Wednesday, September 18, 2013

SRM 591 Recap and editorial

So, another SRM, another editorial.

This was another bad day. Well, I was trying to solve the div1 500 for most of the match, but it seemed quite tricky. I opened div1 275 with 10 minutes left. I took a while to code it. Without the 10 minutes constraint I would have submitted this problem. I got tired of the lame strategy. I am going to try something else next match, but opening the problems in the normal order.

When writing the editorial, I learned about div1 900. This SRM had very odd choices for problem scores and slots. I think div1 900 was at least as easy if not easier than div1 medium. Actually, I wonder what would have happened if I opened it during the match. It is my style of problem, bitmask dp when filling a board. Well, I guess I would have probably taken more than 75 minutes to solve it...

The other very odd choice for this SRM was the div2 hard. I actually wasted more time trying to come up with a solution to this problem than I spent understanding the div1 hard solution. That is odd. The thing that made it complicated is that I really didn't expect a solution for a division 2 hard to be so hacky. Tricks to save memory in dp belong to div1 medium. I think even the div1 easy would have felt less out of place for the div2 hard slot.

I think that admins should go back to enforcing there to be at least one shared problem between the divisions. Making div2 medium and div1 easy the same problem was a good way to constraint problem setters to stop them from making the divisions too different and also limited the difficulty of div1 250 which is a serious slippery slope without safeguards. Also, there was nothing that made me feel the division 2 and division 1 versions of this contest were part of the same contest. Part of the experience should be to have red coders and greens talk about the same problems after the end of the match, this cannot happen when the divisions are so different. Also, it is quite easier to write explanations for 5 problems instead of 6 :)


Regarding the editorial itself, I suspect the explanation for div1 275 really, really sucks.

1 comment :

PinkyPal said...

Thanks! This is definitely a handy script to have.

Here is an updated version that should be platform independent. :)