First of all... wow! Amazing. Thanks for posting this.
I think it's not about calculation though. It's judging what's left on the board. I assume he didn't look much deeper than 5 or 6 moves.
So it's the queenside pawns the bishops vs black's rook. How can black use his rook? So you calculate to see if, for example, they can use the d or e files. Or if they can attack your king... since black's king is loose, that's another consideration. I think white is counting on that it will be really hard to move the knights for example. White's bishops exert a lot of pressure.
Not that I would find this move, it's amazing, and I probably still don't get it, but that's the basic process I think. Just looking at what's left on the board, Black's material didn't matter as much as White's. The rooks never did anything, and white's pawns marched up the board.
I am referring to brilliant rook sacrifice Capablanca style in the following game. I am eager to know how much of calculation needs to go in before a player can boldly sacrifice rook at that level of chess. What if Hari had played a different combination or sequence or not accepted the sacrifice.
http://www.2700chess.com/games/kramnik-harikrishna-r4-shamkir-2017-04-24
Thoughts?