rook ogle in corridor knight, doesnt it I annex us together e8 paint

a poor picture no b5 lack swanz damage ne4 stymmy in wall, flower it 15...bd7 in permanent thorn to pick on a cycle in b6 or came in

borrow a lean build (inclusive so) in f6xd5 position equality oh,

ace lip in bd7 white gets i tantric in g3 e4 knights and roocks

tender feed dig it is take in e4 get pieces off, finally 20...re8

dont upset apple cart mario before the horse leap in d6, counting up

ply it stare in would face be a 21.Nxd6, in for key you see

nincompoop in c8 heck agile e3 has b6 rote ammend knight was needed

to guard in b6 22.kc8 and tactics abound in 5b bull shy it knock off

in c6 look for now light has b5 engage d6 everything kaputz elle

darius king children bereft, black elects e3 on demand, wrestlings

ponos game this re-joinder method allows knightd6 in b7+ to gate

again kingc7 rooke1xe3 on goKxb7 wangle knightc3 one dip in sate

24.Ne4 rook heading to b3+ pawn it jiggle d6 down e4 bind c5 open in

kinga7 eg light has no reiterance in packing 27.Nc5+ in general aims

it signal in double brevity ease cinch a benign king allowed fin d7?

click for larger view

Like almosty everyone else, I went for 21.Nxd6 which seemed safe (R+PP for two pieces in the game line). However I am struggling against Crafty End Game Trainer. Without firing up my silicon friend Herr Fritz, I can't find a clear win. The EGT excels in these messy, tactical positions: it even gives me a free double check and I still can't beat it (unaided). A reminder of the puzzle position:

click for larger view

with EGT interactive link http://www.chessvideos.tv/endgame-t...

I looked in detail at the line 24 Ne4 Nf6 25 Rb3+.

click for larger view

Black has only one move here, 25…Bb5, as 25…Kb7 loses to 26 Rc7# and 25… Nb6 loses to 26 Nd6+ with forced mate.

So in order to prevent mate black has to pin his bishop.

Now white wins after 26 Nxf6 Bxf6 27 a4.

click for larger view

<chessgames.com: <The Rinus Scholarship> For several years, User: chesstoplay (Peter) devised and managed "The Rinus Scholarship", an effort designed to give premium memberships to people who contribute to the World Team's success. The Rinus Scholarship is named in memory of the Chessgames member User: rinus who passed away in August of 2008. Rinus was the first recipient of this award, and remains a shining inspiration to the World Team.

Sadly, due to health problems Peter can no longer continue this effort. Fortunately, some members stepped forward to take over the job.

To help ensure its success, we have now incorporated the Rinus Scholarship directly into our website, allowing anybody to donate a premium account directly to the fund. This fund will then release premium accounts to members of the World Team, at the judgment of the new managers.

From time to time you'll be seeing a small banner ad to the left of the board soliciting donations. The proceeds from donations made to this fund will not only purchase a premium membership for a deserving user who may not otherwise be able to afford one, but also will go into a special account used to finance future Chessgames Challenges. In this way, we hope to hire the very finest grandmasters in the world to step forward and take the challenge. So each dollar donated serves two purposes: you put a smile on somebody's face today, and in the future you put a smile on thousands of people's faces when a new Chessgames Challenge begins.

If you are able, please consider donating to worldwide chess learning through the Rinus Scholarship. Either click on the banner ad when it comes around, or better yet, click on this link right now to make a donation:

21...Nxd6 22.Bb6+ Kc8 23.Nb5+ Kb8 (23...Bc6 24.Nxd6+ and 25.dxc6 ) 24.Bc7+ Kc8 25.Nxd6#

21...Rxe3 22.Nxb7+ Kc7 23.Rxe3 Kxb7

21...Bxc3 22.Rxc3 Nxd6? 23.Bb6#

In my solution post, considering line (3) with 21. Nxd6, Rxe3; 22. Nxb7, Kc7, I extolled as a "key move" the apparently tempo-gaining 23. Ne4† (to which I gave an exclamation point), reasoning that the knight would invade on c5 to bring home the attack.

I have now revisited this line in light of <LTJ>'s "tale of the tape," and to my chagrin I find that my touted finesse appears insufficient for a win. After 23. Ne4?, Rc3!; 24. Nxc3, Bxc3!!, White must recapture on c3 with the rook or permit the closing of the c-file he needs to continue the attack. After he does this, however, he finds that his rooks now trip over one another, as they do not if he plays 23. Rxe3 as in the game.

The critical difference is that with the rook held on c1, the black king can't flee toward the center, and meanwhile the recapture on e3 doubles as a rook lift, enabling that rook to enter the fray via b3 with check. After 23. Ne4, however, Black's zwischenzug removes the rook lift, leaving White in a position in which he must first spend a tempo shifting 25. Rec1 before he can use the rook on c3 to check on the b-file.

What a difference that tempo makes. As far as I can determine, the whole outcome of the game turns on it. For after 25. Rec1, Nb6, there doesn't appear to be a way for White to continue the attack in any useful sense: On 26. Rc7†, Kb8; 27. d6, Nf6; 28. Be2, Nfd5, White has significant pressure, but his capacity to convert it into a win appears problematic.

I wouldn't call this agreeing. I was quoting a depth of 25 (at 27, 0.9 Gnodes, I'm getting +1.4 with a mainline of the enginesque ...Kc7 over the more human ...Nd6), and I'd call a pawn and a half won.