Why would \(\displaystyle -\sqrt{810,000}= -900\) "make no sense"? The problem is asking you how many bits would be required to save "-900" in memory if "each digit requires 4 bit of memory and the sign requires 1 bit". 900 has 3 digits so that would require 3*4= 12 bits and the additional one bit for the sign brings it to 13 bits.

(9, decimal, is 1001, binary, so using the bit 1 for negative, the 13 bits would be 1 1001 0000 0000.)