hi there, we're planning to make a video about tracing the efficient frontier, but that'll be coming later on. the best thing to do is to turn notifications on so you'd know when it comes out.
The short answer is "yes". Especially if you're sample is large using var.p vs var.s would have little practical impact on your results. The reason to focus on var.s is that it gives an unbiased estimator if you're interested in forecasting variance.
Hi there, it depends on your objectives. For example, if you'd like to compute portfolio risk & return to evaluate performance over a specific period, you'd use daily returns covering that entire period. Or, if it's for forecasting future returns, you'd typically need at least one year's worth of daily data.
Daily returns are typically small. So, starting your sample period one day before or after would normally make very little practical difference in your results. Ideally, you'd like to use the price you paid as the starting point. So, if you bought your shares, say, on Jan 1 you can use either the closing price on that day or the exact price you paid during that day as your starting point.