Barely won


tips?

On Move 9 I’d recommend playing D5 for D5 C6 O3, since O3 is an important follow-up to K4. The C5 hane is always eventually gote for Black.

Move 15 is a little overconcentrated. It’s probably best to play away, White isn’t threatening anything definite that you have to be worried about.

Move 17 is very overconcentrated, and White’s hane at C8 is big. No joke, if you learn to respond to B7 with C8, and vice versa play C8 as White against opponents who allow you to, then you’ll be better at playing this particular shape than many, many players below 5k.

On Move 29 you should follow the proverb don’t atari the cutting stone and simply connect at D7, making White move out under his own steam. Perhaps exchange C9 for B8 first.

On Move 35 the proper shape is F5, which captures using a very short ladder.

There isn’t really much point in Move 37 – there’s little territory there.

On Move 41 your hane leaves a cut – you seem to be worried about this because you fix on Move 43. Better to have extended in the first place.

Why did katrain-6k go T5 on move 232?

I’m just super impressed that you - a 20 kyu player - were able to beat a 6k opponent -

even more impressive since White basically grabbed the entire outer edge, and you managed to build a big enough inside moyo, and then save your eyeless group at the end. I commend your fighting spirit!

thanks I also have beaten kugustu and elotest a couple of times

yeah made some obvious mistakes
doesn’t move 29 also protect the cut
move 41 I was expanding the black group wasn’t trying to protect a cut

I don’t think it’s due to @kano4526’s playing (sorry!), because katrain completely lost its marbles on move 232, allowing black to live easily. It looks like a bug. Here’s what the bot said about it in chat:

W T5 caused a significant score change (B gained 39.6 points) → Win Rate B 50.6% Score W+26.0 AI Thoughts: Using policy based strategy, base top 5 moves are Q6 (30.93%), S5 (28.02%), R6 (22.39%), S6 (10.16%), R5 (4.80%). Picked 26 random moves according to weights. Top 5 among these were T5 (0.02%), P7 (0.01%), P6 (0.01%), H18 (0.00%), P3 (0.00%) and picked top T5.

So perhaps it picked 26 random moves, but none of the good moves? I’m not sure how KaTrain works.

yeah made no sense to me as well