Using the MMULT and TRANSPOSE fomulas simplies for the process of calculating portfolio variance as you do not need to build the covariance table outlined in the ‘simple’ method.
The outputs should be exactly the same assuming the inputs are the same.
Also, if you are looking to generate the efficient frontier using VBA, the second method usually is better.
Go ahead and write out the risk equation for 10 assets. The number of terms increases (approximately) with the square of the number of assets. The matrix formulation keeps the terms organized for you so you can work with (almost) any number of assets.
Edit: OK, I went back and looked at how the “simple” method did the sum. That’s a clever way to get the number in excel, and it isn’t too hard to scale up if needed. However, the matrix format is just a lot more flexible… once you know it in excel, you can do the same thing in R or Matlab or a number of other contexts that you might need.