A large one of juxtaposed rows Rank # 9.
Test questions have not changed completely (and not simulated annealing) so they can only mutter a solution to a problem
However, do not want to Tucao T2 is the original title of sjzyz lead to 4 AC there are three of them
Although this results not very good, but this time of the examination process is a new experience, it is worth recording about the (future such cases may be more)
Open question, see Tl, surface stupid question, seemingly simple, the data range 100 000, like a long time from the n- . 3 optimized to n- 2 , is not the first hit, a change of thinking into T2
Look, BFS ah. It seems to be quite water.
But really just looked and did not see a good question, understand meaning of the questions wrong, all bets are off.
After about 45 minutes code, again through the sample 3, pay up.
I think 100 points steady, and then began to look at T3.
A look obvious the n- 2 violence can take 60 minutes.
T3 want 30 minutes, thinking positive solution is more complex, a little idea, but not a system. Temporarily abandoned.
The overall distribution of test scores:
Violence T1 60, T2 can be A, T3 60 violence.
Then 100 + 60 + 60 = 220. The score will not be lower in the rankings (the first 2!)
So I change ideas, to give up the positive solutions T1 and T3, consider how to stabilize 220 points.
First, of course, we require the code T1 and T3 violence out over the sample.
The sample really is dog feces. . . But I thought I was quite right yet
Second, the remaining 60 minutes of course, is to shoot it. (Do not want positive solutions, then the remaining time is really a lot)
I use a very slow speed of the code, the code of a random number generator, code of one pair beat, also waited a long time to debug.
And then I want to come out for a while before T1 of the n- 3 violence, give it a yard out.
And then began to shoot. Then suddenly went wrong! Take a look at the code against his for a long time did not see it.
The data is narrowed to five, hand measurements, only to find the n- 3 violence are wrong!
Change change change. Violence also accounted for the change for some time. Finally change over, continue to shoot.
Wrong again! See table, thieves panic, instantly throw up first violence
Wrong solution as well as 45 minutes of it. . . But I still chuckle on the test I have not staggered solution
还有36分钟，应该能够找出问题。结果是i-j和j-i互换了位置。。。
诶？暴力不是60分吗？为什么多了5分？考后实测：
差在哪里了？
%%%肥人圆脸skyh，骗分乱搞+5分
我还有27分钟。理论上可以闲下来了毕竟什么都干不完的了
但我太闲了所以要找点事干，稍微改造了一下对拍和数据生成器：拍T3！
决策很正确，只有5分可不是闹着玩的。
还没有开始对拍，突然灵光一闪
每次上去讲课是说的头头是道结果自己还在犯。。。
19分钟，不抱什么希望了。但也干不了别的，接着拍吧。
极小点，拍了3个就挂了。吓死我了！
快！来不及了！先交n3暴力！
帅死了。我还以为由发现惊人错误救回来多少分呢。
已经没有心态了，还有14分钟，还能怎么办呢？改我的n2吧。
结果是把1写成了i，我可是真有才。
心率早就上百了，时间那么紧又发现自己有错。。。
呼唉，交上去吧。。。
诶？又是65？考后又测。
%%%大脸大脸不要脸！听他的clock合计骗10分！
放心了。或者说，无法可想了。
还有7分钟，就让它继续拍下去吧。。。我去厕所
考后讲题，稍微看了一下T2的题面，然后我和skyh都是。。。
只是加了一句特判。
那我们还能怎么办呢？已经考后了。
总的来说爆就爆了吧，exp++最重要。
上面废话太多，所以exp简明扼要一些：
对拍很重要！尤其在NOIP等正式比赛上对拍是一定要有的，这是你确保稳分数的重要手段！
及时放弃正解。毕竟不是所有人都像开窍了的LNC一样。
仔细审题。你一整场考试的努力可能还顶不上你多花20分钟重新仔细审题。
要有规划。对自己每场考试每道题的得分要有想法，这样才能更好的进行以后的决策。
每一秒都有希望。不要因为考试刚开始几分钟时间还多就放松，不要因为考试还有几分钟就结束而放弃 。
放低期望。考场上你要承认你的能力水平，考场上贪婪可能就会失去本分（亦指“本来的分数”）
考场上骗分乱搞也是一种能力。很明显瞎输出一个也比全T好，然而并不适用于考后。
还是收获满满的一次考试啊。
顺便声讨不要脸的大脸skyh考后不打正解天天骗AC带起不良风气。