1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Hard Geometry Challenge

  1. Sep 4, 2012 #1
    Well, I found this challenge in another forum (not about math) on the internet, and, originally, there were no 'w', 'z' or 'y' drawn on the pic, it was just "find the x", but I put them on because I know you guys probably would create other variables to solve the problem.
    Another problem I had to deal with is if the Quadrangle is, by fact, a square, I mean, if 'w' plus 'z' really is 1. How the wreck do I solve this?

    Attached Files:

  2. jcsd
  3. Sep 4, 2012 #2


    User Avatar
    Science Advisor

    (y/1) = (1+y)/(1+x) (both are secants of the same angle)

    1 + (1+x)² = (1+y)² (Pthag. theorem)

    I'll let you do the algebra.
  4. Sep 4, 2012 #3
    Could you do (y/1) = (1/x) instead?
  5. Sep 4, 2012 #4

    Yes, in this case x=y[itex]^{-1}[/itex].
  6. Sep 4, 2012 #5
    Just another doubt, how can I know if the biggest triangle (1, 1+x, 1+y) has a 90º base angle? Because it wasn't specified in the problem if there is a square.
  7. Sep 5, 2012 #6


    User Avatar
    Science Advisor

    It is still OK even if is not a square. If it is a rhombus, the triangles involved are similar, so the side ratios {y:1 = (1+y):(1+x)} are equal, even if not the secant of the angle.

    If it is not a rhombus, then I suspect there is no solution.
  8. Sep 7, 2012 #7


    User Avatar
    Science Advisor

    Added note: In case of a rhombus, you need to know the angle, so that you can use the law of cosines to get the secon equation.
  9. Sep 7, 2012 #8
    Ok, by doing the algebra I found the equation on the pic. I don't know if this was right, so I went to Wolfram to solve this.

    Attached Files:

  10. Sep 7, 2012 #9
    No real solutions, so it wasn't right.


    a) w+z = 1
    b) x^2 + z ^2 = 1
    c) w^2 + 1 = y^2
    d) y/1 = 1/x

    use c amd d to get

    e) w^2 + 1 = 1/x^2

    and then a and e to get

    f) z^2 - 2z +2 = 1/x^2

    and finally b and f to get

    g) (z^2 - 2z+2)(1-z^2) = 1

    feeding this to wolfram alpha:


    gets you x ≈ 0.883204
  11. Sep 8, 2012 #10


    User Avatar
    Science Advisor

    Your original equation looks wrong.

    I got: y4 + 2y3 - y2 -2y -1 = 0.

    A quick inspection shows there is a root between 1 and √2, which is expected.
    Last edited: Sep 8, 2012
  12. Sep 8, 2012 #11
    Am I wrong or I can divide this equation by (y-1)?
  13. Sep 9, 2012 #12


    User Avatar
    Science Advisor

    1 is not a root, so you can't divide by y-1.
  14. Sep 9, 2012 #13
    Yeah, sure, I miscounted the coefficients.
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook