I'm having difficulty figuring out if my application of exponential compounding here is reasonably accurate. Please help if you have a second...:)

Lets say I have $1000 invested in an index fund. Historically, it has earned investors say 6% per annum in growth (including dividend payments).

So I could just plug in these numbers into the exponential compounding formula A=P(1+r/n)^nt, and just kind of estimate that the investments growth is similar to it being compounded annually, since on average this was its annual growth. Is that really the best we can do for a spreadsheet-able formula though?

Using a fixed compounding calculation seems like it may not be the most reasonable, since shares are not necessarily compounding my money precisely "annually". Selecting annual compounding seems a bit arbitrary. These are stocks. They are volatile. It is only compounding when growth is occuring. This could be a small amount daily or it could be a big leap after a few months of negative 'growth'

Is using the 6% historical average in an annual compounding formula reasonable or not? If not how can I improve its accuracy?

Note: This is a question more about the application of the mathematical equation. This question is NOT about the merits of using historical growth for future predictions so please do not focus on that, but feel free to still comment.

Also lets leave inflation and expenses/fees out of the equation for simplicity.