Finally beat Iron Helm on my 5th try with a score of 53.
@sojournTime There's some discussion here:
https://www.reddit.com/r/statistics/comments/8de54s/is_r_better_than_python_at_anything_i_started/
@trinsec It doesn't exist; it's just a title suggestion.
@vwbusguy Scientists aren't sure exactly how eels reproduce.
My mother adds, "and when you find out, don't tell them, because I don't want any more."
@ancientjames ... so it's got at least as much computational power as a pregnancy test.
Hint, not a solution
@Placholdr I'm not sure if this will work, but here's an approach:
Use contraction. Assume a_n *is* a perfect square. Consider all possibilities for the square root's last digit. It can't be even, because squaring an even number gives another even number. That leaves 1, 3, 5, and 9. ***1^2 (that is, squaring a number that ends in 1) ends in 1. ***3^2 ends in 9, ***5^2 ends in 5, and ***9^2 ends in 1. So the proposed square root must end in 1 or 9. Explore those possibilities and see if you can eliminate them.
Alternately, maybe the old algorithm for taking square roots by hand is useful.
Again, these may be dead ends.
@celesteh Doppler Barf is the name of my next band.
Honest question: Is there anyone here fluent in both #rstats and #python who prefers R?
I know Python well and am still learning R. It feels kludgy, extremely context-sensitive, and more like an accreted bag of tricks than a designed programming language.
Maybe this is just my fear of change / computer scientist bias or issues with the particular book I'm reading. A part of me feels that students minoring (and someday majoring) in #DataScience should know R because it's widely used, but another part worries that teaching them this inferior tool is doing them a disservice.
Is there an R champion in the house?
CS professor, game designer, and fire dancer ordinaire.