So far I have been investigating 12 problems from Polgars middlegame brick. I wasn't able to solve a single problem without aid. When I read the solution, I often had still no "aha"-experience. Only after 4 hours messing around with Arena, I got the feeling that I began to grasp the essentials of the position.
From the 12 solutions, 6 were busted by Rybka. Which means that even grandmasters are often wrong in complex tactical positions. Hisbestfriend posted about the same phenomenon in Euwe's book.
It is time to draw a few conclusions, in order to direct my future study. After an impressive detour I'm back again at complex tactical situations. I used to have the idea that if I do enough simple problems, that would be the road to learn to do the complex problems. After solving 100k+ problems it is proven that this idea is wrong. I couldn't even solve a single problem from Polgar's book! (since I restarted a week ago)
The coming time I will focus on complex tactical problems from Polgar's middlegame book. I don't know a better method than to use 4 hours per problem. At least I skip the unviable approach with trial and error during the first half hour or so, hoping for inspriration. After a few minutes I go immediately to the solution. Of course I will try to find a method that takes less than 4 hours per problem, but at the moment I haven't a clue how that kind of study should look like.