Alas I have failed to break 20 after all. <Once> is back in top form: P Cramling vs E Tsuboi, 1992 (extract) <34. Kf1? Now Fritz is tut-tutting. "Zis is a mistake!" he cries. "Ze sauce is curdled, zere are bones in ze raw fish. Vot about Kh1?" Or in chess terms, the eval has fallen from over +9 to just +1.2> So I lose a point for preferring Kf1.

Now, I freely admit that I was brought up in the provincial north of England where the closest we got to a cosmopolitan culinary experience was putting horseradish sauce on our roast beef. And serving roast beef without Yorkshire puddings was considered a heinous crime punishable by excommunication to the softy south.

But I have to say that I just don't get the sea cucumber. Paying more for something just because it is rare and difficult to cook, but it doesn't taste of anything?

One time the sushi chef gets a little brown bottle with Japanese writing on it and serves me some of its contents in a little cup. The man sitting to my left asked me what it was and I said (as it often happened) “I have no idea”.

So he asked the sushi chef and the chef at first ignored him. The man persisted and the sushi chef then said it was “special”. The man still persisted and the sushi chef finally said that it was the inner lining of a sea slug.

The man then turned to me and said “Buddy, I used to think of myself as an adventurous eater, but you are way out there!”

I tried to find out if a sea cucumber is the same as a sea slug. At least in the USA there seems to be some controversy as to whether they’re the same or not. I also found some recipes for sea cucumber of which the following is sure to be your favorite:

White has the bishop pair for a bishop, a knight and a pawn.

The black castle has many weak points. In particular f6 can be further weakened with 30.Rc8:

A) 30... Qxd7 31.Bxf6+

A.1) 31... Rxf6 32.Rg8+ Kh6 33.Qxf6 (threats 34.Qh4#)

A.1.a) 33... Be7 34.Qg7+ Kg5 (34... Kh5 35.Qxh7+ Kg5 36.Qxg6+ and mate next) 35.Re5+ Kf4 (35... Kh4 36.Qh6#) 36.Rf8 Qf5 37.Rfxf5+ gxf5 38.Qg3#.

A.1.b) 33... Qf5 34.Qxd6 + - [R vs N].

A.1.c) 33... Kh5 34.g4+ Kh6 35.Qh4#.

A.2) 31... Kh6 32.Qe3+ and mate in two.

B) 30... Rxd7 31.Rxd8 Rdxd8 (31... Rfxd8 32.Bxf6+) 32.Ra8

B.1) 32... Bc7 33.Qe7+ Rf7 34.Bxf6+ Kg8 35.Qe6 with many threats (36.Bxd8, 36.d6, 36.Rc1, etc.).

B.2) 32... Nc4 33.Rxa6 again with several threats (34.Bxf6+ Rxf6 35.Ra7+, 34.Ra7+ and 35.Rb7, etc.).

Since 30. BxP+ is by brute force, there should be a better way.

30. Rc8 is the logical choice to attempt to win the exchange, and since QxR would be too heavy a loss to Black, 30... QxB is again logical and forced. Then White, in order to maintain enough pressure, cannot exchange Q, hence 31. BxP+ RxB.

Now is the critical digression, 32. Qg8 + looks forceful, but the Black King got an escape square h6, then the White Q at g8 would either be forced to exchange (which White doesn't want,) or trapped, or needs more moves to get out to be useful in the attack. Hence 32. Rg8+ instead, and after 32...Kh6, 33. QxR.. White got an easy game, and the rest follows naturally.

White to play (30.?) "Insane"

It took me awhile to see the winning idea, then a little longer to work through to a winning endgame.

30. Rc8! Qxd7 31. Bxf6+!

This is the key move that makes this line work.

31...Rxf6 32. Rg8+ Kh6 33. Qxf6 Bc5+ 34. Kf1 (Kh1) Qxd5

(White wins brilliantly after 34...Be7 35. Rxg6+! hxg6 36. Qh8+ Kg5 37. h4+ Kf4 38. Qe5#.)

35. Qh4+ Qh5 36. Qxh5+! Kxh5

(White wins back the piece, thanks to 36...gxh5?? 37. Re6#.)

37. Re5+ Kh4

(Or 37...Kh6 38. g4! (threatening 39. Rh5+ gxh5 40. g5#) Be7 39. f4 and mate can't be stopped. In this line, Black last longer with 38...Rg7 39. g5+ Kh5 40. Rxg7, but that's hopeless too.)

<johnlspouge> is right, this puzzles causes the engines to calculate for an “eternity". I’m running a “tournament” using the Sunday puzzles to find out how quickly various engines can find the winning move in Infinite Analysis mode since I think that this may be a better way to assess various engines’ relative performance in analysis rather than look at ratings or rankings based on engine tournaments at finite time controls. And I consider not only how long it takes the engines to find out the winning move, but to determine that the move indeed represents a winning advantage, with I define as an eval greater than [2.00] (Arena GUI’s threshold is [1.50]).

I’ve only run 6 puzzles prior to this one and most engines can find the winning move and determine that the player has a winning advantage in about 10–15 plies in less than 2 secs. But on this puzzle Rybka 4 didn’t find 30.Rc8 until almost 32 secs at d=7, with an eval of only [+0.73]. After 7.5 mins at d=21 Rybka’s evaluation never reached higher than [+1.19].

Houdini 1.5a wasn’t much better. It found 30.Rc8 at d=12 in less than 1 sec but with an eval of only [+0.48]. After 1.7 mins at d=25 its eval never exceeded [+1.19].

Stockfish 2.0.1 was better but borderline per my winning advantage criteria. It found 30.Rc8 at d=19 with an eval of [+2.02] after 5.5 secs. But at d=20 its eval dropped to [+1.97] and after 16 secs its eval was only [+1.87].

On my computer Toga II 1.4 5c beta found 30.Rc8 at d=13 with an eval of [+0.88] after a little over 3 secs. After 3 mins at d=21 its eval was still only [+1.69] but steadily rising.

Spike 1.4 seemed promising. It found 30.Rc8 at d=11 with an eval of [+1.48] after only 1.2 secs, even faster than Stockfish. But after almost 7 mins at d=27 its eval was only [+1.26] although in between it got as high as [+1.78]

Anyway, I was curious as to how you get prawns out of the North Sea. I thought much of the seafood we eat is from warmer climates.

Bill Wyman, ex-Rolling Stones bassist has a fish and chips shop somewhere in England. Is this fish traditionally fried, or can one order it broiled (for health reasons)?

Here's a breakout:

<30. Rc8!! Qxd7>

If 30... Rxd7, then 31. Rxd8 Rdxd8! 32. Ra1! Bc5+ 33. Bf2! Rd6 34. Bxc5! Rxe6 35. Bxf8+ Kxf8 36. dxe6 looks sufficient.

<31. Bxf6+! Rxf6 32. Rg8+ Kh6 33. Qxf6> All forced up to this point, and possibly played as an intuitive positional attacking sacrifice.

<31...Bc5+!~> This avoids a forced mate.

If 33... Qf5, then it's mate after 34. Qxf5 gxf5 35. Re6+ Kh5 36. g4+ fxg4 37. fxg4+ Kh4 38. Rh6#.

<34. Kf1!?> This wins, but much stronger is 34. Kh1! which results in an inescapable mating web following 34... Qxd5 35. Qh4+ Qh5 36. Qxh5+ Kxh5 37. Re5+ Kh4 38. g3+ Kh3 39. Re4 Be7 40. Rg7 Bd8 41. Rxa7 .

<34... Qxd5> Maximizes resistance and avoids the quick mate 34... Qf5 35. Qxf5 gxf5 36. Re6+ Kh5 37. g4+ fxg4 38. fxg4+ Kh4 39. Rh6#.

<35. Qh4+ Qh5 36. Qxh5+ Kxh5>

Not 36... gxh5?? 37. Re6#.

<37. Re5+ Kh6?!> This makes it easy.

Putting up more resistance is 37... Kh4 38. Rxc5 Nb3 39. Rd5 Rc7 40. Re8 Rc4 41. Ree5! when White will win slowly and with difficulty, and will wish she had played 34. Kh1! instead of 34. Kf1!? earlier.

<38. Rxc5> This is the obvious "winning" move, but much stronger is 38. g4! Rg7 39. g5+ Kh5 40. Rxg7 h6 41. Rxc5 .

<38... Nc4> This allows White the clever winning move that will force immediate resignation.

Putting up more resistance is 38... Nb3, but White is still winning after 39. Rd5 when play might continue Rg7 40. Ra8 Rc7 41. Rxa6 .

<39. g4!> 1-0 This forces immediate resignation.

I'm a New Zealander, and we do fish 'n' chips in the English fashion. I'm sorry to tell you that traditionally, it's deep-fried.

If 32...g5, e. g. 33.Rxa5* Rfe8 34.Qg4 h5! 35.Qd4! (35.Qxh5 Rh8 36.Qg4 Rxh4 =) 35...gxh4 36.Ra1

click for larger view

White should probably win in the long run (Rybka: +0.60)

*Or better 33.Bf2 Rfe8 34.Qh3 Bb4 35.Qf5 Re5 36.Qb1 Bd2 37.d6 Bc3 38.Qc2 b4 39.Rd1 Nc4 40.Bd4

click for larger view

Yes, I know of your predilection for Toga. And you may be right. Who knows? There may be even better engines out there for solving puzzles. I don’t know of any systematic study of engines’ performance on puzzles, so I decided to try and find out. And I didn’t know of any objective way to rank them other than speed in finding the winning move AND determining that the move is, indeed, winning. So that’s what I’m trying to do. My biggest problem so far is that the engines are too good and they find the winning move too quickly! A difference between finding the winning move in 0.2 secs vs. 0.4 secs just doesn’t seem to have much practical significance, so I’ll have to figure out a way to present meaningful results (provided there are any) in a reasonable way.

Here in Hawaii, we have the equivalent of the Soup Nazi, although he is a sushi chef. He has been written up in several articles.

On the door are the words: TRUST ME.

You sit down at the bar and he brings you whatever he wants you to have. You must eat it correctly or else he will kick you out.

As the tail landed near this guy, he keeled over, had a heart attack and was pronounced dead at the scene (sorry for being a Friday night downer). Anyway, the guy's family sued the sushi restaurant, claiming that the actions of the chef, the shock of seeing this flying object, induced the fatal heart attack. Only in the good 'ol USA, where everyone needs a law degree to protect themselves from frivolous law suits.