Another newbie question-this one about routing

I am using Garmin's nRoute software, which as I understand it draws on
MapSource for its routing.

I have noticed what seems to me like a strange anomaly in the
production of routes. I asked for a route from my home to a certain
golf course, with a certain motel as a way point. (The direct
route.) I then set separately routes from my home to the motel and
from the motel to the golf course (the segmented routes).

When I compared the two routings, I noticed that the direct route was
not the same as the sum of the segmented routes. The direct route
took me to the motel, then put me on the road to the golf course. The
segmented route put me initially on the same road to the golf course.
But at the very end the two routes diverged; and while the divergence
was, essentially, going up opposite sides of a rectangle, nRoute
computed the trip as part of the direct route as requiring seven
seconds less than the segmented route from the hotel to the golf
course.

Now, seven seconds is not going to change my life; but I can't
understand why, both routes having been set for "fastest time", the
segmented routing would compute a route seven seconds slower than a
clearly available alternative.

The anomaly was more significant when I looked at the longer route
from my home to the motel. In this case again different routing was
chosen although in both cases I asked for the fastest route. This
time the segmented route was seven minutes shorter in time than the
direct route; and that is enough to notice.

Why would routing software select two different routes from the same
point of origin to the same waypoint/destination when fastest has been
requested in both cases? And when planning a trip, must I put in a
variety of routes segmented differently and then compare them to get
the optimal routing (of course, this process would have taken longer
than the seven minutes the longer computed route would have cost me)?

There is in fact a difference between the specifications for the two
routes; one has some intermediate waypoints listed, along the portion
of the route that is identical in both calculations. I've found by
experimenting that inclusion of just one intermediate waypoint,
whether 400 miles from the destination or 50, results in the more time-
consuming route being computed. I haven't done the experiment, but
think this may have happened with another route, which changed after I
inserted an intermediate waypoint. But why?

And the use of intermediate waypoints was not involved in the slightly
different routes between the motel and the golf course.

Gary Williams

williamsg4713@juno.com wrote:
I have also noticed some serious anomalies in routing. It is a tough
problem and far from solved. I get some truly ridiculous routings if
divided hiways are involved.

Ted

Ted Edwards wrote:
Ted and Bill;

I don't have a definitive answer to your question but most likely the
difference is a result of round in the math computations. They
accumulate differently.

Dave N