Type checking in dynamic language

Below is the program written to perform the type check in addition to primary functionality(i.e., square root).

from math import sqrt
def isnumber(thing):
    try:
        int(thing)
    except:
        return False
    return True


def make_safe(f, isnumber):
    def type_check(datum):
        if isnumber(datum):
            return f(datum)
        else:
            return False
    return type_check


safe_sqrt = make_safe(sqrt, isnumber)
print(safe_sqrt('hello'))

If an abstraction is designed for any functionality in dynamic language(like python), there is more chance of runtime type check failures compared to static-type language like java.

In fact python is more strictly typed in runtime compared to java at build time. Python has it's own type hierarchy.

So,

What advantage does a programmer have in using dynamic language when type checking is any ways getting deferred to runtime? Is it not crucial to take control of type checking by a programmer rather than interpreter?