NavList:

TO JK: Thanks for your fuller explanation of your objections to TRF in the MPP of CRF documents. What become clear to me is not so much that TRFs are inadequate from Bayesian point of view but rather that traditional DR plot is inadequate from that viewpoint. Others have contributed to that discussion based on DR plotting vs EP plotting so that we are moving forward in understanding. Good.

Thanks to Greg we have a plot which shows the difference between a traditional DR plot and a plot which continues from an EP. Dutton in section 1120 (14th ED)tells us to draw a lite line from EP so that we see if our uncertainty presents a danger. So really not much has changed. At sea I'm not likely to care about the difference shown in Greg's plot. Note we have LOPs with substantial cuts here and JKs exposition pointed out, as we know, that without having quite independent information in two LOPs we can make an error.

Finally I note that neither plotting scheme demonstrated in Greg's plot satisfies Bayesian theory. By reducing EP to a point we have thrown out the information we have about the probability distribution function. Which would need to be renormalized at each step and need to incorporate uncertainty in a cross track direction for leeway every step of the way.

----------------------------------------------------------------
NavList message boards and member settings: www.fer3.com/NavList
Members may optionally receive posts by email.
To cancel email delivery, send a message to NoMail[at]fer3.com
----------------------------------------------------------------