EXECUTIVE SUMMARY:   
   - We extend a s/w that correlates day by day planetary parameters with   
     (lagged) daily UFO sightings.   
   - We use the s/w to matches patterns found in the observations against   
     a 2nd model of a "UFO fleet" operating across the solar system   
     according to simple rules. All the simplest variations on the theme   
     were pre-computed to match against the actual correlations we observed.   
   - Looking at UFO types "all", "Lights" and "Non Lights" we find the   
     simplest apparent set of assumptions that matches the observations   
     is: the probability of a flight from A to B depends on the present   
     distance between A and B; most UFO's originate nr Saturn; "Light" UFO's   
     seem to originate nr Neptune (and maybe even the Sun, the s/w finds);   
     Non Light UFO's originate mostly from Saturn with some from Neptune.   
   - Other origins can't be eliminated because only the simplest set of   
     scenarios were used in this study. Even more slightly complex   
     assumptions are likely to produce different results given the   
     chaotic and inter-correlated nature of planetary movements.   
      
      
      
   We've seen in a previous post how the positions of the planets seem to   
   correlate highly with day by day UFO sightings across N America.  In   
   particular we found the distance between Saturn and the Earth seemed   
   to explain a big chunk of sightings and other data showed it was not   
   likely a significant number of UFO sightings were just a matter of   
   confusing Saturn with a "real" UFO.  (E.g. the same patterns were seen   
   for day/night sightings and sightings of objects that don't seem to be   
   confusable with a planet).   
      
   To extend that work I've added a more planetary parameters to the   
   mix. The apparently position of each planet -- declination and right   
   ascension -- boost the list to 13 parameters for each "normal"   
   planet, and 15 in the case of Saturn with its rings.  We can now   
   correlate this set of time series against UFO data for various types   
   of object to see how well each planetary param can predict future   
   sightings. It's assumed a lag between the change of a planetary   
   position "now" and UFO activity in N days might indicate "something"   
   is travelling between that planet and Earth and taking maybe around N   
   days to do it. In this way we found last time a lower bound on   
   putative UFO movements between planets was substantially sub-FTL.   
      
   Working with an AI s/w has been overall a boon to this work. While a   
   complex program spitting out results can be tedious to check and   
   almost impossible to debug, it can also help in its own   
   development. In this case the s/w, which uses some standard stats   
   tools to do some of the heavy lifting, discovered one package gave   
   unusual results in some cases. This is often the case when packages   
   are pushed (by statisticians) further than their designers envisioned.   
   But a "problem" with AI's is they tend to push whatever tools they are   
   allowed to use *well* past design limits. They have generally no   
   concepts how a package is typically used or what kind of data people   
   normally give it; they just use it if, when, and how they see fit.   
   And sometimes the package obviously breaks or spits out an answer that   
   is wrong but without printing a warning message with it the AI might   
   pick up on.   
      
   After much nail biting the problem in question was tracked down and   
   fixed. The up side is the stats s/w now outputs numbers with a known   
   and hand-checked (:) error bound -- generally +- 5%.  If 2 R2   
   correlations (normally I use the R2 statistic from various time series   
   regressions) differ by say 10% from their common average then the   
   larger one is "most likely" really the larger. If they are closer than   
   10% then the ordering can't be unambiguously determined.   
      
   So the AI can now "confidently" print out a list of correlations of   
   each UFO type (shape/color) against day by day (and hour by hour if   
   necessary) planetary movements at the time and order them from largest   
   to smallest.  The larger correlations are then the "most likely" ones   
   to be "real" and not due to some luck-of-the-draw in the data.  And   
   the order of the correlations show e.g. which planets are "more   
   involved" in the particular kind of UFO activity, and which are less   
   involved.   
      
   For example. We can correlate the daily UFO sighting counts against   
   planetary movements from an ephemeris s/w and find the table:   
      
   Planet	Parm	R2   
   		(in TS regr of planet param predicting   
   		daily UFO sightings)   
   mercury	Dec	0.56812030   
   neptune	rg	0.50386267   
   venus	Dec	0.47351440   
   uranus	rg	0.45647785   
   saturn	rg	0.43902437   
   venus	RA	0.41095631   
   jupiter	Dec	0.27301541   
   pluto	rg	0.23986765   
   mars	RA	0.21302930   
   Dec == declination (celestial latitude) of planet in degrees   
   RA == right ascension (celestial longitude)   
   rg == distance between planet and Earth in AU   
      
   The table shows the "most involved" param with daily UFO sightings in   
   total is the declination of Mercury.  The next most correlated is the   
   "geocentric distance" to Neptune.  (Note the diff in R2 between 1st   
   and 2nd is slightly more than 10% of either).  Etc.   
      
   And now the complication. The AI has warns me already that we cant   
   just assume most UFO's are coming from Mercury because of the first R2   
   in the list.  They may be coming VIA Mercury, for example.  And in the   
   latest twist the AI s/w has proved to itself and me the correlation   
   between Mercury's declination and UFO sightings may be "induced" by   
   something else because the movement of Mercury is synchronized with   
   other planets in the solar system. The distance between planets and   
   their periods are not random numbers. They are correlated by Bode and Kepler!   
      
   Up to recently I had assumed planetary parameters were "more or less"   
   statistically independent, close enough. But it turns out some are   
   way way NOT independent of others.   
      
   It is therefore necessary to get the AI to look at the pattern of   
   correlations and decide which set of simple assumptions would results   
   in numbers than most look like the correlations we found.  We need the   
   s/w to essentially "get inside the head of" UFO captains and simulate   
   flying between planets under different kinds of assumptions, and pick   
   out which set produces results most like the correlations we have   
   above. Only in SOME cases will the largest correlation straight out   
   point at the most interesting planet.  Sometimes the real interest   
   will be lower down on the list or maybe be "hidden" and not appear on   
   the list at all.   
      
   To that end I quickly made another model that takes the planetary   
   parameters output from the ephemeris s/w, takes a set of assumptions   
   and strategies a UFO captain might use to decide when and where to fly   
   from their current location, and collect the correlations as we have   
   measured above for "real UFO sightings", and determine which strategy   
   looks to be closest to those observations.   
      
   While there is no guarantee we will find the *actual* strategy used   
   to fly between different planets of our solar system, we will likely   
   end up with *a* simple set of assumptions that produce a similar   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)