4

Offhand, I say the Shepherd's Mate. But let us just consider "serious" tournament play. Still, the game must be rather short to be repeated extremely often. Could be a "Will they ever learn" opening trap or a prearranged draw line. Anyone care to research it in a megabase?

Note that just because Lichess lists 187×10^6 e4 vs. 100×10^6 d4 doesn't mean that an e4 game will be the winner! (By playing always the most often played move on the analysis board, I landed in the Lasker Pelikan Sveshnikov.)

lodebari
  • 3,323
  • 2
  • 22
  • 44
Hauke Reddmann
  • 14,860
  • 2
  • 24
  • 61
  • Use "×" instead of "`*`" in SE. – user21820 Jul 12 '21 at 18:40
  • 3
    That repetition draw in the Zaitsev Ruy Lopez is a good contender. Some spectacular draws like Haupke-Meitner, Vienna 1889 have been repeted many times also... – Evargalo Jul 12 '21 at 20:28
  • 1
    @Evargalo: You say :-) (Caught two youth players myself. Me: Oh, Hamppe-Meitner! Them: Who??? Their luck it was none of my business...) – Hauke Reddmann Jul 13 '21 at 07:15
  • 1
    [Are there any repeated games between GMs?](https://chess.stackexchange.com/q/29129/7297) – Glorfindel Jul 13 '21 at 10:19
  • @Glorfindel: Judging from Lichess (which, I dunno, might include blitz etc.), this opening draw is somewhat less frequent than the Zaitsev. BTW, Hamppe-Meitner is more than one order of magnitude less frequent. (Maybe because of the fear there is a trick for one side...Stockfish declares it as +10.0, until Nxc6, 0.0.) – Hauke Reddmann Jul 13 '21 at 14:47
  • like the berlin variation where 3 fold occurs. – CognisMantis Jul 24 '21 at 03:36

1 Answers1

7

The most played game: a drawing line

The most played game in the TWIC database (issues 920 to 1438, containing about 1.8 million tournament games) is the following one, which appears 111 times. I add also a variation of the same (a change in move order) that is the second most played game (87 times). A couple of variations ending a move earlier or a move later also are amongst the most played games, with 38 and 39 repetitions respectively.

[FEN ""]
1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6 4.O-O Nxe4 5.d4 Nd6 6.dxe5 Nxb5 7.a4 Nbd4 8.
Nxd4 Nxd4 ( 8...d5 9.exd6 Nxd4 10.Qxd4 Qxd6 11.Qe4+ Qe6 12.Qd4 Qd6 13.Qe4+
Qe6 14.Qd4 Qd6 {2nd most played game, 87 times} ) 9.Qxd4 d5 10.exd6 Qxd6 
11.Qe4+ Qe6 12.Qd4 Qd6 13.Qe4+ Qe6 14.Qd4 {15th most played game, 38 times
} 14...Qd6 {Most played game, 111 times} ( 14...Qd6 15.Qe4+ {14th most 
played game, 39 times} ) 1/2-1/2

How I found the answer

In case anyone is interested, I used a custom made script to parse the games in pgn format and extract the movetext section (the section containing the actual moves and result).

After saving the results as a dataframe in R, I grouped the data by the column movetext (equal games are grouped together), calculated the number of times for each game and sorted in decreasing order:

    games <- mdata %>% group_by(movetext) %>%
        summarise(n_times = n()) %>% arrange(desc(n_times))

And I only took into account the games where there was at least one move by each of the players. Otherwise, the most played game would be "0-1", black wins because white didn't show up, followed by "1.e4 1-0", white wins because black didn't show up.

lodebari
  • 3,323
  • 2
  • 22
  • 44