I believe that the e6 sacrifice is rather thematic sacrifice in the Najdorf; for example, in this game, interesting but between a Fritz8 centaur and Fritz8 (both banned, of course).
Game 2072053 Analyzing with the aid of Fritz8 analysis indicated that the 10…Bb7 and 10…g6 lines are refuted, but 10…Nf6 holds. For example: 10…Nf6 11.Be3 d5 12.f5 Be7 and Black intending to castle; the sac option isn’t there – as in this travesty of a game:
Game 1754562. 🙄
I recall reading once that the Nsac is good when white may gain three pawns for the piece, and maintain an attack; however, I do not know such Najdorf positions offhand to add credibility to this general assertion.
Tal-Botvinnik WC match 1960, Classical CK…5. Ne2 is new to me, which isn’t saying much, although I recall the use of the move in this CK
Game 1397656
http://www.chessgames.com/perl/chessgame?gid=1032520
Ivanchuk vs Sergey Volkov (French gone weird) 🙂
http://www.chessgames.com/perl/chessgame?gid=1361556
Tal-Suetin (Sicilian Kan)
http://www.chessgames.com/perl/chessgame?gid=1134883
Tal sometimes did not play the technically best move in order to induce complications [obviously it appears he always aims for these]. I believe I read in an on-line article that Karpov claimed that he occasionally didn’t always play the best move; undoubtedly, in some positions, technically worse moves may give an opponent more chances to go wrong - many gambits and sacrifices, for instance.