Computed score priority

[Andrey Tkachuk (MLO)]

Here is the place to continue discussion of changes in computed score
priority algorithm started in this thread:

http://groups.google.com/group/myLifeOrganized-ppc/browse_thread/thread/79d1599c17443ef1

Bob Pankratz (ratz) the author of the algorithm made some fixes after
your comments. Here you can find new beta with his changes:
http://www.mylifeorganized.net/downloads/files/beta/mlo_beta4.zip

I can not answer questions regarding these calculations so please
address them to Bob :)

Andrey.

[dbcoyer]

Hmm. Perhaps I don't fully understand the implications of date based
urgency...

Simple Example:

Item A has no date set and importance of 1.5 and urgency of 1. Because
it is deep in an outline (1.2.3.4.5.A) and item 2 has priority 2 an
urgency 1, I know some tweaking goes on to make sure that the depth of
the outline doesn't artificially inflate the importance/urgency. So, on
the statistics for item A, the importance is 1.0000 and the urgency is
.33333 and the IMP & URGENCY is 1.33333.

Now, Item B has a due date of 2 days ago with IMP AND URGENCY both 1.
It is in a tree of 1.2x.3x.4x.B. 2x has importance of 1.5 and urgency
of 1. B's statistics show IMP=.7346 URG=.5679 and IMP & URG=1.0679.
Shouldn't B's IMP & URG = 1.3025?? Why isn't B's IMP & URGENCY a sum of
IMP & URGENCY the way A's is?
Granted, this will not affect the placement of B in relation to A in
this example, but if I modified 3x's importance to 1.5, this should
case B to move ahead of A. But it doesn't because the new value for IMP
& URG of B is only 1.3182

What am I missing?

--David

[dbcoyer]

Forgot to mention, I am using the new beta 4 posted earlier...

[Chris]

Bob: I kinda get the why for the calculation method, except for the
way urgency and importance are combined together. It seems that
treating them as vectors would yield a more useful weighting. Kinda
like Covey's quadrant model but not ass-backwards for the X-axis.

Using 0-2 ranges, you could use something like this:
SQRT(((0-Urgency)^2)+((0-Importance)^2))

Here's a graphic that illustrates the difference:
http://static.flickr.com/25/135993298_3ac2494645_o.jpg

Whatcha think?

--
Chris

[scoobie]

What is the difference between this beta and the other one that has
been posted recently?

[Mark Levison]

It would be nice if the Computed score etc in the task statistics updated as I made changes to the urgency. That would allow me to gauge whether my changes are sufficient.

Mark

[Bob Pankratz]

The combining of urgency and importance was changed with the latest
beta and the shift to .01-2 as the range. Sqrt is a bit heavy of a
computation give that the algorithm is recursive; and would be slow on
even a new ppc and hard on the battery.

The major and only change to the published forumals in the help is that
it's now:

PriorityByImportance := (mi + dw) /3;
PriorityByUrgency := (mu + dw) /3;
PriorityByBoth := (mi + mu + dw) / 3;

and the ranges were change to 0.01-2. for importance and urgency.
before they where 0-1 and 1-2 respectively.

[Bob Pankratz]

You have to go to the todo list; and move off the item before it's
stats are updated did you do that?

See the forumlas I posted else where in the thread.... yes p=(i+u +
d)/3

[Bob Pankratz]

Andrey once upon a time told me why that would make the program a lot
slow. I don't remember the explanation; but I do remember being
convinced and dropped the subject.

[Bob Pankratz]

Oh and I should also mention that the fomula what changed with three
goals

1) Give the users a normal default priority in the middle of the slider
range
2) Generate lists of similar order to the current version
3) Minimize the programming impact so as not to slow down the ppc sync
development.

So you get the changes as they are.

The number that it spits out for the priority is different but the
relative order of the list for outlines less than 5 levels deep is the
same.

If you go deeper than 5 levels; importances of the high side of the
slider will gernerate new and interesting results. Once you comprhrend
them you see they are a result of goal (1) and are desirable; you just
need to learn to use them.... :)

The orginal change to the algorithm was huge; then interesting I
factored the changes through the whole alogorithm and found out that we
could make the simple change posted and get the same results. So
that's the direction I went.

[Chris]

Wouldn't you just have to compute the score one at a time when the
task is created or changed? ...then the weighting would take place as
normal.

--
Chris

[Bob Pankratz]

It's recursive up and down the tree; so a change on any node cascades
to all children.

[RHersh...@gmail.com]

Can you explain in laymen's terms how the computed priority should be
used? I am very confused (i.e. difference between importance and
urgency etc.)

[Richard Musselwhite]

There are others on the list who understand the differences between
urgency and importance much better than I do, but here are three ways
that I think about it:

1) Importance is a measure of how significant a task is to me or how
essential it is to one of my core values or commitments. Urgency is a
measure of how time-sensitive a task is.

2) Importance is the priority that I place on a task. Urgency is the
priority that others place on my tasks.

3) Moving the importance slider will change the position of a task on
the to-do list within the limits set by the importance of its top-level
task in its hierarchy. Urgency will elevate the position of a task on
the to-do list even higher than those limits. So if a task is buried
deep in a hierchy with a fairly low importance, it will appear low on
the to-do list unless you set its urgency quite high, in which case it
can appear even higher on the to-do list than other tasks of greater
importance.

Again, there are others on this list who know a lot more about this
stuff than I do, and I hope they chime in so I can learn more from them
as well.

So the best advice I can give is to play with the sliders and see how
they affect the order of tasks on your to-do list.

Yours,

-RichardM