Neoclassical Chess Variation Feature Request

Recently I discovered this chess variation, that is just chess but beginning after the third move. The first moves are chosen probabilistically from a random distribution of the masters games. In this way, opening theory, ie. memorization, loses importance in favor of pure chess ability.

In would be interesting to have this modality available as a new variation in lichess. I think that there is no chess portal implementing this variation.

Web describing the details:
http://neoclassicalchess.com/

Paper with technical details:
http://neoclassicalchess.com/wp-content/uploads/2016/06/Paper-Neoclassical-Chess-June-1-2016.pdf

Nigel Short on Neoclassical Chess:
twitter.com/nigelshortchess/status/739084175748280320

I like this idea. Also seems pretty simple to implement.

Related (though perhaps somewhat difficult to implement) is "Bronstein chess", via http://www.quantumgambitz.com/blog/chess/cga/bronstein-chess-pre-chess-shuffle-chess

• The positions of the pieces are decided entirely by the players, not by a computer program. Strategic chess thinking therefore begins with the first piece placement. The two players place their pieces alternately, one at a time. White does not necessarily have any advantage here; in fact, Black may have the advantage because Black gets the first look at the opponent’s placements.

• The pieces may occupy any square as long as the bishops are on opposite colors. The kings do not have to be placed between the rooks.

• Castling is permitted only if the unmoved king is on e1/e8 and an unmoved rook is on a1/a8 or h1/h8; orthodox castling rules apply. The possibility of castling is up to the players, who may or may not place their kings and rooks appropriately.

• There are 8,294,400 possible opening positions.

I actually disagree with the premise of this variant. Because the game starts on the third move of standard chess, knowledge of deeper variations in certain lines would be more valuable. Essentially the player who is lucky enough to be given a third move position in a line he is very familiar with, or another who is unluckily given one he is not familiar with, starts with an immediate advantage or disadvantage, respectively.

Seems to me the premise is more sound than not, as the likelihood of someone running into a line they know well would be more than offset by the lines they don't know well overall.

Also, if you really know a line that well you have a much better than random-ish (as per the distribution of allowed 3-move positions) chance of getting into that line by trying to get into it by making the first moves yourself than getting it by chance, unless perhaps it is a very unusually variation of moves in the first place.

It would be interesting to get some master+ level opinions on this I suppose, I'm certainly no expert on opening theory or statistics..