SRM 458: Please be the bottom

Okay, so it looks like I'm going to have to reset my expectations here. I am not, it turns out, going to rocket into the top tier on TopCoder within a few months. Instead, it appears that I'm going to sink deeper into the depths—737 after today's match! 80% of people on TopCoder are better than me! The pie is humble.

I thought the easy problem was harder than before. At least, I took a lot longer, and I didn't even have a hope of submitting the medium problem. What is really taking the wind out of my sails this time is that I failed the system test for the easy problem! That is, I passed the few tests they give you as part of the problem, but tripped up on one from the more extensive set of tests that get run after coding is over. My policy has been to submit the problem immediately when the few problem tests pass. If this happens often, I'll have to rethink that policy.

So what does this mean, in the bigger picture? Well, I really don't think I'm fooling myself when I say that these matches are way harder than the coding problems I got in my Google interview. As I've reflected further on that, I think it was the algorithm analysis part that tripped me up more than the coding. That interview was the first time that I actually did asymptotic analysis out loud with another person, and I cringe more when I remember that part of the interview than the coding part.

So I don't think I have to be first on TopCoder in order to be a good fit for Google, but I do feel like I have to keep plugging away at it this year. I would say that if I could get into Division One by the end of the year, I will feel like I've accomplished something. At the same time, I'll need to find a way to get more comfortable with algorithmic analysis, and let's not forget to get some distributed experience as well.

I wonder, how well would Steve Yegge do on TopCoder?