Optimal value is outside range when doing bruteforce optimization using Scipy

I'm following the example given in scipy's optimize documentation to do brute-force optimization on a function with 3 parameters.

This is the function I wish to optimize:

def entropy_of_hyperplane(z):
    w0, w1, b = z
    # do some calculations..
    return -1 * ent

To find the global minimum of this function using grid-serach, I do:

def find_max_entropy_hyperplane(X):
    # grid-search on all the possible w in R^2 , b in R
    # and calculate the entropy of each

    rranges = (slice(0, 2, 0.05), slice(0, 2, 0.05), slice(-4, 4, 0.25)) # ranges for w0, w1, b

    resbrute = optimize.brute(entropy_of_hyperplane, rranges, full_output=True,
                          finish=optimize.fmin)

    w0, w1, b = resbrute[0]
    print(w0, w1, b)

And the result is: w0 = 1.56; w1 = 1.1; b = 1842.73906893

What is going on here? How is the optimal value of b outside the range I specified? I can't understand this behavior. Any help appreciated!