Somewhere, somewhen, Arlo Belshee suggested not paying attention to story cost, only to value. Since story value probably varies from some large negative number to some large positive number, value is certainly very important.

However, cost matters also. Suppose without loss of generality that your highest value stories have value 100.  Suppose that stories vary in cost between 1 and 3. And suppose we have time to do 30 points of work.

We can do 10 3-point stories and get value 1000. Or we can do 30 1-point stories and get value 3000.  

3000 would be more. Story cost does matter. 

Left to the reader ...

Suppose that, like most organizations, you don’t really know much about the value of your stories. Should you do high-cost ones, low-cost ones, or what?