LCZero-Stockfish Superfinal Variation: Investigation Opened (Part 1)

We are past the mid-point of the "TCEC Top Chess Engine Championship" event, a 100-game Superfinal match between the computer chess programs Stockfish (Stockfish dev-20230409-b36d39) and Lc0 (LCZero 0.30-dag-a9b25c2b-BT2-).

Yury V. Bukayev pointed out to me that two of the games featured the opening 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.0-0 h6, and he suggested that the line be named "LCZero-Stockfish Superfinal variation".

I figured a little investigation might be in order.

It turned out that Stockfish played the unusual 4...h6 in one of the games, and Lc0 played it in the other. For both programs, the moves were described as being in their "book" of prepared moves.

How unusual was the line of play?

I consulted my ChessBase database (2016) and uncovered 320 games, the earliest from 1978.

Given that the Giuoco Piano is one of the oldest recorded openings - Damiano played it in the early 1500s and Greco played it in the early 1600s - that late date of 1978 seemed odd.

I checked Dr. Tim Harding's comprehensive UltraCorr 2023 correspondence database, and did not find an earlier game reference there, either.

In the meantime, I kept thinking about the related Semi-Italian Opening, 1.e4 e5 2.Nf3 Nc6 3.Bc4 h6, examined in Chess Master vs. Chess Amateur (1963) by Max Euwe and Walter Meiden. It is much easier to find early examples, such as Bruijn, W. - Kloos, H., Amsterdam, 1851 (1/2 - 1/2, 39) and Kloos, H. - Blijdensteijn, WJ., Amsterday, 1851 (0-1, 39).

You can march the Semi-Italian Opening one step further with 4.0-0 and find an 1856 game (Willberg - Wolff, Balduin, Berlin, 1856 [1-0, 31] ) as well as 973 games in ChessBase database (2016).

Then, add 4...Bc5 to the Semi-Italian Opening and we are back to 1978 as the earliest game example of the "LCZero-Stockfish Superfinal variation".

From the time of Paul Morphy to that of Anatoly Karpov, the line seems to have disappeared?!

Clearly I need to do more research, and expand my sources.