NavList:

You ask whether there is any kind of mathematical approach for this.

The answer is yes!

The most interesting part of solving this is coming up with a likelihood-of-error function. That is: what is the likelihood that your measurement is off by one minute? by two minutes? by 2.34 minutes? Etc. This is a probability density function. Assuming that this is a constant function (that is, on every sight you are equally likely to err by +1.5'), which seems reasonable, you can then take the map and for each point on it assign it a likelihood based on the product of the error functions for each sight. So, for instance, a particular point on the map might be associated with a +0.3' error in one sight, a -2.4' error in a second sight, and a -0.6' error in a third sight. Then (probably quantizing your density function to tenth-of-minute bins) you might find that that's a 0.2 * 0.03 * 0.1 = 0.0006 likelihood.  Then, take the point with the maximum likelihood probability. Done!

Your question might have been: is there an effective technique for this that can be performed in reasonable time by a navigator with a calculator, a pencil, and a piece of paper. That's a different question. Which I have no useful information on. smiley!

John Clements