[Retros] Comments on your article "Castling and En-passant capture in the Codex 2009"

andrew buchanan andrew at anselan.com
Mon Mar 1 13:54:07 EST 2010

Dear Mr Keym,

This is a wonderful, historic article. Well done! You have achieved the awesome task of unifying the concepts of PRA & RS, and you have done it in a way which allows the chess to do much of the talking. There are also some wonderful compositions directly included, and referenced. Most importantly, you have made the head-ache which I've had for 10 years go away! A must-read for any serious retro fan.

I have a number of detailed comments, which I would appreciate if you could respond to:

MINOR ERRORS (very few that I could find :o)

(1) No.1: Loyd #3
If Black cannot castle at all (e.g. K has moved), there is actually a *third* solution 1.Qc5. (This also needs fixing in the Retro-Variants section at the end.)

(2) No.11 Ceriani #2
"Sa1 is not a promoted piece" replace with "Sa1 did not promote on a1".


(3) I don't understand why you take an existing (if newly unused) term
Retro Variants and redefine it completely. Isn't this just a recipe for
confusing later generations? Terms aren't like houses: when it becomes
unoccupied it doesn't have to be occupied again. What's the relationship between the old and new terms. Unless it's very close I suggest you
invent a new term for the concept you have.

(4) How does "new RV" relate to the other conventions? When you go back a move, how do you decide what's legal? Castling & e.p. conventions?
PRA? RS? Or can you use new RV again to go back another move? :o). This
should be spelled out. Really the RV section was the only part of the article that I didn't enthuse over. It seemed a bit rushed.

(5) Were there any compositions that you stumbled across which didn't work for the new approach? All I can think of is if there was a problem which had an intended RS solution, but could also be viewed from an unintended PRA perspective hitherto suppressed because RS was the default.

(6) Is there a current listing of the best PRA task problems?


(7) You introduce a formal methodology in problem 13. This is excellent, and to be honest, it should be introduced right at the beginning, so that readers are crystal clear what you are actually doing. The beginning of the document up is the hardest part to read. I think also your syntax (relying on underscores, fonts, apostrophes etc) is also a bit arduous. Here for example is what I would propose for problem No. 1:

"There are four potential parts to the solution:
(a) no castling rights,
(b) qside castling rights only
(c) kside castling rights only
(d) both sides castling rights.
Retroanalysis tells us that part (d) is impossible.
Part (a) is dominated through the Castling Convention by (b), and also separately by (c). We can therefore eliminate (a) from consideration.
If (d) were legal, it would dominate both (b) and (c), but it is not. So (b) & (c) survive as the two mutually exclusive parts of the problem."

As you go through later problems, you can speed up a bit - here are the notes I wrote for no. 12:

"With two castlings and two eps, there are now 16 potential parts. There is at most one ep. If Black can castle on both sides, then there is an ep. The 11 remaining possibilities are:
(a) Castle both sides, ep d file
(b) Castle both sides, ep f file
(c) Castle kside, ep d file
(d) Castle kside, ep f file
(e) Castle kside, no ep
(f) Castle qside, ep d file
(g) Castle qside, ep f file
(h) Castle qside, no ep
(i) No castling, ep d file
(j) No castling, ep f file
(k) No castling, no ep
The castling convention eliminates: (c),(d),(f),(g),(i),(j),(k)
The en passant convention eliminates: (c),(d),(f),(g),(i),(j)... i.e. nothing that the castling convention didn't already eliminate.
So this leaves (a),(b),(e),(h) as the mutually-exclusive parts."

Just a suggestion - this is certainly the pedagogical approach which I think we should use in the Retro Corner glossary entries.


So, thanks for this. In summary:
- I am OK with this document, apart from the RV stuff
- There are still concerns that I have about the Codex, but not about Article 16
(except wording could be clearer - maybe by putting the methodology into a footnote?)
- The Retro Corner glossary needs to be amended, for the benefit of newcomers.

All the best,

More information about the Retros mailing list