YoungForest's blog

This contest was moderately difficult. Because of an issue with the judging program, many people were trapped by the third problem. The test case was corrected after the contest. This is already not the first LeetCode incident.

5083. Occurrences After Bigram

LeetCode opened its first biweekly contest, every Saturday night from 10:30 to 12:30. The goal may be to make it easier for students in Europe to participate. The normal weekly contest is usually in the early morning in Europe. With the duration extended to two hours, they can also set harder problems.

Since I had already taken ByteDance’s summer camp written test from 19:00 to 21:30, and that written test was also very hard, with only 30% of the second of three programming questions passed, the later biweekly contest also went badly. My result was not ideal. I only solved two Easy problems.

When I got up on Sunday, it was already past 11, so I joined the contest 40 minutes late. I solved the first two problems smoothly. I took some detours on the third problem and only solved it after the contest. If there had been enough time, solving the third problem should not have been an issue.

1071. Greatest Common Divisor of Strings

Although the problem quality this week was quite good, the difficulty was not high. It was a contest about speed.
Because the second problem had a relatively long statement, my problem-solving order was 1->3->4->2.

1051. Height Checker

This week’s problems were harder than usual. You could also say they happened to hit my blind spot: DP problems. To be honest, I have not studied DP problems deeply. This contest had especially many DP problems, especially the fourth problem, which can be solved with the classic knapsack problem.

1046. Last Stone Weight

The quality of recent contests has been decent. Even the simplest warm-up problem requires careful thought. There is also a DP problem every time. DP is the type of thing where even after doing many problems, you may still fail to write a new one when you encounter it.
This time I returned to my previous level and dropped to 200+.
For this contest, it probably required finishing the first three problems within 55 minutes to enter the top 200. First, I solved problems relatively slowly. Second, I got an index issue wrong in the DP for the third problem, which caused one penalty. So unfortunately I did not enter the top 200.

1041. Robot Bounded In Circle

My book review

This Sunday was a workday in China, so the number of people participating in LeetCode weekly contest directly dropped by 1,000, which shows how enthusiastic domestic participants are about this contest. Chinese contestants are also generally among the strongest in the world. So this time I ranked 70 and entered the top 200 for the first time. Besides the credit for racing against the clock and AC’ing the last problem before the end, there was also the reason that fewer strong contestants participated.

5051. Valid Boomerang

The quality of this contest continued to improve on top of last week. The corner cases in particular caused me to have 4 incorrect attempts, which means 20 minutes of penalty. But from the leaderboard, everyone seemed to be in a similar state, with many wrong attempts.
The focus of this contest was solving all the problems. It happened to require all 4 problems to enter the top 200. Because my idea for the third problem was wrong, even if I had another half hour, I would still have had difficulty solving it. So I lost convincingly.

1033. Moving Stones Until Consecutive

This week’s problems improved quite a lot in quality compared with the previous few weeks. There were fewer trivial problems, and more algorithmic knowledge was tested. Even the first two Easy problems tested enough programming ability. In this contest, I lost half an hour at the beginning because of a stomachache, so I started late. The last problem was actually close to AC. The overall direction was right: using a Trie. But after submitting in the last 10 minutes, it TLE’d, and there was no time left to modify it. The construction direction of the Trie was reversed. For matching problems, we can go from front to back, or from back to front. For this problem, going from back to front is not only simpler to implement, but also better in time (even if the worst-case time complexity is the same).

1030. Matrix Cells in Distance Order