Register
to add
comments.
Game number: Main Page
White: nie_wiesz    Black: Alan Hensel
twixt.ch.14.1.1 (LG) | This game (LG) | Download JTwixt file
On 2007-06-25 at 16:37, Alan Hensel (info) said:
The opening is a little bit hard to explain. According to my stats, 1.c8 is strong, so I swapped it. His h15 peg is a downstream peg, for example 7.o12 8.n15 9.m13. My m8 peg is the beginning of a very misshapen dipper formation, with my Y-shaped potential path starting on the left, and skirting around his p10 peg on either side. His 7.l10 half-attacks the stem while half-attacking a branch of my Y, forming his own Y in the process. His branch threats are, more or less, n9 and k7. The interesting thing is the threat of n9, because that gets him to the top by going to the right. So, anyway, in typical Twixt opening form, I look for a way to make a new Y threat, this time with the stem on the right and the branches on the left. However, the only move I find leaves the branches pretty tall and airy, and my opponent's response, 9.h10, takes advantage of that, balanced high in the branches. I'm running out of room to continue the Y rotation, and 9.h10 turned out to be a very strong move, because most of the obvious replies to it don't work; for example, 10.k7 11.j9 12.n10 13.i7 14.j4 15.f4 or 10.h8 11.k8 12.k11 13.j11. I don't remember all the variations, but it took me a long time to come up with 10.l3, but I felt pretty good about it.

Then the n9 shoe dropped. At that point, the Y rotation is over. I'm committed to a line across the top, and now it's nie_wiesz's job to poke thru my line, while I defend it. The next few moves were pretty obvious, and by 16.t8, I didn't see anything he could do. Then he played 17.h6. That didn't seem to make any sense. I almost just laughed it off and played the obvious 18.f4 defensive move, but I got a bad feeling about it. I started looking for threats that could lead to his h6 downstream, and I finally saw it: q6!

I had considered q6 earlier, and brushed it off with an easy p7 answer, but on closer inspection, that doesn't work: 17.q6 18.p7 19.s7 (tilt parallelogram). So whatever my answer is, it must make the other side, 17.q6 18.r7, work, and as it stands with 17.h6, it doesn't! 17.h6 was just strengthening his left threat, but he might have been able to do without it anyway: 17.q6 18.r7 19.p4 20.n4 21.n3 22.p6 23.n5 (double threat with j5 and n9). This was a crisis. The obvious response, |18.f4, leads to defeat. I don't remember all the variations; it gets deep. I finally thought I found the answer with |18.h12, but I was deluded. I was thinking |18.h12 19.k12 20.j11 21.j14 and then my m8 peg has its own double threat that works (for example, 22.k9 23.q6 24.r7 25.p4 26.n4 27.n3 28.o9). But his actual answer, |19.I13, doesn't allow 20.j11, because his easy reply would be 21.j9. But I continued to use my h12 peg threat, hoping to throw this game into chaos. |20.k6 shouldn't work because of 21.k8, which renews the threat of White's connection to n5 that k6 broke (21.k8 22.j8 23.j11 24.f4 25.q6 26.r7 27.p4 28.n4 29.n3 30.p6 31.n5). So I breathed a sigh of relief when he played |21.j11. Victory seemed assured. A slight feint to the left creates more certainty for me, and then |24.f4 25.Q6. The other shoe has dropped.

I said I wasn't going to explore all variations, but I'll mention that I played 26.o9 instead of 26.r7 because 27.p4 28.p6 29.o6 (instead of 29.n5) and that has a strong coign connection with n9. I couldn't play 29.o6 30.o9 at that time because of 31.m7. So 26.o9 first, while m7 has nothing to connect to on top. At this point, I should have the game locked up, and we're just playing it out to the end.

But I got overconfident, and careless, and perhaps a little impatient. Instead of conservatively renewing the double threat with 34.r4, I went straight for 34.p8. Then 35.u8 shocked me. Suddenly, my winning move, w6, didn't work anymore. I didn't see any way around that.

Later, David Bush pointed out the tricky winning sequence, 36.u5 37.v4 38.w5! I still had a winning position! And I resigned, not knowing it! In a Championship game! You have to be very careful in Twixt.

(David also pointed out that 21.r11 might be a winning move for nie wiesz. But I didn't see it at the time, or anything like it; I think we were both focused on Q6.)

On 2007-08-09 at 16:33, dushoff (info) said:
Is 36.w6 also a winning move for black?

On 2007-08-09 at 17:45, Alan Hensel (info) said:
No, if |36.w6 then 37.t6 38.resign.

You must be logged in to add comments.