/e1 and B/f4 attacks interfere with each other) is not yet obviously unsound as far as I've looked. I have to take a break now, or I'd finish it myself.

- -

[B]29 Rb1[/B] wins us at least a piece or Exchange-plus-a-pawn in all branches I've looked at, so [B][B]29 Bf4[/B][/B] will have to do better than that to convince me that [B]29 Rb1[/B] isn't best.

As of now, I vote for [B]29 Rb1[/B].

IMO the idea that [B]29 Rb1[/B] isn't as aggressive or direct as [B][B]29 Bf4[/B][/B] is absurd, once one sees how easily Black complicates the latter with [B]29 ... e5[/B].

[LaurV]

[QUOTE=cheesehead;374162]As of now, I vote for [B]29 Rb1[/B].
[/QUOTE]
I went through all your analysis today. Very good! I don't believe we can move anything better than 29.Rb1. No matter how they play this, we end up with at least a figure. If they beat the bishop we can escape the horse with a check Nd6, and take the rook, then if Kc7, Nxc8, Kxb8, Nxe7, and we have horse and make free pawn, this is white win. If they play 29...anything else except Nx(d2|c5), then we still may be able to play Bf4 and stay much better.

The only part which is was uncovered was 29 Rb1 Nxc5 30 Rxb8 Kd8 31 Nb6 Kc7 Nxd8 brings us to the same position, and black is somehow forced to play it, any other path bring loss of additional figure or pawn.

So, we play final without rooks, and with a figure (and maybe a free pawn if we clear a5) more. There is no way we could do better here. Or lose the game.

My vote is Rb1 too.

[WMHalsdorf]

While both 29. Rb1 and 29. Bf4 lead to a material advantage the position arising out of 29. Rb1 seems clearer to me. So unless someone comes up with something better than 29. Rb1 I'll go with it.

[cheesehead]

[QUOTE=LaurV;374171]The only part which is was uncovered was 29 Rb1 Nxc5 30 Rxb8 Kd8 31 Nb6 Kc7 [strike]Nxd8[/strike]Nxc8 brings us to the same position, and black is somehow forced to play it,[/QUOTE]That is the result of our various positional advantages, building up on each other. Again, I thank you guys for persuading me to agree to 15 e5 and 18 a4, both of which strengthened our position, but neither of which I normally would have played on my own. Both were examples of properly-supported pawn pushes that probed/provoked weaknesses in Black's position.

[quote]any other path bring loss of additional figure or pawn.[/quote]... because we control important squares on Black's side of the board.

[quote]So, we play final without rooks,[/quote](* sigh *) I'd love to have an excuse to dig out my [I]Rook and Pawn Endings[/I] (British author, early 1970s) and [I]Rook Endings[/I] (Levenfish and Smyslov, English translation 1973) books. Just after I read the first one and learned the key positions and rules, I got into a rook-and-pawn ending versus one of the stronger senior players in a local club tournament.

At one point he offered a draw, but I could see from my recent book-learning that I had a win (he said after the game that he had not yet realized that I had a win at that point, or he wouldn't have suggested a draw). I felt awkward refusing to draw with a player rated 350 higher than me, but I was as polite as possible, and continued on to win.

I notice that there's now a [URL="http://Wikibook on rook and pawn endings"]Wikibook on rook and pawn endings[/URL], but haven't looked at it.

[cheesehead]

[QUOTE=cheesehead;374194] [I]Rook and Pawn Endings[/I] (British author, early 1970s)[/quote]
[strike]Perhaps[/strike] Probably I'm misremembering that one. At least two lists of important endgame books don't mention it, an online search doesn't find it (though it easily finds the Levenfish and Smyslov book), and I can't find it in a quick search of my chess books.

But the Levenfish and Smyslov book doesn't look like the one from which I learned the Lucena position!

[quote]I notice that there's now a [URL="http://Wikibook on rook and pawn endings"]Wikibook on rook and pawn endings[/URL], but haven't looked at it.[/QUOTE](* ahem *)

I meant

... a [URL="http://en.wikibooks.org/wiki/Chess/The_Endgame/Rook_and_Pawn_Endings"]Wikibook on rook and pawn endings[/URL] ...

but it's just nine paragraphs with one diagram!

(* snort *)

[cheesehead]

[QUOTE=LaurV;374171]My vote is Rb1 too.[/QUOTE]

[QUOTE=WMHalsdorf;374179]So unless someone comes up with something better than 29. Rb1 I'll go with it.[/QUOTE]

Great! Now that we've agreed about [B]29 Rb1[/B] ... it's time to extend the vague parts of [B]29 Bf4[/B] analysis (and of [B]29 Rb1[/B] analysis) before committing ourselves.

- - -

[QUOTE=cheesehead;374162]< snip >

If [B]30 Bxe5[/B] ????[/QUOTE]

[B]29 Bf4[/B] [B]e5[/B] [B]30 Bxe5[/B]

[b]30 ... Bxe5[/b] seems to be the only reasonable Black response here.

If [b]31 Nxe5[/b]
If [b]31 ... Nxc5[/b] ????

If [b]31 Rxe5[/b]
If [b]31 ... Nxc5[/b] ????

If [b]31 Nxb3[/b]
If [b]31 ... Rxb3[/b] ????


(to be continued)

[cheesehead]

[QUOTE=cheesehead;374229]

[B]29 Bf4[/B] [B]e5[/B] [B]30 Bxe5[/B]

[B]30 ... Bxe5[/B] seems to be the only reasonable Black response here.

If [B]31 Nxe5[/B][/QUOTE]
If [B]31 ... Nxc5[/B], we're down a knight for a pawn here, so we'd better threaten to capture something.

If [B]32 Nc6[/B] (forks the R/c8 and pinned N/e7)
If [B]32 ... Rb7[/B]
If [B]33 Rxe7+ Rxe7 34 Nxe7 Kxe7 35 Bc6[/B] leaves us a knight-for-pawn down, headed for loss, as does [B]33 Nxe7 Rxe7 34 Rxe7+ Kxe7 35 Bc6[/B].
If [B]33 Nxa5 Rc7 34 Bc6+ Bd7 35 Bb5 Kd8[/B] shows that we don't have enough firepower left to menace the Black king. Our queen-side pawns will fall (e.g., ... Ra7).
I don't see how we can avoid loss after [B]32 Nc6[/B].

If [B]32 Bc6+ Bd7 33 Nxd7 Nxd7[/B] we're still down a piece and our c-pawn will fall soon after ... Kd8.

If [B]32 Nxg6 Be6[/B] un-pins the N/e7 long enough for Black to reorganize.
E.g., 33 Bh3 Rb6 34 Nf4 Kf7 and ... we're still down a piece.

I don't see any other promising line for regaining the bishop we sacrificed at move 30.

- - -

From post #13 above, [B]30 Nxe5[/B] [B]Nxc5 31 Nxg6[/B] (where we've sacrificed a knight for two pawns)

allows Black to hold with [B]31 ... Rb1 32 Rxb1 Nxg6[/B] because we've given both our knights for a rook and two pawns ... actually, rook and one pawn because we can't hold onto both our a-pawn and c-pawn.

[cheesehead]

Another way to respond to [B]29 Bf4 e5[/B] is simply to re-position our dark-square bishop. In view of our N/c5's predicament, [B]30 Be3[/B] may be suitable ... except that then we'd have blocked our R/e1's file and have no remaining threats to gain material .... which takes all the fun out of why we played [B]29 Bf4[/B] in the first place.

We also still have [B]30 Nd6+[/B]. After [B]30 ... Kd8[/B] we could do [B]31 Ncb7+[/B] to get our N/c5 into play, or [B]31 Nf7+[/B] to see whether capturing on e5 from a different direction after pushing the Black king around might be better than 30 Nxe5 was.

If [B]30 Nd6+ Kd8 31 Ncb7+ Bxb7 32 Nxb7+[/B]
[B]32 ... Kc7[/B] threatens 33 ... Rxb7 34 Bxb7 Kxb7.
[speculation] 33 Bxe5+ Bxe5 34 Rxe5 Nc6 35 Rb5 Rxb7 and we're still down a piece-for-a-pawn.

- - -

I doubt that further analysis would reveal a winning line for us after [B]29 Bf4 e5[/B]. The [B] ... e5[/B]'s combination of blocking our B/f4's attack on the rook, attacking our B/f4, and freeing Black's B/c8 is too much help for Black. Our positional advantages have left Black rather constricted, and [B]29 Rb1[/B] takes advantage of that without sacrificing material and without letting Black uncramp.

(All this is not to say that we should post [B]29 Rb1[/B] immediately. Let our collective subconsciousnesses ponder the possibilities a day or two.)

[cheesehead]

Continuing to tie loose threads of the [B]29 Bf4[/B] analysis:

[QUOTE=cheesehead;374229]

[B]29 Bf4[/B] [B]e5[/B] [B]30 Bxe5[/B]

[B]30 ... Bxe5[/B] seems to be the only reasonable Black response here.

< snip >

If [B]31 Rxe5[/B][/QUOTE]Unlike 31 Nxe6, [B]31 Rxe5[/B] has the advantage of protecting our N/c5 ... [I]but that protection is illusory:[/I]

If [B]31 ... Nxc5[/B], then
32 Rxc5 preserves our one-pawn material advantage (and attacks their a5-pawn a second time), but releases the pin on the N/e7 ... and, most damagingly, [I]leaves our king to be attacked by Black's newly-freed rook and bishop, starting with[/I]
32 ... Rb1+.

Here, 33 Bf1 allows 33 ... Bh3 (threatening mate) 34 Nd2 Rd1 (threatening ... N/e7-f5-e3, while Black's king can eventually escape checks by our rook by journeying via g5 to h4).
If 35 Kf2 Rxd2+ 36 Ke1 Rxh2 and we're down a knight.
If 35 Rc4 (intending Rh4) g5 36 Rd4 (protecting the N/d2 so that Kf2 is not met by ... Rxd2+) Nf5 37 Re4+ Kf7 38 Re2 Nd6 (threatening 39 ... Nc4, where 40 Nxc4 would allow 40 ... Rxf1++) 39 Rf2+ Kg6 and now we need to uncramp [I]our[/I] position in order to exploit our one-pawn advantage. Any move by our N/d2 allows ... Ne4 or worse. E.g. 40 Nf3 Ne4 41 R-move Rxf1++.
Or 40 Nb3 Ne4 41 Rf3 Nxc3 42 Nxa5 or Nc5 (not Rxc3 Rxf1++) Ne2+ 43 Kf2 Bxf1 and we're a bishop down, or
42 Nd4 Bxf1 43 Rxf1 Rxd4 and we're a knight down.

Our only other response to 32 ... Rb1+, 33 Kf2 allows 33 ... Rb2+ and if 34 Kf3 Bb7+ we lose our remaining bishop and the game.

So, after [B]29 Bf4[/B] [B]e5[/B] [B]30 Bxe5[/B] [B]Bxe5, [/B][B]31 Rxe5[/B] is a losing line for us.

Another possible response to [B]29 Bf4[/B] [B]e5[/B] [B]30 Bxe5[/B] [B]Bxe5[/B] is
[QUOTE=cheesehead;374229]
< snip >

If [B]31 Nxb3[/B][/QUOTE]Suppose Black then captures with [B]31 ... Rxb3[/B].

Now Black's rook and light-square bishop are still free to attack our king's position after [B]32 Rxe5[/B]:
32 ... Rb1+ 33 Bf1 Bh3 34 Nd2 Rd1 is familiar by now, right?
35 Re2 Kf7 36 Rf2+ Kg7 37 Nc4 (intending Nc4-e5-d3-f4) g5 38 Nxa5 Nd5 39 Nc4 ( not 39 c4 Ne3 threatening 40 ... Nxf1) Nxc3 (threatening 41 ... Ne2+ 42 Kh1 [not 42 Rxe2 Rxf1++] Rxf1+ 43 Rxf1 Bxf1 and we're a bishop-for-a-pawn down with their bishop being the right color to prevent our a-pawn from queening) 40 Ne3 Re1 41 Ng2 Ra1 42 Rc2 Nxa4 43 Kf2 even material and drawish.

So, [B]29 Bf4[/B] [B]e5[/B] [B]30 Bxe5[/B] [B]Bxe5 [/B][B]31 Nxb3[/B] joins the list of non-winning lines for us.

Does anyone see any better move for us after [B]29 Bf4[/B] [B]e5[/B] [B]30 Bxe5[/B] [B]Bxe5[/B]?

[cheesehead]

[QUOTE=cheesehead;374162]< snip >

[B]29 Rb1[/B] wins us at least a piece or Exchange-plus-a-pawn in all branches I've looked at[/QUOTE]Guys,

I now see that I hadn't looked at as many of the potential branches as I thought I had. There are several other branches to examine, not only for Black's 29th, but also for ours!

Consideration: 28 ... Nb3 attacks both our N/c5 and B/d2. Two alternatives are for one of those to protect the other: 29 Be3 or 29 Ne4. Another is to capture: 29 Nxb3.

Here's the all-legal-moves list for our move 29:

[B]29 Ne3[/B]
[B]29 Ne5[/B]
[B]29 Nd6+[/B]
[strike]29 Nb6[/strike] Rxb6
[strike]29 Nxa5[/strike] Nxa5
[B]29 Na3[/B]
[B]29 Nb2[/B]
[B]29 Nd3[/B]
[B]29 Ne4[/B]
[B]29 Nxe6[/B]
[strike]29 Nd7[/strike] Bxd7 or Kxd7
[strike]29 Nb7[/strike] Bxb7 30 Bxb7 Rxb7 31 Nd6+ Kd7 32 Nxb7 Nxd2
[strike]29 Na6[/strike] Bxa6
? 29 Nxb3 Rxb3 30 Nxa5 ????
[strike]29 Bc1[/strike] Bxc3
[strike]29 Be3[/strike] Bxc3
[strike]29 Bf4[/strike] e5
[strike]29 Bg5[/strike] hxg5
[strike]29 Bxh6[/strike] Bxh6
[strike]29 Ra1[/strike] Nxa1
[B]29 Rb1[/B]
[strike]29 Rc1[/strike] Nxc1
[strike]29 Rd1[/strike] Nxc5
[strike]29 Re2[/strike] Nxc5
[strike]29 Re3[/strike] Nxc5
[strike]29 Re4[/strike] Nxc5
[strike]29 Re5[/strike] Bxe5
[strike]29 Rxe6[/strike] Bxe6
[strike]29 Rf1[/strike] Nxc5
[strike]29 Kh1[/strike] Nxc5
[strike]29 Kf2[/strike] Nxc5
[strike]29 Kf1[/strike] Nxc5
[strike]29 Bh1[/strike] Nxc5
[strike]29 Bh3[/strike] Nxc5
[strike]29 Ba8[/strike] Nxc5
[strike]29 Bb7[/strike] Nxc5
[B]29 Bc6+[/B]
[strike]29 Bd5[/strike] exd5
[strike]29 Be4[/strike] Nxc5
[strike]29 Bf3[/strike] Nxc5
[strike]29 Bf1[/strike] Nxc5
[strike]29 g4[/strike] Nxc5
[strike]29 h3[/strike] Nxc5
[strike]29 h4[/strike] Nxc5

- - -

The non-struck ones are:

[B]29 Ne3[/B]
[B]29 Ne5[/B]
[B]29 Nd6+[/B]
[B]29 Na3[/B]
[B]29 Nb2[/B]
[B]29 Nd3[/B]
[B]29 Ne4[/B]
[B]29 Nxe6[/B]
[B]29 Rb1[/B]
[B]29 Bc6+[/B]

and dubious:

? 29 Nxb3 Rxb3 30 Nax5 ????

I think we ought to do some analysis for all these non-struck moves, not just [B]29 Rb1[/B]

- - -

Note:
To our previous [B]29 Rb1[/B] analysis, add:

[B]29 Rb1[/B]

If [B]29 ... Nxc5[/B] [B]30 Rxb8[/B]
If [B]30 ... Nxa4[/B] (threatening 31 ... Nxc3) ????

- - -

(to be continued, with all-legal-moves list for Black after [B]29 Rb1[/B])