First, the training of a great harvest, much faster than the self-study to be
Second, the new knowledge is not how to tell when (A *, except number theory), but still deeply felt the difficulties of training, the rhythm of seven days is this: say this is what (10min), we stress the basic algorithm (10min), after immortal title is a long time, average difficulty purple, so hear foggy. . . There is a twice-simulation game, feeling the way of thinking between five days has been significantly improved.
This training is the biggest gain for the mode of thinking of it adapted. I used too dependent algorithm, templates, template but for some memories but not impressive enough, the examination room may not be able to play out, resulting in a large part of the flexible use of check-points can only playing, has been set to the board, if the topic not designed to test the algorithm, but to be transformed by it to maintain something, then I would absolutely gg up. So, I think said before "50min thinking, 10min knock Code" kind of thing very right, which should not think of where to play, should first clarify ideas a problem, and then implement the code, the code is more important than thinking.
Secondly, the code forces. For some ideas, might simply can not achieve, or need to adjust for a long time, causing the game to waste a lot of time, some violent dfs the subject is difficult to bring up, so before the race I should write about data structures and search violence issues, to raise my code force.
There is dynamic programming, the teachers feel very equation of metaphysics, profound, and then look back, the equation does not seem so profound, would like to direct state violence find transferred on the line, the focus lies brush turned a deaf ear, and more are summarized.
The focus may be on the future review board, do simulation questions, brush dp, engage in the data structure.