Politicians like to get elected. They will do just about anything to get elected. Including changing the rules of the voting: who can vote, where you can vote, when you can vote, etc. One very common practice is called gerrymandering, where the boundaries of "ridings" are redrawn to favour a particular candidate (the one doing the redrawing, of course).

Your task is to help determine how to do some simple, linear gerrymandering.

You will be given the information about N ridings (2 ≤ N ≤ 1000) where there are candidates from p (2 ≤ p ≤ 10) different parties. These N ridings are linear, in the sense that they are side-by-side; there are two ridings (on the ends) that have only one adjacent riding, with the rest of the ridings having two adjacent ridings. A picture is shown below for N = 4 and p = 2 (which is also the sample data):

Note that Riding 1 and Riding 2 are adjacent, Riding 2 and 3 are adjacent, Riding 3 and 4 are adjacent. No other ridings are adjacent.

You have some financial backing that will let you bribe the people in charge of setting the boundaries of ridings: in particular, there is a fixed rate to merge two adjacent boundaries. When you merge two ridings, the votes of the ridings merge together, in the sense that the number of votes of party 1 is the sum of the votes of party 1 in each riding, and likewise for all other parties.

Your task is to merge the minimum number of regions such that the first party (your party!) has a majority of the ridings. Note that to win a riding, the party must have more votes than any other party in that riding. Also note that to have a majority of ridings, if there are Q ridings (where Q ≤ N), then your party has won at least ⌊ Q ⁄ 2 ⌋ + 1 of the ridings.