Обсуждение @ 1413. Марсопрыг @ Timus Online Judge

Показать все сообщения Спрятать все сообщения

Probably tests are wrong... Burunduk1 29 окт 2005 15:16

I've sent two solutions.
On the test one million of '1'

the first one outputs:
-707106.7811865476 -707106.7811865476

the second one outputs:
-707106.7811865475 -707106.7811865475

Of course both solutions get AC.

Re: Probably tests are wrong... Alexandrov Sergey 1 ноя 2005 03:22

Well, I've got two different solutions for this problem too. BUT!

on the testcase one million of '1'
the first one outputs:
-707106.7811914912

and the second (!)
-11992.5310089238

both AC..
I think the difference is very big...

Re: Probably tests are wrong... Akshin Salimov 9 янв 2006 02:34

Mine outputs exactly the same as
Sergey's -707106.7811914912 BUT I HAVE WA#2! WHY!?

Here is my code:

[code deleted]

Edited by moderator 22.02.2006 22:07

Re: Probably tests are wrong... Victor Barinov (TNU) 9 янв 2006 13:24

I think z must be sqrt(2)/2
And real is not so exact, use double or extended

Re: Probably tests are wrong... botl_ayhan 12 янв 2006 22:32

there is something strange i got WA with cos(45) but when i changen it to sqrt(2)/2 i got AC:

[code deleted]

Edited by moderator 22.02.2006 22:08

cos(45) ??? wwwwww 12 янв 2006 23:49

cos(pi/4) = sqrt(2)/2
but cos(45) is something strange.

Re: cos(45) ??? botl_ayhan 13 янв 2006 09:48

i didn't know this. thank you very much

Re: cos(45) ??? Koryakov Anton_ghost_Anigus 15 фев 2006 22:47

cos(45) - but you must change it into cos(pi/4)- degree must be in radian^)

Обсуждение задачи 1413. Марсопрыг