[skimdesk]

Consider a sentence like "We will finish project x next week with 75% probability".

I'm not sure it is any more clear than "It is probable that we will finish project x next week".

The point estimate is of little value unless we know how it was made. In this case "Jim just came up with a number".

Instead of talking of probability, an estimate of the number of days/hours/tasks would seem more grounded and easier to come up with.

In this case I'd go for, "We will finish project x after we have completed four days worth of work".

[necovek]

That certainly helps in eliminating some hard-to-control-for variables (like how much work time you will really have, as in what things will pop up that need urgent attention), and it also gives you "error bars" by default (4 days means 3-5 days), but lots of projects involve a bunch of unknowns, making them impossible to estimate with any sort of precision before you actually dive into them.

As with anything, estimations should be iteratively improved, but usually nobody bothers ;)