This is really a poorly-worded question. When it says "only at its lowest", or "only when horizontal", or "only at its highest", it effectively means "exactly" at those angles (0, 90, 180, or 270°). There is zero tolerance on "only". Therefore, the wire is infinitely more likely to break within any non-zero range of angle centered around the lowest point than it would be at any exact angle.
It would be like setting up a lottery where, in order to win, you have to guess the "exact" random number between 0 and 1.
Customer: "Well, how many digits is that?"
Seller: "Well it's not a number of digits, you have to get it exactly right."
Customer: "Okay, I'll give it a shot."
Seller: "Do you want to pick your own random number or would you like a quick pick."
Customer: "Just give me a quick pick."
Seller: "This is amazing! I've never seen anyone come as close to winning as you got. You matched the first 1, 579,687 digits of the random number. Unfortunately, you missed the next digit. Want to try again?"
Also, the angle that it breaks depends on the rate of the angular acceleration. If I chose an infinitesimally small acceleration, I could guarantee that the wire would always break at the lowest point. And if you were so generous as to put a tolerance on "lowest", I could even perform the experimental verification is less than infinite time. :)
Sorry to go on so long about this, but this question just kind of bugged me. :)