1. Standard memberXanthosNZ
    Cancerous Bus Crash
    p^2.sin(phi)
    Joined
    06 Sep '04
    Moves
    25076
    23 Nov '06 05:362 edits
    I'm writing a series of number labels and I need to ensure that the label is not read upside down so I underline any label that could be read upside down and if doing so would give a different number than reading it right way up. As I'm doing so I wonder, how many labels will need the underline?

    Note:
    A number can be read upside down if and only if it consists of only 0, 1, 6, 8 and 9. Leading zeroes are allowed in the upside down versions (so 1000 can be read upside down).

    So how many labels do I have to underline when my series goes from 001 to 999? What about to 9999? Is there a general answer for a series ending at 10^N - 1?

    EDIT: Although I did mention it, it should be properly stated that numbers will be padded with leading zeros to all have N digits (the same N as in the 10^N - 1) and that labels which when read upside down are the same do not need underlines (so if going to 999, 111 doesn't need an underline but 011 does).

    EDIT2: The real thrust of the question will follow once the preliminary answers have been worked out.
  2. Standard memberpawnfondler
    no edits
    Behind you
    Joined
    14 Mar '05
    Moves
    10935
    23 Nov '06 08:14
    so any number in the 200,300,400,500 and 700 sections would be out? as well as any number with 2 3 4 5 and 7 in it. so would i be correct by saying find out how many numbers can be upside down in a single hundred section and multiply by 5? and same would apply for the 000-9999?
  3. Standard memberXanthosNZ
    Cancerous Bus Crash
    p^2.sin(phi)
    Joined
    06 Sep '04
    Moves
    25076
    23 Nov '06 08:35
    Originally posted by pawnfondler
    so any number in the 200,300,400,500 and 700 sections would be out? as well as any number with 2 3 4 5 and 7 in it. so would i be correct by saying find out how many numbers can be upside down in a single hundred section and multiply by 5? and same would apply for the 000-9999?
    You're forgetting that numbers that read the same either way up don't need an underline. The rest of your logic seems fine.
  4. Joined
    28 Nov '05
    Moves
    24334
    23 Nov '06 08:43
    Originally posted by XanthosNZ
    Is there a general answer for a series ending at 10^N - 1?
    probably.
    The number of candidates for underlining is 5^N - 1
    Then we need to weed out those that are symmetrical, like 010, 818, 619 which should be formulaic

    for N=3 symmetricals = (5x3)-1
    for N=4 symmetricals = (5x5)-1
    for N=5 symmetricals = (5x5x3)-1
    for N=6 symmetricals = (5x5x5)-1
    and I guess
    for N=7 symmetricals = (5x5x5x3)-1
    for N=8 symmetricals = (5x5x5x5)-1

    for N=3 underlines = (5^3)-1 - ((5x3)-1) = 124 - 14 = 110
    for N=4 underlines = (5^4)-1 - ((5x5)-1) = 624 - 24 = 600
  5. Joined
    28 Nov '05
    Moves
    24334
    24 Nov '06 13:18
    for N=1 symmetricals = (3)-1
    for N=2 symmetricals = (5)-1

    for N=1 underlines = (5^1)-1 - ((3)-1) = 4 - 2 = 2
    for N=2 underlines = (5^2)-1 - ((5)-1) = 24 - 4 = 20
  6. Joined
    28 Nov '05
    Moves
    24334
    28 Nov '06 11:00
    5^n - ( 5^INT(n/2) x 3^INT((n+1)/2) )
Back to Top