Absolutely! Accuracy in estimates is overrated. If estimates are accurate it’s not an estimate anymore
That’s why i am suggesting story estimates are good enough even for iteration monitoring and tracking. In my opinion the best tool to monitor progress through an iteration is a physical story wall. The wall can still include tasks for stories, they are just not estimated. The movement of cards on the wall will tell everyone if an iteration is doing well or not. How much undone work is on the left side of the wall? How much completed task are there on the right side? How many tasks are hanging somewhere in the middle and why? How many really completed stories by passing all acceptance tests are on the far right sight? The wall tells the team much more than a task burn down chart. At the end of an iteration all that counts are completed stories.
By far the worst thing about task estimation and having the culture of updating them everyday is effectively introducing micromanagement. Unless it’s driven by the team itself to improve selecting the right capacity for iterations. Otherwise it’s a direct expression of mistrust into the capabilities of the team self organizing their own work.
Re-estimating every day also strips away the possibility to use task estimates for iteration capacity. The concept of velocity does not work when re-estimating partially completed tasks or stories.

Cheers
Kai

IMHO, the estimates of task time are most useful for monitoring the progress through the sprint. The estimates are updated and charted daily to reflect the team’s belief on how much time is remaining for the sprint. This number should be going down as days pass, but if it is not, this is sign of an impediment which must identified and resolved.

Without tasks and task estimates, monitoring progress through the sprint becomes guesswork — been there, done that, thank you, not again.

I do think accuracy in estimating is overrated. Estimates should give us a ball park figure for time and costs. Hitting the target on time and budget is a matter of staying focused and keeping your eye on the ball of making a great product within the resources available.

Cheers
Peter

From my opinion there are two situations where task estimates are still helpful:
- teams new to agile and inexperienced with relative estimates and INVEST. Once the team has a thorough understanding of these principles task estimation becomes more and more obsolete
- long iterations like the standard 30 days Scrum. Some teams struggle to plan for such long iterations thus task estimation can give the team a better understanding of capacity. An alternative is to introduce two week iterations. It is easier to look just two weeks ahead and instead of wasting time in long iteration planning meetings the time is better spend on delivering working software. Even one week iterations work but i am more comfortable with two week iterations.

I just like to emphasize that, whether you do task estimation or not, story estimation and story tracking is much more important. I have seen teams which solely relied on task estimation and had no visibility of project progress at all.

see http://blog.mountaingoatsoftware.com/?p=15

Stefan