I think one of my canonical patterns shifted during this game.

Normally, I would think |34.q17, and then 35.s15 36.u18 37.t17 or 36.u17 37.q18 38.p19 39.u16 40.r20 41.u20. The problem I have with both of these variations is that the wrong guy wins.

Because of this game, from now on, I'll also consider |34.s18. It looks strange, like Black is heading downward instead of to the right. It looks like White should just be able to cut right by that, on either side, but in practice, neither side works: |35.r19 36.p16 37.s15 38.t16, or |35.t19 36.r16 37.s15 38.t15.

This surprised me so much, I wrote down every variation I could think of. Here is a copy of my notes:

if |35.p16 then 36.q18
if |35.t15 or |35.u16 then 36.q17
if |35.t19 then 36.r16
if |35.r19 or |35.q18 then 36.p16
if |35.r16 then 36.r20
if |35.q16 then 36.t16

I think it works. There is nothing White can do.

The |35.r16 variation looks most likely to cause trouble in similar situations. Black's drop-back must be above the line (as 36.r20 is), or there has to be extra support to make something like 36.t16 or 36.q17 work. Even 36.r20 needs some distant support on the left - in this case, for example, 37.o19 38.p19 39.n21 40.n18 41.l20 42.m16.

Maraca's response |35.p16 may look weak, but it was actually dangerous: 36.r20 looks like a safe response, but then 37.p20 38.o20 39.n21 40.q19 41.l20 42.l16 43.k14 44.j17 45.f18 46.??? (Notice the unusual strength of that m15 peg thanks to its good neighbors: 42.m16 43.o16 44.k15 45.m17 or 42.m14 43.k14 44.n16 45.m13.)

There are other interesting things about this game that I'd like to write about, but I'll stop here for now.

The moves you are analysing are all too late in the game I think, I already knew I've lost the game, but as you've seen you can still make some mistakes.

Apart from the already mentioned line to win, I hoped for this draw:
|40.r19 41.t19 42.t20 43.u21 44.r16 45.t15 46.s13 47.s14

Now to the mistakes:

|15.f9 and white wins. Actually I've planned that move but thought: "okay, Alan probably knows what he's doing". Something you never should do.
For example: |15.f9 16.f7 17.f6 18.g9 19.h10

Then |19.m15, another stupid move. Imho I should first play something like this: |19.g11 20.f10 21.e10 22.d9 23.c9 24.b8

|21.p12 doesn't look too bad, but there's no reason for letting black cut between N11 and N8.
Probably better: |21.m10 22.j12 (forced, if I did at 19 what I said before) 23.n12

And finally move |25.h19 was stupid too, after 26.i19 it was over. Maybe |25.e19 or |25.f19 or |25.d19 would work better, but it's anyway hard to say, because I already messed up the top left corner.

And I forgot to say, the first interesting move is |8.p8 I think. Now there are 3 possibilities:
- 7.r11 wasn't that good and I shouldn't allow Alan to play p8
- 9.n8 wasn't strong enough, there are also other lines, for example: |9.n7 10.m7 11.p6 12.n9 13.s9
- 7 and 9 were okay, but I messed it up later.
I actually don't know, but I tend to believe that 9.n8 is the problem. Even if I have a win at 15, it was only because Alan made a mistake at 14.

I wasn't so sure I was winning until I found |34.s18. I struggle with "my opponent probably knows what he's doing". This board looks pretty intimidating at move |21.p12, especially with the name "maraca" in the corner!

I agree with most of what you wrote, but if |15.f9 16.f7 17.f6, then why not 18.f11?

I think the game was lost later. My prime suspect is the failure to press the |19.g11 20.f10 21.e10 line, but it could be even later.

The most interesting thing about my |24.j20 is that it was a desperation move. It was only after I played |26.i19 that I realized that the k15 threat existed and that it proved I'd actually won the path to the left! I should probably make a point of analyzing my moves before I play them.

Also, I wondered why |27.s17 was so far out. What was wrong with |27.r17? For example, |27.r17 28.r19 29.q19 30.p20 31.n21 32.o18 33.k14 34.k20 35.f18 ? I must be missing something. Maraca probably knows what he's doing ;-)

Oops, that last line starting with |27.r17 28.r19 29.q19 30.p20 doesn't work. After 31.n21, 32.l21 33.o20 34.o22 ends it. Instead, 31.m21 32.n21 33.j18 looks like the win for White. Or 32.o18 (threatening n16, so then) 33.k14 34.k20 35.f18 (same ending as before). So I maintain the hypothesis that the mistake that cost maraca the game was |27.s17.

This game had interesting battles in all 4 corners. The upper-right corner was the most standard of the 4, so I can talk about it in pretty absolute terms.

If White had played |21.p10, then it's completely hopeless and useless for Black: 22.o10 23.t7 24.s7 25.r6, but in this game, that would have led to a much better position for Black along the bottom, which maraca wisely avoided.

The standard parts of this pattern are the link leaning parallel to the crucial diagonal (o6-p8), a peg at the standard blocking distance from the link (t7), and a support peg that is a 2-4 gap from the blocking White peg (r11). And the standard line that follows, to the bitter end, goes like this: |24.s7 25.r6 26.t9 27.t10 28.q6 29.u8 30.s5 31.u5 32.v5 33.w4 and in this distance from the corner, White wins. From this you can see that if the pattern were shifted even a little bit to the right, so that the White pegs touch the line, that final peg can't be placed, and it turns into a win for Black.

As it is, its value for Black, despite being a losing corner battle, is that it guarantees that Black can place a peg at t11 or u11, that connects to the right side. For the t11 peg, the sequences are:

|24.s9 25.s8 26.t11 27.q9 (the sequence maraca wisely chose later in the game),
|24.s9 25.s10 26.t11 27.r8 28.r10 29.q9 (which gives me a little extra influence at r10), and the trickier
|24.s9 25.u10 26.v10 27.s8 28.t11 29.q9 (which gives me an extra peg at v10).

And the u11 peg is achieved like so:
|24.s7 25.r6 26.t9 27.t10 28.u11 29.s8

Those are such forced moves that I could avoid getting my t11 or u11 until I knew one of them would help. There was still a lot of empty space in the lower right corner, and I had to see what was going to happen there first.

@Maraca : >>> 7.r11 wasn't that good and I shouldn't allow Alan to play p8

I don't understand why you say that. The line you mentionned (|9.n7 10.m7 11.p6 12.n9 13.s9) is nice for white...

@Alan : >>> So I maintain the hypothesis that the mistake that cost maraca the game was |27.s17

I agree with your analysis. The point is that it allowed you to play in a key place both for you (regarding T11 and N17 pegs) and for Maraca (regarding R11 and S17 pegs). I have noticed this is always good being able to do so.

...and thank you for your explanation of the top right corner. Optmization is mandatory at this level of playing.