I don't see why 1.F4 is so strong in my first move statistics (58). It's surrounded by weaker moves. Even after this game, it still doesn't seem that strong. I think that number will come down, but this game didn't help it. I probably made my first mistake with |6.m13.

When I played |12.o19, my plan was, roughly, |18.q10 19.s11 20.r12 21.r13 22.q14 23.q15 24.p16 25.r17 26.s19. But when the time came, and Gyorgy actually played |17.r8 like I had hoped, I suddenly noticed that my plan had a fatal flaw: |18.q10 could be followed by 19.q11, and then 20.p12 21.r9, with a threat of p10. Still, I thought there was maybe some hope of following that with 22.p8 23.q12 24.s9, but if 23.p13 instead, I'm screwed.

So I started searching for alternatives. |18.q11 looked mildly promising, but after 19.o9, then what? Some support like o7 would be nice, but how could I sneak that in there? That would require a some kind of diversion on the top. I came up with |18.r6, hoping for something like |19.q6 20.q11 21.o9 22.o7 23.t9 24.p5 25.t5 26.p9 27.t14 28.u9. That would have been something to behold. But the flaw was that |19.q6 was not a necessary response for Gyorgy. He could play |19.p7 instead, and then the trick moves don't work anymore because of the strong threat o9. Even so, this position lets me fight my way toward the edge.

So my dilemma was between |18.q10, with a slight chance of a blunder by Gyorgy at the very end, or |18.r6, with a slight chance of a blunder by Gyorgy on the very next move. Neither way was technically winning, but I had to estimate the chance of a blunder by Gyorgy. That's a rough thing to estimate, especially against the #1 ranked Twixt player. But |18.r6 was off the beaten path, so I tried it.

Unfortunately, it didn't work, but this is a combo to keep in my back pocket, in case I have the right support in the lower right corner next time.

(Some endgame lines: |26.r14 27.s15 28.q12 29.r12 with a double threat at p11 and q14, or |26.q12 27.r12, or |26.p13 27.q14 or |26.t13 27.r12 28.p13 29.q14.)

18.r6 is an excellent combination! The problem with 18.r6 is that it _looks like_ a bad move. So I immediately knew there must be a trick behind that... :) The sequence 18.q10 19.q11 20.p12 21.r9 22.p8 23.q12 had a higher chance in my opinion.

Maybe 1.F4 is strong because it is so efficient at helping to secure that corner battle.

Instead of O19, how about 12.L20 13.J17 14.J19 and now:

15.F18 16.K12 threatens O14 or i17.

15.G18 16.K12 17.O13 18.i7 19.L6 20.H18 21.E17 22.i16 looks good for black.

15.H18 16.H20 17.D19 18.K14 19.O13 20.i7 21.L6 22.G18 23.E17 24.H16 and maybe white has a chance, but I wouldn't want to be in that position.

Okay what did I miss?

In that last variation 25.G14 looks strong for white, so let me take back a couple of moves and instead play 24.F16 25.F15 26.H15. Okay now show me what I missed :-)

I'm not sure, but I have doubts that the response to |12.l20 13.j17 14.j19 15.h18 16.h20 17.d19 18.k14 would be 19.o13. This is not a very clear position; maybe 19.h14?

I believethat after 12.l20 13.j17 14.j19 15.h18 16.h20 17.d19 18.k14 the best answer is 19.o13. Then 20.g18 21.o19 22.m19 23.r16 wins the game for white.

Therefore, I feel that 10.k18 was definitely a mistake, 10.l18 is better. But even after the sequence 10.l18 11.j17 12.j19 13.h18 14.h20 15.d19 16.k14 17.o13 18.g18 the move 19.s17 wins it for white (at least that was the sequence I expected during the game, but it is always possible that I miscalculated something :).

12.L20 13.J17 14.J19 15.H18 16.H20 17.D19 18.K14 19.O13 20.G18 21.O19 22.Q20 two -- branches

-- 23.Q18 24.O21 25.U19 26.S18 27.R20 28.S21 29.V21 30.U22 31.U23 32.T20 33.T16 34.S17 two --++ branches

--++ 35.T17 36.Q16 two --++-- branches

--++-- 37.Q14 38.U16 39.R18 40.P18 41.M18 42.P19 looks like a draw

--++-- 37.U14 38.S13 two --++--++ branches

--++--++ 39.T11 40.V14 41.Q14 42.U16 is the same draw

--++--++ 39.T13 40.Q12 41.S11 42.R9 43.T8 44.S4 45.Q5 46.P3 great googly moogly. I must have made a mistake. White has at least a draw, anyway.

--++ 35.V15 36.U16 37.U17 38.W18 39.V17 40.U15 Could this be another draw?

-- 23.N21 24.M19 looks winning for black.

Fire away!

10.L18 11.j17 12.j19 13.h18 14.h20 15.d19 16.k14 17.o13 18.g18 19.s17 20.q15 what can white do? For example 21.q18 22.p13.

12.L20 13.j17 14.j19 15.h18 16.h20 17.d19 18.k14 and now after Alan's suggested 19.h14 20.o14 looks strong for black, for example 21.h10 22.f16 23.e17 24.g14

So, in summary, it looks like Gyorgy was right that 10.K18 is a mistake (possibly leading to a draw?) and 10.L18 is better (winning.)

Here's more detail about the branch 12.L20 13.j17 14.j19 15.h18 16.h20 17.d19 18.k14 19.o13 20.g18 21.o19 22.q20 23.n21 24.m19 25.r16 26.p18 27.u19 28.t19

-- 29.s20 30.u17

--++ 31.t18 32.s18 black wins

--++ 31.r18 32.q16 black wins

-- 29.s18 30.u21 black wins

In my post I should have said "summary so far" not "summary." That's too much analysis to be error-free.

Hm, I knew I can't beat David in counting :). But let's make another try:
How about 12.L20 13.j17 14.j19 15.h18 16.h20 17.d19 18.k14 19.o13 20.g18 21.r16 with keeping an eye on the aji of o19.

If 22.q20 23.u19 24.t19 25.s18, white wins.
If 22.r18 23.o19 24.q20 25.n21, white wins.

So what would be the way to win by black?

(I will get back to the 10.L18 11.j17 12.j19 13.h18 14.h20 15.d19 16.k14 17.o13 18.g18 19.s17 20.q15 variation later. Even if s17 doesn't work, it still looks like a fair game for white on the left side.)

12.L20 13.j17 14.j19 15.h18 16.h20 17.d19 18.k14 19.o13 20.g18 21.r16 22.r19

-- 23.o19 24.n19 black wins

-- 23.r18 24.q17 black wins

-- 23.p20 24.p18 etc. etc.

Everything looks ugly for white.