21 August 2010

The Bernanke Positions

When a few of the chess960 fans at Chess.com were discussing how to convert the character representation of a random start position to its chess960 numeric identifier -- e.g. BRNQNKRB to no. 419 (see forum post Chess960 Starting Position ID) -- I pointed out that search engines can do this.
You can use Google to translate start positions into the corresponding ID. For example, if you search on 'BBQNNRKR' you'll see without even clicking through the results that it's position 0. On top of returning pages that are lists of all start positions, the search will turn up various odds and ends on chess960.

It's a useful tip, which is why I'm repeating it here. By 'odds and ends', I mean the two sites I mentioned at the end of my first post on the subject, Searching for BNRKNBRQ, where I left off with the time honored phrase 'I'll look at those sites in another post'.

One of those sites, 64squar.es, is currently returning the message 'Down for maintenence. Sorry the site's been getting too slow, we're working on a new version', so I may have waited too long. The other site, Wildchess.org, is currently saying 'Games in database: Fischer Random 11338' for its Fischer Random database, up from 9028 on the BNRKNBRQ post. I was wrong to pigeonhole it as 'a database site', because the FRC page also lists recent live games.

What other chess960 odds and ends are available on the web? I repeated the search on two start positions recently discussed on this blog. The first, BNNRKBRQ, returned about a half-dozen pages that list all 960 start positions, including my own SPs & Twins, linked in the sidebar. Following other links to their primary chess960 pages, I found MySchach.de, along with a handful of sites that I had already seen, most of which haven't been updated in years. The second, BRNQNKRB, led me to the usual lists of 960 positions plus MyChess.de, an alias for the MySchach.de domain.

More interesting on the BRNQNKRB search was Google's question, 'Did you mean: BERNANKE', the current Chairman of the United States Federal Reserve. My first response was, 'Huh? What's that got to do with chess960?'. My second was, 'Now I get it. Removing the vowels in BERNANKE gives BRNNK, which are five of the pieces in BRNQNKRB and in the correct order.' Google, you're trying too hard!

Given the first five letters BRNNK, the missing pieces are QRB. Since the missing Bishop must be on a light square, there are exactly four chess960 positions that start with BRNNK. I nominate these as the Bernanke Positions. What other English words match nontrivial sequences of legitimate chess960 positions?


Ichabod said...

Of course, there's more than one way to number starting positions.

Mark Weeks said...

Re 'more than one way to number starting positions', I saw somewhere that Chessbase uses a different numbering system. Why don't other systems show up on my searches? Google is, after all, a disinterested party and returns any page having the random letters that denote a particular chess960 start position. Where are the advocates of the other systems? - Mark

Ichabod said...

My information comes from Milener's book. He mentions two that he doesn't really go into in depth. One is Ferdnand Joseph's system, which he says is based on an itterative computer algorithm (?). It can be found at http://www.xs4all.nl/~marcsmet/ebbs/download/FRC_960Pos.pdf (note that the link in Milener's book is out of date). It assigns RKRBBQNN to position 1 and the standard position to position 518.

He says chessgames.com uses a different system, based on the number 5. It won't give me a specific position, but it gives position 19 as being RNBNKRQB, which (if I'm doing it right) is position 471 under Scharnagl's system (the standard).

Milener also describe's his own system, which gives position 0 as BQRBKNNR and assigns the standard position to 362. Then there is the standard (Scharnagl) system and the Fritz system, which I think are both described in Wikipedia.

Then there is my system, which I just invented after looking at all the other systems. They all seem to have complicated systems for generating numbers. I say, make fen strings of all of them, and alphabetize the fen strings. That gives BBNNQRKR as position 1 (only computer programmers start counting at 0), and assigns the starting position number 829.