OJ Board

The answer is just the value of C(n+v-1,v-1)+C(n+v-3,v-1)+C(n+v-5,v-1)+...+C(v or v-1,v-1).
I caculated it directly and my program got ACed in 1.9x(sec). rather slow
But many people's programs used much less time. I wonder how they made it.

It should be "Print a blank line after each case."

To Whinii F. :

Yes, there is a more simple method.

Let Pos[ch(Char),i(Longint)] be the position of i-th ch in S. Because Pos[ch,i]<Pos[ch,i+1], Binary Search can be used here, too.

Let Length(ch) be the length of array Pos[ch]. Obviously, Length('A')+Length('B')+...+Length('Z')+Length('a ...

Yes, use recursion.

I dunno why i've got wa. :(
If the number of different x is smaller than the number of different y, I will turn the plane. Maybe it can make the program more efficient.
[pascal]
Program Counting_Rectangles;

Const
Maxn=5000;

Type
TPoint=Record
x,y:Longint;
End;
TLink=^TNode;
TNode=Record ...

My algorithm is:
1. Since covering the entire strip and covering the upper bound are equivalent, each sprinkler can be represented by an interval.
2. Use GREEDY algorithm to solve it.
Is it correct? I think so, but why did I got WA.
[pascal]
Program Watering_Grass;

Const
Maxn=10000;
Zero=1e-3; {I ...

I got WA at about 10.x sec.
Can anyone tell me why or give me some testcases?
[pascal]
Program Shogi_Tournament;

Const
Maxn=50;
Maxmatch=Maxn*Maxn;

Type
Link=^Node;
Node=Record
data,c,f:Longint;
next,g:Link;
End;

Var
a:Array [1..Maxmatch+Maxn+2] of Link;
res:Array [1..Maxn,1..Maxn,1..2 ...