NavList:

Indeed, I have had lighthearted jousts with the statistics folks before....  Recalling the statement attributed to Mark Twain in 1913, but of earlier origin. Figures don't lie, but liars figure.

Discarding points taken on a data set in research is admittedly frought with hazard. To wit: when supporting or discarding a hypothesis even aberrant data may have significance. Natural human frailty is to discard data that doesn't fit our interpretation of the mean.

The beauty of Celestial sight variation is that we are comparing our data set to an accepted mathematical value.  I hope the statisiticians grant the premise that the nautical almanac is such a value (at least statistically it seems to be).  Mathematics is delightfully arbitrary.  We can discard all points happily based on a standard arbitrary deviation, without fear of concluding some scientific heresy.  This tyrannical act is made plausible by the acceptance that our data set is, in practical terms, non subjective in the presence of an arbitrary definition & slope.

The argument stems from the fact that statistics may used for evaluating unknown patterns, vs against accepted reference standards. A true unknown creates peril of discard values. 

It's like the task of counting elephants. Traverse the world left and right, and compare every animal caught to a known accepted elephant. If your definiton of an elephant is precise, you will end up with a pretty good count of detected elephants. If you used a sloppy definition, such as "really big gray animals", you might end up with a Rhinocerous or two and perhaps some gray whales.  

Precise Definitions (or as close as we can get to them) save the day...A numerical mean is a thing of beauty. Most of us can agree that a 2 = 2. 

The value in this contemplation....statistically... so far...lol..none to speak of...It just proves its a long day and I'm sliding into insanity.

The value in graphing sights I have come to appreciate is: It allows me to evaluate more sights in less time, with greater  accuracy. (and also to discard dock-bound sights that went wrong while talking too much and sipping a cocktail).

So I clutch my statistical teddy bear, and claim the accepted Lat/Lon (.....neglecting continental drift.....) on the concrete pier aside my boat acts as a fixed definiton...by which to um.... statistically verify my statistical methods.....oh geez...