1.h4 seems to be David's new favorite first move. It's one of my favorites, too. I'll skip the rest of the opening moves, because I can't remember back that far... Okay, on to the mid-game!
But somehow, something David did on the left, starting with his |17.g11 and |19.g15, made that not work. I wonder if he knew that, and if so, how he thought it up. In any case, it wasn't until after I played |20.h18 that I realized the original plan was broken. The new, broken plan looked more like this: |22.f1923.k1624.k1725.r1026.q1327.p1128.o1229.n12, which would have been a sad ending indeed. The Twixt roller-coaster had bottomed out.
Despondent, my next few moves, |22.h13 and |24.h8, were purely defensive "what the heck, it can't make things worse" moves. I wonder if David thought I had a plan. Hoping to come up with a working plan later, I looked for an alternative to |26.f19 that would give me the most strength on the right. That's what that crazy-looking |26.h23 is all about, in case you were wondering.
Right around |30.i13, it occurred to me that the ladder chase from |17.r10 might not be so bad, after all, because at the end of it I could play s21, which is also a downstream peg from my large group on the left. Score! The roller-coaster is up again.
Sudden dip in the roller coaster as I realized that my plan might end in |40.m2041.g1942.f2243.j22, and I started to regret |26.h23. But then I realized that |40.l20 closed that gap, and had the same effect. That was actually a pretty quick realization, because I know that the coign defense, |40.m2041.m2142.l2243.n1944.n21, has a similar effect to just advancing the 0-2 ladder chase: |41.n1942.n21. But in this case, without the dangerous gap.
Unfortunately, David had the final clever move, which I didn't see coming until he played it. It's a double threat: |42.s2243.m2044.n2145.o2146.p2247.q2248.q2349.p24, or |42.q2243.u20. I took some time to scout for all the downstream pegs from L20, but I proved to myself that there is no overlap between them and the moves that break the u20 threat. Game over.
In hindsight, |32.p12 was probably better than |32.p10, but it's not close enough to that 2-5 gap (m15-r13), so it probably didn't make a difference. If I made a mistake, it might have been as far back as my response to |17.g11 or |19.g15.
At move 16, I knew that the ladder chase to the lower right from the threat of |17.r10 didn't work, so I formed a plan along the lines of |17.r10 18.q13 19.p11 20.o12 21.n12 22.l14 23.o14 24.m20 25.m21 26.n18 27.k20 28.m16. I felt I was winning.
But somehow, something David did on the left, starting with his |17.g11 and |19.g15, made that not work. I wonder if he knew that, and if so, how he thought it up. In any case, it wasn't until after I played |20.h18 that I realized the original plan was broken. The new, broken plan looked more like this: |22.f19 23.k16 24.k17 25.r10 26.q13 27.p11 28.o12 29.n12, which would have been a sad ending indeed. The Twixt roller-coaster had bottomed out.
Despondent, my next few moves, |22.h13 and |24.h8, were purely defensive "what the heck, it can't make things worse" moves. I wonder if David thought I had a plan. Hoping to come up with a working plan later, I looked for an alternative to |26.f19 that would give me the most strength on the right. That's what that crazy-looking |26.h23 is all about, in case you were wondering.
Right around |30.i13, it occurred to me that the ladder chase from |17.r10 might not be so bad, after all, because at the end of it I could play s21, which is also a downstream peg from my large group on the left. Score! The roller-coaster is up again.
Sudden dip in the roller coaster as I realized that my plan might end in |40.m20 41.g19 42.f22 43.j22, and I started to regret |26.h23. But then I realized that |40.l20 closed that gap, and had the same effect. That was actually a pretty quick realization, because I know that the coign defense, |40.m20 41.m21 42.l22 43.n19 44.n21, has a similar effect to just advancing the 0-2 ladder chase: |41.n19 42.n21. But in this case, without the dangerous gap.
Unfortunately, David had the final clever move, which I didn't see coming until he played it. It's a double threat: |42.s22 43.m20 44.n21 45.o21 46.p22 47.q22 48.q23 49.p24, or |42.q22 43.u20. I took some time to scout for all the downstream pegs from L20, but I proved to myself that there is no overlap between them and the moves that break the u20 threat. Game over.
In hindsight, |32.p12 was probably better than |32.p10, but it's not close enough to that 2-5 gap (m15-r13), so it probably didn't make a difference. If I made a mistake, it might have been as far back as my response to |17.g11 or |19.g15.