-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
conclusion -- choose smaller tasks, because mean time of larger tasks is infinite?
-
Mean of all tasks is infinite according to my model :)
- Show replies
New conversation -
-
-
As always, such a nice read. Also very interesting topic, IMO.
I have a pet theory that task distribution/assignment is a submodular function over the set of tuples (what, who, when). Do you have any hunch about this? Or data...Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Oh boy this feels correct. Uncertainty has always been the killer in a project for me. It’d be interesting to see if the PERT distribution gave a better fit yet:https://en.wikipedia.org/wiki/PERT_distribution …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Hey Erik, interesting article. You have a small logical error in your argument, having the median (log) residual be 0, does not imply estimating the median well.
-
For example if I make my estimate by flipping a coin and choosing 1hr or 1,000,000hrs randomly, overestimate and underestimate are equally likely, so median residual will be close to zero. But it's a terrible estimate of the median.
End of conversation
New conversation -
-
-
I'd also add that if there are any non-trivial dependencies between tasks, the "long tail" time consuming tasks will make things *even* worse. This may be partially responsible for the success of "agile" development and "time-boxing" if done right you can cut-off the right tail.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
My initial thought was, is adding up estimates something people do? Then I realised agile tools does if you use estimates in them (also why that sucks), since the idea is it should work. Most with experience know this, but nice to see some in-depth.
-
Argh. Agile tools *do*, ffs.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.