NavList:

Dave, you wrote:

It depends where you take the bearings from and how you plot them.  I don’t want to get into any discussions on statistics, because that’s above my head, but with 12.5 degrees constant error, what you suggest would work if you picked three points roughly 120 degrees apart and plotted the fix using a station pointer or piece of tracing paper.  However, if you pick three points on one side of the vessel, take a bearing on each and plot the reciprocals, you’ll get a cocked hat which the vessel lies outside of.  It’s the classic example used at basic nav-school to show the advantages of three 120 deree cut radio/visual PLs c.f. three 60 degree cut radio/visual PLs.

 I think I probably misunderstood Andres' problem, but I was commenting on a fix by horizontal angles, rather than by intersecting compass bearings.  Horizontal angle fixes are usually associated with the use of a sextant, but a pelorus, or a compass with a constant, unknown error will also do the job.  Of course, a satisfactory choice of objects for a good angle of cut between the circular LOPs is a good idea.

John