[Retros] Fwd: FW: How many different positions?

joose norri joose_norri at hotmail.com
Wed Jan 2 13:26:38 EST 2019

Olavi Riihimaa published the same position, except a4<>b4, d4<>e4, g4<>h4, in the British Chess Magazine November 1954, see also Schach und Zahl. He arrived at 192 positions.

As to your second question, Olli Heimo constructed a position with I believe something like wKe1 Rh1 - bKe8 pe2, where obviously white can never castle, even though he hasn't lost the right to do so.


Subject: FW: How many different positions?
Date: Wed, 2 Jan 2019 11:21:20 +0000

From: Mestel, Jonathan
Sent: 02 January 2019 11:20
To: The Retrograde Analysis Mailing List <retros at janko.at>
Subject: How many different positions?

Happy New Year All.

Someone must have done this before. How many different positions can look
the same? We define two positions which are identical in appearance as
different if a sequence of moves could occur in one but not the other, so
that castling or en passant rights may lead to different positions. For
example, the attached position 3k2r/8/8/pPppPppP/pPPpPPpP/8/8/R3K2R I think
could represent 200 different positions:

192 = 4 * 4 * 6 * 2 = (White castling)*(Black castling)*(en passant)*(whose

Then change a1 to h8  and add 2*4 = (whose move)*(en passant)

Giving a nice round total of 200. Is this maximal?

And a related question. According to the current 3fold repetition rules of
Fide, in the game

1 e4 d6 2 Bc4 f6 3 Bf7 Kd7 4 Bd5 Ke8 5 Bf7 Kd7 6 Bb3 Ke8 can White claim a
draw on the grounds he intends to play 7 Bf7 ? The 3 positions are not
identical in the sense that in the first Black could castle, and yet
obviously no legal continuation would enable him to castle. So the
positions are logically the same.

  Best wishes,


Jonathan Mestel
Professor of Applied Mathematics
Imperial College London SW7 2AZ
Huxley Building 746
j.mestel at ic.ac.uk<mailto:j.mestel at ic.ac.uk>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://pairlist1.pair.net/pipermail/retros/attachments/20190102/69031c00/attachment-0001.html>

More information about the Retros mailing list