Agile and numbers

In his recent post, Seth Godin writes:

When you measure the wrong thing, you get the wrong thing. Perhaps you can be precise in your measurement, but precision is not significance.

On the other hand, when you are able to expose your work and your process to the right thing, to the metric that actually matters, good things happen.

We need to spend more time figuring out what to keep track of, and less time actually obsessing over the numbers that we are already measuring.

This is the essential part of the "agile development".
In my current team we spend lots of time on "Backlog grooming sessions" arguing on numbers, getting asked "why is it 5 not 8?"
On the same time, nobody cares on getting the Burndown Chart to look like burn down. The reason is "We have many important and urgent tasks with vague deadlines emerging meanwhile in the sprint."
So on our "Sprint planning" meeting we are basically putting the "precisely measured" tickets to a heap of immeasurable ones. What we get is precisely immeasurable. Like dropping 10 carat diamonds to a bucket of soil and trying to estimate the bucket weight.
What precision we might get?

Moreover, the measurement does not influence our actions in any way.

Our the most precise scale of estimations probably should have 3 values:
- "It's OK to try to do this now"
- "Better not to try now"
- "We don't know"

Having the priorities in place, this simple scale will give us enough information to plan sprints, and dramatically reduce time spent on estimation and planning.