[seqfan] Re: phi-antipalindromic numbers

[Hans Havermann]

Douglas McNeil:

> FWIW, I find 25983 such phi-antipalindromic numbers less than 10^6,
> and I agree with Hans' half-sequence. They have a cute pattern,
> probably related to the observations in the comments.

I see that Neil has added my two bisections of A178482 (A171070,  
A171071). Differences of adjacent terms of either bisection are terms  
of A107328 (which appears to be A065034 with an initial 3 added) and I  
have used that fact to translate the "cute" pattern into code that  
will calculate term #n of A171071 fairly quickly - without first  
calculating any of the previous terms. Assuming that I've done it  
correctly,

term #1 = 3
term #10 = 54
term #100 = 1183
term #1000 = 24324
term #10000 = 771278
term #100000 = 17680149
term #1000000 = 364243723
term #10000000 = 11583057054
term #100000000 = 238175057907
term #1000000000 = 5402072437445
term #10000000000 = 173192141696518
term #100000000000 = 3578032635908428
term #1000000000000 = 81149986784320483
term #10000000000000 = 2600997386890746294
term #100000000000000 = 53482527632200745828
term #1000000000000000 = 1213745387417082498749

term #10^100 =  
16243248183500325534450089584837853567324809329361973207684488292891935207090205003873914101783289817252931340703388640057477985409132564969