Chu Shogi Puzzle #1

Sort:
dax00

Eyes must have skipped a diagonal. In any case, I missed an easier mate: K-7a DH-1g Ph-6b DK-7c (if BTx7c Ln-5c mate, so...) G-7b Ln-5c K-8a Lnx6bx7b mate

I think my brain saw the shape and mate pattern so vividly, it wanted so badly for that beautiful mate to be real. I normally just play chu shogi on auto-pilot, letting my brain analyze the shape and make the instinctual call based off that. Saves a lot of time and energy, rather than having to calculate.

evert823

The alert flags are messy again. Sometimes there is a new message and I don't see a flag. And all the time I see flags and no new message.

dax00

Yesterday, I did not receive notifications. Based off of the notifications that have popped up today, it appears to me that they are arriving belatedly in quick succession, in order to catch up.

Anyways, the reply to K-5a was SE-6c. K-7a has been established as a blunder, since it gives up forced mate. If you wish to continue, the puzzle is yet unsolved.

evert823

For my own convenience:

evert823

1. K-5a SE-6c

(2. DKx6c Lnx6c would be mate so I'll play something else)

And I see a possibility that the SE takes on 5b with igui which gives check.

Let's try - DKx6c


evert823

This was quite stupid, LnxDK #

 

dax00

Yeah, that's not the move. Very complex options available for both sides - makes for a good puzzle. My opponent resigned after SE-6c in the real game.

evert823

Ìnstead of DKx6c, maybe B - 3d ?

dax00
  • K-5a SE-6c
  • B-3d Lnx3dx3c
  • VMx3c SEx!5b
  • S-5b SEx5b mate
HGMuller

HaChu thinks K-7a is better defence (mated in 6).

 12 -79.94 137.0M 4:11.33 g12f12 h10h8 g8h8 j4l6 f11g11 e10h10 f12f11 i8h8,h8g9 f11g12 h10g11,g11f12 e12f12 g9h10
11 -79.94 54.1M 1:45.97 g12f12 h10h8 g8h8 j4l6 f11g11 e10h10 f12f11 i8h8,h8g9 f11g12 h10g11,g11f12 e12f12 g9h10
10 +0.92 21.2M 0:48.09 g12f12 h10h8 f7g9 h8g9+ g8g9 e10h10 f12g12 h10g9 g12f12 g9b9
9 +5.61 9.2M 0:24.05 g12f12 h10h3+ f7g9 e10g8 g9i11 g8g9 i11h12 g9b9
8 +5.97 2.8M 0:07.00 g12f12 h10h4+ f7g9 e10g8 i10g8 j4l6 f11g11 i8g9 f6h4+ g3h4 g8i6
7 +8.72 1.5M 0:04.27 g12f12 h10h4+ f7g9 j4l6 f11g11 h4f6 g8f8 f6f8 g9f8,f8f7 i8j7,j7i7
6 +12.71 465336 0:01.35 g12f12 i8j7,j7i7 d10d9 h10h7 g8h7 e10h7
5 +12.71 266970 0:00.85 g12f12 i8j7,j7i7 d10d9 h10h7 g8h7 e10h7
4 +12.71 46727 0:00.15 g12f12 i8j7,j7i7 d10d9 h10h7 g8h7 e10h7
3 +12.44 30488 0:00.09 g12f12 i8j7,j7i7 d10d9
2 +12.05 415 0:00.00 g12f12 i8j7,j7i7
1 +12.05 150 0:00.00 g12f12 i8j7,j7i7

After K-5a it is mate in 4:

7 +79.96 1.0M 0:01.91 e10f10 h12g12 h10g11 h11g11 f10g11,g11h12 g12h12 i8h10
6 +79.96 397900 0:00.85 e10f10 h12g12 h10g11 h11g11 f10g11,g11h12 g12h12 i8h10
5 +79.96 183541 0:00.48 e10f10 h12g12 h10g11 h11g11 f10g11,g11h12 g12h12 i8h10
4 -5.38 53752 0:00.15 e10f10 h12g12 f10f7 g8f7 h10g9 g12f12 g9b9+
3 -12.08 19830 0:00.06 i8j7,j7i7 f11h9 c2k10+ l10k10
2 -12.18 10436 0:00.03 i8j7,j7i7 d10d9
1 -11.79 247 0:00.00 i8j7,j7i7

 

evert823

Your coordinates are like this:

HGMuller

Indeed, normal algebraic coordinates.

Oops, forget what I said. I had not noticed that 5c was a +R, instead of a DK.

There also seems to be something wrong with HaChu's 'promote on entry' option that no one ever noticed before: in some of the lower-depth variations it seems to think it can promote by entering its own zone!?!

dax00

@HGMuller : I challenge HaChu to a correspondence game to be played on these forums. In addition to the engine, you can use any other resources you wish.

HGMuller

Well, at the moment I (and my computer) are tied up rather heavily in the correspondence championship Tenjiku Shogi, at gamerz.net. But I will keep it in mind for when most of the games there have finished.

BTW, gamerz.net has also a Chu Shogi section. It would be far more convenient to use that, than improvise here.

HGMuller

OK, after 1... K-7a 2. DH-1g it is definitely mated-in-6:

11 -79.94 134.0M 5:14.23 k10k7 l6k7 j10j8 i8j8,j8i8 e12d11 h10h11 f12e12 h11f11 e12d12 k7f12+ e11f12 f11f12
10 -2.71 77.4M 3:27.59 k10k7 l6k7 j10j8 i8j8,j8j9 f6i9 j9i10,i10i9 h11g11 h10l10 l11l10 c2m12
9 -1.38 36.2M 1:44.67 k10k7 g3k7 j10j8 i8j8,j8j9 e12d11 h10f10 h11g11 e10f11,f11g12 f12g12 f10f7 g8f7 j9i10,i10h9
8 -1.50 20.2M 1:02.15 k10k7 l6k7 j10j8 i8j8,j8i8 f6i9 i8i9,i9j10 g8g7 h10i10 f11g11
7 +2.29 9.0M 0:27.62 k10k7 l6k7 j10j8 i8j8,j8j7 f11g11 j7k6,k6j7
6 +5.05 3.2M 0:10.56 f6i9 i8i9 d10e9 h10h7 g8h7 i9j10,j10i10
5 +6.95 1.2M 0:03.96 k10k7 i8j7,j7k7 d10e9 h10h7
4 +7.32 336526 0:01.20 f6i3+ h10f10 i3l6 f10f7 g8f7 e10f11,f11e10
3 +14.42 3865 0:00.01 g8i8 h10f10 f6i9 e10f11,f11g12 f12g12 f10f7 i8i6
2 +14.39 1729 0:00.00 g8i8 h10f10 i8i9 f10f7 i10f7
1 +15.59 156 0:00.00 g8i8 c2i8

After 2... Ph-6b it is mate in 3 because you can move the +R away with check (3. +R-7c) to make room for the Lion:

4 +79.97 61599 0:00.18 h10f10 e12f11 i8h10 f12e12 h10g11,g11f11
3 -2.18 46449 0:00.15 h10f10 e12f11 i8h10 f12e12 f10f7 g8f7 e10f11,f11e10
2 -9.62 11933 0:00.04 i8j7,j7i7 d10d9
1 -9.22 319 0:00.00 i8j7,j7i7

After 1... Ka5 2. SE-6c HaChu still favors to simply take the Lion (2... DKx4e). I can let it search somewhat deeper, but don't see good prospects for achieving checkmate after that.

dax00

K-5a SE-6c wasn't under the impression of a mate. It simply allows sente to trade off a bunch of stuff to make gote's king eternally weak and eventually win a long grind. And if gote blunders, then there are possibilities.

Indeed, gote should take the lion. 

  • 1... K-5a
  • 2. SE-6c DKx4e
  • 3. SEx!5b B-5b
  • 4. SEx7bx8a BTx8a
  • 5. DKx5b Sx5b
  • 6. Bx4e DH-9b

...and sente should be very happy


The point of this puzzle was that I would have offered the mistake 3. DH-1g, to which there is a clearly good reply. It's more fun to think it through than to use an engine. I may not be as tactically adept as a fair amount of chess players (much less an engine), but I can at least see 4 move tactics, and my positional knowledge of chu shogi is by far unparalleled.

gamerz.net is a bit too fiddly for my liking. How about on chessvariants?

HGMuller

I never used Game Courier. Last time I tried (just for looking at games others had played)  it was pretty much unworkable. The pages contained links to ads that were extremely slow to load, and the whole thing is based on server-side scripts, so that you cannot do anything without contacting the server.

Anyway, I have another puzzle here. It is a mate in 7. (Tsume, so checks only!)

dax00

It will take me forever to see the pieces, if they look like that. If everything were in kanji, I could analyze the board in seconds

HGMuller

Well, take a piece of paper and write them as kanji, then. Or set up the position in woodware.

dax00

Let's see

  • Bx4f+ FKx4f
  • Ln-3i FBx3i
  • ...

By the way, how am I supposed to tell which pieces are already promoted? There are 2 DEs, which means at least 1 of them promoted from a GB.