David Hobby wrote: > >> 1 >> 11 >> 21 >> 1211 >> 111221 >> >> Not enough data to extrapolate the rule, but it could be >> >> 2113211 or 3113211 or 211121211 or 221121211 >> >> then >> >> 1221131221 or 1321131221 or 1221112111221 or 12122112111221 >> >> The problem with first two rules would come when the >> algorithm requires a digit greater than 9. So maybe the >> simplest rule is the third or fourth >> >> Maia per Lamai Maru > > You want to give us a hint as to what kind of rule > you think it is? There are of course an infinite number of > possible rules for any finite initial sequence. > Those tests usually require "the simplest rule", meaning those that takes less words to explain. So, for example, in a sequence like: 3, 31, 314, 3141 the "correct" rule is "integer part of Pi times the next power of ten" instead of some convoluted rule that could be devised with these numbers.
For that sequence, the "simplest" rule seems to be: (a) replace a single digit _d_ with _1d_ (b) replace a sequence of n digits _d_ with (n+1) and a sequence of (n-1) digits _d_ with variants for rule (b) to take into account the problem when n = 9. Alberto Monteiro _______________________________________________ http://www.mccmedia.com/mailman/listinfo/brin-l
