Analyzing human game

I've played a lot against the computer here lately.
The following game is i think a nice example of how people play. After playing, i reminded the computer and i know from it, that it makes around 3 serious errors in the game, but some people (sorry) seam to do it all the time. What i mean is, people often believe to long in a strategy because they've played it before. However more often board situations are new and we tend to forget that...

Here white tries to dominate the white center, while my defense doesn't care for such center, my bishops strike the center without pawns. So my game is to fool around with white, i can offer pawns to clear such a defense, and meanwhile make space to move my peaces around. Compared to the computer lev1200 easy game i got many more chances, it comes down to a poor pawn opening game by white, or sticking to long with a wrong idea.
In contrast computers adept better to opening types then we do, as we stick to what we've played before. I too stick to fianchetto, but mainly to the first two moves, the rest is (and should) be decided by the game you see, not what you've played before and felt safe in the past....