It looks like the function you link to is using an approximation to the Normal CDF. If Excel is using an exact calculation, the results will of course be a little different, especially at the extremes; but they shouldn't be a lot different. Here's my version of the same function, which appears to be using a different approximation. Maybe it works better for you.
If you can post a sample file, that would help us better see what you're seeing.
The StDev function does not need another parameter. It's also available as a summary field type, if you want to be consistent with your cAverage field.