[seqfan] yet more primes

[Chuck Seggelin]

> I'm convinced that eventually I should be able to find a prime tree of
depth
> 16, making it deeper than the record holder Cunningham Chain.

3n+/-531440 at 84119351 (Depth: 16, Population: 18)

[84119351]
   - = [251826613]
       - = [754948399]
           + = [2265376637]
               + = [6796661351]
                   + = [20390515493]
                       - = [61171015039]
                           + = [183513576557]
                               + = [550541261111]
                                   - = [1651623251893]
                                   + = [1651624314773]
                                       - = [4954872412879]
                                       + = [4954873475759]
                                           - = [14864619895837]
                                               + = [44593860218951]
                                                   + = [133781581188293]
                                                       - = [401344743033439]
                                                           - =
[1204034228568877]

That's one DEEP tree! :)  The offset (531440) is 3^12-1.  I've found that
3n+/-(3^12-1) appears to be highly generative.  I'm seeing many more prime
trees out of this formula than out of hundreds of others I've tested.  I
don't understand why it is so generative, but its high generation rate made
it a good target for pushing out into larger root primes in search of a tree
with depth>15.

                -- Chuck Seggelin