Assuming that you have a generous tolerance over one bend, the error in bend deduction can be spread out, and a good part still can be produced. The total error, or amount of difference between the top and the bottom entries, was 0.039 in.
Figure 1 shows data from five of those charts selected at random. In reviewing a couple dozen bend deduction charts from a variety of sources, I couldn't find any that agreed with another, with the exception of a single number here and there.
