From: Colin Paul Adams COLINA DEMON CO UK> Date: 30 nov 1999 Subject: CHU: Black's quick attack on the fifth file >>>>> "Nick" == Nick Bardsley ICC CO UK> writes: Rikard> Forget about the bare king rule! There is in practice no Rikard> bare king rule. At least the bare king rule would never Rikard> be used and it is very specific. Agreed. I think that is quite enough of rules discussion. How about some real Chu discussion? Back in October, I was playing Thomas Majewski, when I decided to try a strategy that I had been mulling over in my mind for some time, but didn't really believe it was safe in the face of determined counter-play. But since Thomas was registering people for the German Shogi Lightning championship at that time, I thought it was worth a try. Subsequent reflection convinces me that it might be a very good strategy indeed for Black, as I have been unable to think of a telling counter by White. So this article is to see what people think White should do. The strategy is to threaten a quick attack on the fifth file, with the aim of getting a secure space advantage in the centre. To demonstrate the threat is a real one, here is a game I have just played against my computer program (I wouldn't normally inflict White on the computer, but I was checking that I had managed to cure a particular bug). Time limits were six hours each (but no time consumed on the clock if a move is played in no more than 60 seconds) (White's time used is given so you can see how slowly the computer plays - this is the version that I will release by this weekend at the very latest - promise): 1 Ln 7j - 6h, Ln 6c - 7e 6 seconds 2 P 8i - 8h, P 5d - 5e 41 seconds 3 Ln 6h - 6g, Ph 7b - 5d 51 seconds 4 Ph 6k - 8i, P 8d - 8e 77 seconds 5 P 5i - 5h, P 10d - 10e 86 seconds 6 P 10i - 10h, P 3d - 3e 105 seconds 7 P 3i - 3h, DH 9c - 11e 115 seconds 8 DH 9j - 11h, DH 4c - 2e 166 seconds 9 DH 4j - 2h, P 6d - 6e 224 seconds 10 P 7i - 7h, C 10a - 9b 263 seconds 11 Ky 7k - 7i, C 3a - 4b 239 seconds 12 C 10l - 9k, C 9b - 9c 193 seconds 13 DK 8j - 7k, S 9a - 9b 183 seconds 14 C 9k - 8j, C 9c - 8d 176 seconds One thing I have realised since playing the game against Thomas, is that it is safest to NOT play P - 6h for as long as possible. This way, the Kylin can always retreat to 6h or 7k, (whichever square the Copper leaves free) Anyway, the moment that 8d is occupied, is the best time to attack the Lion. 6d is a terrible square for the Lion, guaranteeing that the attack will suceed. But the Lion has nowhere else to run, so: 15 Ky 7i - 7g, Ln 7e - 6d 265 seconds 16 C 8j - 7i, P 7d - 7e 223 seconds 17 P 5h - 5g, S 9b - 9c 327 seconds The computer is using an augmented 2-ply look ahead (that's two moves by the Japanese count). The augmentation consists of looking ahead one more move if the last move was a capture, and then continuing to look ahead as long as the next move is a capture on the same square. This helps for the common situation of many pieces bearing down on the same square, but is of little use when a multi-square dogfight starts, as is about to happen. Anyway, with only a 2-ply look-ahead, the computer cannot see the danger, so it continues to develop. 18 C 7i - 6h, C 4b - 4c 317 seconds 19 P 5g - 5f, S 4a - 4b 492 seconds 20 C 6h - 5g, P 12d - 12e 469 seconds 21 P 5f x 5e, Ph 5d x 5e 290 seconds I guess that objectively, C - 5d is the best move, just suffering the inevitable loss. 22 C 5g - 5f, P 6e - 6f 356 seconds 23 Ky 7g x 6f, Ln 6d x 6f 686 seconds 24 Ln 6g x 6f - 5g White's clock stood at 4:32:20 Clearly White must evolve a strategy that does not force his Lion to go to 6d. Any suggestions? -- Colin Paul Adams Preston Lancashire