The statistical term "degree of confidence" is not quantifiable for extrapolation, and will get you flunked from most good statistics programs. Extrapolation is predicting results beyond the spread of the data, like trying to predict how much liquid water would be on earth if the average temperature was 120 degrees. We have no experience in or around those parameters, so we cannot make valid predictions there.
Carbon dating is extrapolation, because we do not have standards going back 50,000 years to calibrate from. While it is a useful theory, it is not proven over that interval. Use it if you want, but the idea that it is a provable, calibrated measuring system is simply not true.
Like 99.93% of politicians and their pollsters do!