Applying Vacant Spaces to Distributional Hands

I apply the Principle of Vacant Spaces in bidding to evaluate highly distributional hands:

From a recent bidding poll by Richard Lawson:

http://bridgewinners.com/article/view/bidding-problem-2-llct3h9tmq/?cj=832018#c832018

West
K109632
AJ87643
W
N
E
S
P
2
2NT
4
?

Assume partner's 2N overcall shows 15-17 and 1+ spade stopper.

I divide HCP by 3 to determine cover cards (A/K/Q), so 15/3 = 5 covers.

I assume 1 known cover in spades and 4 unknown covers distributed over the remaining 9 vacant cover spaces (AKQ, AQ, AKQ, KQ).

With 5 more wasted spaces in /, and 4 useful spaces in /, partner should have about 2 useful covers in / (4/9 * 4).

With my 4 losers and partner's 2 covers, I expect our median contract is 5 or 5.

So while my heart says bid 6, and my head says bid 6, my calculation says bid 4N (if available to me as 2 places to play).

I use this method in many different auctions, and am happy with the results to date. But I have no deeper mathematical or simulation support for my approach. Do you think this is mathematically sound and/or practical?