Go with your original attribution. Bull does say 4?, so presumably the example he examined was fairly dire. Don't lose any sleep over the number of strings as they can easily get filled. My 1697B halfcrown for example has 1 & 1/2 harp strings, and I'm absolutely certain the die didn't start out with that number.
Another reason for not getting too attached to the number of harp strings is that Jackson-Kent did a 10 year study from 1955-1965 on the harp strings of William III coins and concluded at the end of 10 years that you couldn't conclude anything. Just take the number of strings indicated by Bull to be what he saw and nothing more. He didn't see every coin, so all options are available.