R, C, T = gets.split.map &:to_i
Sy, Sx = gets.split.map &:to_i
Gy, Gx = gets.split.map &:to_i
Board = $<.map{|s|s.chomp.chars}

# 偶奇が合わない
man_dist = ((Sy+Sx) - (Gy+Gx)).abs
if man_dist & 1 != T & 1
  p 0
  exit
end

DIRS = [[1,0],[-1,0],[0,1],[0,-1]]
# スタート地点から動けない
if DIRS.all?{|(dy, dx)| Board[Sy+dy][Sx+dx] == ?# }
  p Sy == Gy && Sx == Gx ? 1 : 0
  exit
end

# 行き止まり
used = R.times.map{ [p]*C }
dfs = ->y, x{
  return if used[y][x] || Board[y][x] == ?#
  used[y][x] = true
  DIRS.each{|(dy,dx)|
    dfs[y+dy, x+dx]
  }
}
dfs[Sy, Sx]
if !used[Gy][Gx]
  p 0
  exit
end

dirss = R.times.map{ [p]*C }
(1...R).each{|y|
  (1...C).each{|x|
    next if Board[y][x] == ?#
    dirss[y][x] = DIRS.select{|(dy,dx)|
      Board[y+dy][x+dx] != ?#
    }
  }
}

probs = R.times.map{ [0r]*C }
probs[Sy][Sx] = 1r
[T, 1000+T%2].min.times{
  next_probs = R.times.map{ [0r]*C }
  (1...R).each{|y|
    (1...C).each{|x|
      next if Board[y][x] == ?#
      prob = probs[y][x]
      dirs = dirss[y][x]
      dirs.each{|(dy, dx)|
        next_probs[y+dy][x+dx] += prob * 1r / dirs.size
      }
    }
  }
  probs = next_probs
}
p probs[Gy][Gx].to_f