I looked at the atwood article, which really didn't give useful numbers, except the one citation. I opened that, found the fit number, converted to ~ .5 errors

I looked at the atwood article, which really didn't give useful numbers, except the one citation. I opened that, found the fit number, converted to ~ .5 errors per year and thought, eh, skipping ecc is fine. I didn't notice this: "From the graph above, we can see that a fault can easily cause hundreds or thousands of per month." Now I want ecc again.