While I agree the probability is low as compared to how the gloom-and-doomer portray it, I can immediately see a few major issues with your analysis.
1) The CME doesn't have to directly hit the Earth since disrupting the magnetosphere
, which is many times the size of just the Earth, is what would be required.
2) I don't believe CMEs are uniform in the direction they occur since they are created by anomalies in the Sun's magnetic field, which like the Earth's, has poles. I could not however readily find any breakdown about distribution versus latitude
3) Your caveat is a big one. Your analysis is treating the CME as if it is a single point in space, equivalent to if the Sun fired a bullet at the Earth. The reality is, as you mentioned, CMEs have width, breadth, and height, and these dimensions are big. A CME may be many times the size of the Earth. CMEsalso spread out as they travel the 1 AU it takes to get here. That last part is both good and bad, since the original strength of the CME at the Sun would devastate the Earth, while the greatly weakened version that reaches this far could at worse cause havoc, not devastation.
In short, the Earth has been flying around this neighborhood for a few billion years, including hosting animal life for a good chunk of that, and so far we haven't seen any CME calamities. The game changer is of course our use of satellites and long haul electrical lines which are prone to disruption or damage from a strong CME, but based on the number of known events, the odds of a massive CME causes widespread damage is very low, though not as low as you calculated (0.0028% in 100 years). There may be a handful of CMEs a year that the Sun puts out that if they were to hit Earth could break things, as you pointed out the Earth is a small target in a very large shooting ranges. If I had to guess based on known statistics, a major ground-based disruption will probably happen about once every 100 years. (reference solar storms of 1859 and of 1989)