All more recent "immortals" must disambiguate somehow. Usually they add the winner's name, e.g. <Kasparov's Immortal> Kasparov vs Topalov, 1999. Sometimes they add the winner's nationality instead, e.g. <Polish Immortal> Glucksberg vs Najdorf, 1929.

1. 12/24/2002, Pogo, 5+3, GuaranteedGem-GilmoyFirewyrm, position after 17.Bd2, Black to win:

click for larger view

I thought my continuation was perhaps overly theatrical.

2. 08/09/2007, Pogo, 5+3, moodychess-GilmoySeesDedPpl, position after 22.Be3, Black to win:

click for larger view

Piffle or wow? I saw it in a flash, though it was just silly, almost tantamout to saying "I think you are a patzer, you will blindly save your rook". (That he did is beside the point :) But in retrospect, I see nothing else for White!

<At 12.e5 I was already envisioning the sac. This pawn thrust has two points: (a) boots out the stalwart Nf6; (b) clears e4 for a K-side swarm, e.g. Ne4, Q sideways, and possibly the b1-h7 diagonal. Black usually must struggle to prevent this!>

13..Nc7 <threatens Nb5>

click for larger view

<I saw that the immediate 14.Ne4 momentarily impedes my Q, after which 14..Nb5 drives her off, winning the e-pawn. So I wanted to move my Q first, without losing the pawn. Including time for the sac itself, I spent almost a minute on:>

14.Qf4 b5 15.Ne4 Bb7 16.Nf6+ Kh8!?

click for larger view

<I calculated that gxf6 17.exf6 isn't winning -- Black has time for Kh8-Rg8. I vaguely planned c3-Bc2 and a rook lift for pressure against h7. But having spent almost half my game clock preparing the sac, I was determined to have some fun!>

17.Rd3 <Now my N is safe (proof thereof is roughly Tuesday level)> Nd5 18.Bxd5 Bxd5 19.b3 Rac8 20.Rh3 g6 <20..h6 21.Rxh6+ and Q mates> 21.Rxh7# 1-0

[Site "pogo.com"]
[Date "2002.12.24"]
[White "GuaranteedGem" 1632]
[Black "GilmoyFirewyrm" 1756]
[Result "0-1"]

1.e4 e5 2.Nf3 Nc6 3.d4 <I hate the Scotch -- so unimaginative> exd4 4.Nxd4 Bc5 5.Nxc6 bxc6 6.Bc4 Qh4!? <Sometimes I do this on move 4 just to see 'em squirm> 7.Qf3 Nf6 8.Nc3 O-O 9.O-O d6 10.Qg3 Qxg3 11.hxg3 <opens h> Ng4 12.a3 Ne5 <a pretty mini-tour idea in some defenses> 13.Be2 a5 14.Kh1 <Aha! he unpins his f-pawn -- thin ice ahead --> f5 15.f4 <he gives himself King-in-the-corner syndrome -- can be fatal!> Ng4 16.Bxg4 fxg4 17.Bd2

click for larger view

<Like most other chess enthusiasts, I've gawked at Rubinstein's Immortal. One nugget I mined from it was a deep appreciation for King-in-the-corner. The canonical form has a B covering g1, as here, which implies that White's f-pawn moved. It can arise from the outright check Bc5+ Kh1. Rarely, White plays Kh1 specifically to unpin his f-pawn -- dangerous because he's inviting this mate pattern.

Here, I realized with compounding amazement that White has one fatal weakness -- he can't double anywhere along the a7-g1 diagonal in one move. That means he can't stop my B -- only spite-block to delay mate. But how to get a line-moving R to h?>

18..Bf5! <mistreaver [reply 16] is correct: the point is to blockade f4, so that White's B cannot control h6.>

19.exf5 <now f4 is blocked by his own pawn> Rf6 0-1

click for larger view

[Site "Pogo"]
[Date "2007.08.09"]
[White "moodychess" 1699]
[Black "GilmoySeesDedPpl" 1890]
[Result "0-1"]

1.d4 d5 <Before my KID jag> 2.c4 e6 3.Nc3 Nf6 4.Nf3 Be7 5.e3 a6 6.Bd3 <yaay! the clockwork line> dxc4 7.Bxc4 b5 8.Bd3 Bb7 9.a3 Nd7 10.O-O c5 11.Bc2 Rc8 12.e4? <Harrumph -- now I re-engage my brain> cxd4 13.Nxd4 O-O 14.Be3 Ne5 15.f4 <the f-pawn departs> Nc4 16.Bc1 Bc5 <the B arrives> 17.Nce2 Nxe4 18.b4 Bb6 <homage to Rubinstein> 19.Kh1 <syndrome!!> Qf6 20.Bb3 Rfd8 21.Bxc4 Rxc4 22.Be3

click for larger view

<OK, he's down a pawn, and his hapless Nd4 is pinned and outnumbered 4-3, so it's toast. The only question is which piece I want to end on d4. Rote application of minimal cost suggests ending with Rxd4, which wins NN(+B) for R(+B) and boots his Q out -- but after that, it's blah.

Much more exciting was to hammer the King-in-the-corner syndrome, and end with Bxd4 -- a subtle fork, Ra1 and g1. What I saw "in a flash" was the Greco sac -- which is matish if I own g1!>

22..Rcxd4 <Ob. joke: "Oh, but"> 23.Nxd4 Rxd4! <"I insist"> 24.Bxd4 Bxd4 <"on controlling g1">

click for larger view

<So I got the NN for R portion, but I sacked an exchange with my other R. During the game, I thought my whole combo was "silly" because it relied on White reflexively saving his Ra1 -- perhaps he could have refuted it by elegantly returning an exchange, and doing something on the K-side? But on hindsight, I think he doesn't have any such move! My Bs are just too, ah, "rubenstian".>

25.Rc1 Ng3+ <h opens up> 26.hxg3 Qh6+ 0-1

click for larger view

[Site "Pogo"]
[Date "2003.02.10"]
[White "GilmoyDragsAWing"] (my 27th Gilmoy)
[Black "gggessas"]
[Result "1-0"]

1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nf3 Nc6? <Oops -- now he's down a gambit pawn!> 5.d3 Be7 6.Be2 O-O 7.O-O Re8 8.Nc3 d5 9.e5 Ng4 <Inevitably this will shatter his K-side pawns> 10.d4 f6 11.h3 Nh6 12.Bxh6 gxh6 13.exf6 Bxf6 14.Qd2 Nb4? <A free tempo> 15.a3 Nc6 16.Rad1 Bg5 17.Nxg5 hxg5 18.b3 <Hidden attack on d5> g4 19.Nxd5 gxh3 <Here I calculated that half-open h is more useful to me than g, and my K could hide from Black's Rg8+> 20.Bc4 hxg2? 21.Kxg2 <Elegantly clears 1 -- Black should have declined the pawn trade just to prevent this> Kh8 22.Rh1! Rg8+ 23.Kf1 Bg4 <Now I see mate, so material doesn't matter -- a tempo is worth more than a rook!> 24.Qh6 Rg7

click for larger view

25.Nf6 Bxd1? 26.Qxh7+ Rxh7 27.Rxh7# 1-0

click for larger view

<<Position after 21.Re5 1-0>>

By sheer coincidence, just 3 days earlier, I had an isomorphic configuration of pieces, and I found the same combo OTB (before I ever saw the Gurgenidze game). That's pretty encouraging!

[Site "Pogo"]
[Date "2008.05.24"]
[White "GilmoyLoses512MB" 1894] (my 88th Gilmoy -- still active :) [Black "RogDuaneChess" 1686]
[Result "1-0"]

1.e4 b6 <bleah> 2.d4 Bb7 3.Bd3 g6 4.Nf3 Bg7 <Looks like a Dragon, so I steer toward Yugoslav> 5.Be3 a6 6.Nc3 h6 7.Qd2 e6 8.O-O-O Ne7 <I start freelancing> 9.g4 b5 10.Rg1 b4 11.Na4 Bc6 12.Qxb4 <It's risky to give him half-open b, but pawn is pawn> Bxa4 13.Qxa4 Nc6 14.c3 <To transfer her to K-side> d5 15.Ne5 Bxe5 16.dxe5 O-O 17.f4 d4 18.cxd4 Rb8 19.f5 exf5 20.Bxh6 Rb4 21.Qc2 Re8 22.gxf5 Nxd4 23.Qg2 Qd7 24.Bd2 Qc6+ 25.Bc3 Rbb8 <Now it's a race -- his Q-side pressure is quite serious> 26.f6 Nc8 <I see the double-deflection, winning by a nose> 27.Qg5 Ne6 28.Qh6 Nb6 29.Rg5! <Nxg5?? 30.Qg7#> Na4

click for larger view

<<Position after 29.Rg5 Na4 -- compare to above>>

30.Rh5! <threatens 31.Qh8#> gxh5[] 31.Rg1+ Ng7[] 32.Qxg7# 1-0

You're right, both lines win -- in fact, 29.Rg3 mates 1 move sooner. I saw 29.Rg3, but I preferred 29.Rg5 for five reasons:

1) Sets a trap -- Black could lose instantly.
2) Burns some clock -- Black must think a bit to avoid #1.

3) Complacency -- Black might think he's safe, and miss the second threat.

4) Practice -- In future games, I probably won't have the luxury of two ways to win, and I'll need to be able to (see and) calculate a sac line. I had already invested the ~20 seconds to work out the sac to mate, so I decided to have some fun. (Also, I enlisted his help to double-check my work!)

"... Replace costly computation with fast pattern recognition. ..."

The usual line is 3.Nf3 g5 4.Bc4 Nc6 5.d4 d6 6.O-O Bg7 7.c3 h6, and Black plans Ne7, O-O, "with a Swedish Defense" (except that I've never heard of a Swedish Defense in any other context). If White deviates, I freelance.

Fish in my waters are generally confuddled by a non-Muzio positional grind, and sometimes trade away all the tension, esp. with an early e5 without support. I used to feel that KGA should produce a KID-like pawn storm, but that never works.