Good shapes and a direction of play are totally missing in the early game, which then results to a “Free for all” game.
For example:
We are just a few moves in and already the shapes are over-extended and greedy.
From the two stones you extend 3. ( the ideal extentions are 2+1 from the amount of stones in your “wall”. 3 stones, then 4 gap extention. 4 stones, then 5 gap extention, and so forth)
Similarly the stone a C9, doesn’t really connect with D4 either since there is no shape that is viable like that.
D4 and C6 is a knight’s move
D4 and C7 is a large knight’s move
D4 and C8 is a not any knight’s move that has a name
and you are even further at D4 and C9.
So, C9 is just a pre-placed pincer stone. That’s the “direction of play” for it. In this case, it needs to be used like that (a pincer is mainly played to restrict the potential base/extension of the opponent and is not meant to be immediately defended/expanded upon, unless you desire to totally change direction of play). However, two moves later:
You abandon the idea of this being a pincer stone and you try to defend/build it up.
@Shen0927 suggestion is the correct one. Worth noting is that if the proper shapes where followed, you’d still have a nice pincering stone with no real weaknesses and the opponent would have been stifled in overconcentrated shapes, while you raked in free points. E.g.:
White is not alive yet, both your groups are alive and gathering points while white struggles.
Next:
What does H3 do?
It doesn’t really expand the wall you just built. (it is diagonally placed and too close for that)
It doesn’t really mesh/work together with Q4 (it is too far)
Where is the direction of play here? If you want to leave the fight on the bottom left and expand, that is reasonable, however it is a good idea to locate where is the larger gain.
The AI suggestion of sealing White in with E3 is excellent.
K4? is also good since you are now threatening to build a moyo.
Q10? a similar concept.
K16? You advance your influence while reducing the potential that White has build on the top.
O17? Yo secure your corner, you build up the left side for a potential moyo in the next moves and reduce White’s potential up top.
Next:
After the bizarre fight on the bottom right corner you end up with a good result.
A solid wall facing another nice wall of yours and sente!
Here are two very basic ideas that the direction of play has:
a) Extend/expand towards the largest empty area on the board (not by coincidence the main AI suggestion does that)
b) Extend/expand your own stones (also not by coincidence those are almost all the other AI suggestions).
A whole quarter of the board is about to become your moyo, just by attacking the O3 stone. If the opponent panics and tries to get it out, you are well on the way of winning.
The most basic/obvious move which I’d probably play is the standard “base” (and in this case also base-stealing) extension at L3. I’d play there or L4 at most. The other ones seem to operate on some more complex ideas/fights/goals.
Here is how that would have looked:
This is a mediocre result, which is why the AI isn’t considering L3, but it is a perfectly playable move for SDK level.
a) White reduced your points but is not alive yet. (while your stones are safe and connected White M3, Black M2, White captures, Black connects with N1 under the stones.)
b) The red squares are on the way to becoming points.
c) The blue line is a very scary sector line of influence that spans half the board.
d) There are still plenty of empty spaces to expand and get even more points. If white invades/approaches the top right corner now, you will definitely gain points elsewhere with relative ease.
SIMPLE SOLUTIONS FOR IMPROVEMENT:
a) Only try to play correct/connected shapes (see Haengma for details) and extensions, until you see where there is room for greediness.
b) Follow the aforementioned two simple tenets of direction of play.
c) Try to think “what does this move really do, in the context of my other stones and the rest of the board?” when you are ready to move out of a local fight and want to play elsewhere on the board.
Disclaimer:
All the above need to be taken into the context of improving from DDK to SDK. It is obvious that the above advice is not valid for anything above/beyond that.