Friday 10 July 2015

Chess Detective

Now I'm retired from full-time chess magazine editing, I can indulge my various passions (family history, local history, photography, guitar playing, etc) with a clear conscience. And also pay some attention to Britbase, the British chess game archive which I founded nearly 20 years ago.

If I were asked to choose my favourite pastime, I think I would have to summarise it as "looking things up". This encompasses family history, chess database work, and one or two other things. It dates well before the advent of the internet (I used to spend days in musty records offices, tracing ancestors in ledgers back in the 1980s and 1990s, long before Google). There's nothing I like better than browsing through the index at the back of a book - almost any book - or trying to discover arcane facts on the internet. I wouldn't dignify it with the word "research" since it tends to be unstructured and serendipitous, but it occasionally leads me to doing something useful, almost by accident.

I think this love of looking things up dates back to my discovery of chess literacy. I can't quite remember how or when I learnt to read, but I do remember the beginnings of my chess literacy. It was when I opened a book called Chess for Children by Raymond Bott and Stanley Morrison, probably around 1961 when I was eight. I had already been initiated into chess, thanks to my elder brother, so I didn't really need all the stuff about how the pieces moved but I was amazed to learn from the book that a chess game could be written down. That really was a light bulb moment for me. The thought that games could be recorded and played back at one's leisure greatly impressed me. I'm still not entirely sure why but it was a revelation. It still is. Imagine if one could watch historic football or cricket matches, from a hundred years ago, in full colour and all-round stereo sound - that's how it feels for me to play through some ancient chess game, witnessing the exact moves on the board that Staunton, Morphy or Steinitz played. Fifty years on the sense of wonder still hasn't worn off.

A year or two later, still before I had played in any formal competitions or joined a club, I started writing down my own pitiful chess efforts, as well as cutting and pasting chess columns from newspapers into a sort of scrapbook. We took the Manchester Guardian, so that meant Leonard Barden and his wonderful Saturday column, and also the Sunday Times, for which Hugh Alexander was an equally impressive chess writer. The recording was, if anything, more important to me than the playing, so it was almost in order to have an excuse to write more games down in my scrapbook that I felt the need to start playing competition chess in 1967.

I still have this itch to record chess games, hence my creation of Britbase and my work over the years as a game inputter for various competitions. Of course, there are plenty of other people now engaged worldwide on this important work and I salute their efforts. These days I see myself more as a chess detective and a proof-reader, trying to cudgel the mountains of games being input into some semblance of order. I have to try and keep my tendency to perfectionism in check when I witness games being input and disseminated in industrial quantities, with an inevitably large number of mistakes being made. Rather than lambast sloppy inputters (as I admit I may have done on occasions), I think it is generally better to be constructive, bridle back the criticism and help game inputters in their Sisyphean task.

We shouldn't be too surprised or shocked that commercial databases are not as accurate as we would like them to be. I know only too well the time pressures that game inputters are under to get games digitised to accompany press reports and website publicity. And once a given day's pile of scoresheets have been carted away to make room for the next batch it is often too late to fix any errors that may have crept in.

The next stage of the process is when games are collated and published on disk to accompany database software. I'm not quite sure how this process is conducted as I have never personally been involved it, but I suspect very little resource is put into it. I would guess ChessBase (or whoever is producing the disk) hire a handful of editors, who then have to trawl the net and otherwise collate four or five million games, standardising tournament and names, adding ratings, etc, etc. I doubt that they tinker very much with the actual moves as recorded but other things tend to go awry: important bits of data (whether a player lost on time or by 'phone death', a note of the source of the game, etc) can be flushed away, and dodgy decisions made as to which Smith, Jones, Garcia or Sokolov was involved.

Part of my self-imposed Britbase role is to go through old UK bulletins, looking for games which are missing completely from commercial and online databases, but also for games which might already have been recorded but are deficient in some important respect (e.g. misattributed, short on data, such as round numbers and dates, or more importantly with wrong moves). I may also generate 'stubs' or 'blanks', i.e. game records showing players, ratings, round number and date, but with no moves or only a brief textual overview of the game. This can be useful for the generation of crosstables, and also prompts future researchers to go looking for the game moves.

Here's a case in point: I came across this game in Mega Database 2014. N.B. please don't add the following to your database as it contains a number of mistakes as it stands:

[Event "Lloyds Bank op"]
[Site "London"]
[Date "1973.??.??"]
[Round "3"]
[White "O'Kelly de Galway, Alberic"]
[Black "Penrose, Jonathan"]
[Result "1-0"]
[ECO "A56"]
[PlyCount "73"]
[EventDate "1973.08.??"]
[EventType "swiss"]
[EventRounds "9"]
[EventCountry "ENG"]
[Source "ChessBase"]
[SourceDate "2000.11.22"]

1. d4 Nf6 2. c4 c5 3. d5 e5 4. Nc3 d6 5. e4 Be7 6. Qd3 O-O 7. g3 Na6 8. h4 Nc7
9. a4 Na6 10. Bh3 Nb4 11. Qe2 Bxh3 12. Nxh3 Qd7 13. Kf1 Ne8 14. Kg2 f5 15. exf5
Qxf5 16. Ne4 Nf6 17. Nhg5 h6 18. Nxf6+ Rxf6 19. Ne4 Rf7 20. Ra3 Raf8 21. g4 Qd7
22. g5 Kh7 23. Rg3 Rf5 24. gxh6 g6 25. h5 g5 26. Bxg5 Bxg5 27. Nxg5+ Kxh6 28.
Ne4 Rf4 29. Rg6+ Kh7 30. Ng5+ Kh8 31. Rh6+ Kg8 32. Ne6 R8f6 33. Rxf6 Rxf6 34.
Qg4+ Kh8 35. h6 Rf7 36. Qg7+ Rxg7+ 37. hxg7+ 1-0

I found this one amongst another batch of games attributed to Lloyds Bank op 1973. Which brings me to the first error: there was no Lloyds Bank Masters in 1973 (that wonderful series of tournaments didn't start until 1979). The other games I quickly figured out were played at the Islington weekender in December 1973 (which will be coming to Britbase very soon - my thanks to Richard James for lending this and several other bulletins) but this one clearly had nothing to do with it as nobody called O'Kelly or Penrose took part.

The identity of the white player also made me suspicious. ChessBase had already confused the Belgian GM Alberic O'Kelly de Galway (1911-80) with the English player Rory O'Kelly (born 1950) several times in the 1990s, which I had reported to them and they had fixed a few years ago. It was even possible that Black was someone else - Jonathan has a brother Oliver who is a good player, albeit unlikely to play in London since he is based in Edinburgh. I noted that Jonathan Penrose had never played the Czech Benoni before, but of course players can change their opening repertoire at any time.

My next step was to have a look through old magazines for 1973 but I couldn't find any trace of pairings between Penrose and either O'Kelly. Then I discovered that Ray Keene had annotated the game for Informator 17. Ray's version showed white as Rory O'Kelly, Black as Jonathan Penrose, the occasion simply as "England" and the game moves ending at 31.Rh6+.

The next step was an easy one - contact Ray! The Times' chess columnist responded to my email within minutes and helpfully copied in his former school colleague Rory O'Kelly who, a few minutes later, contacted me to confirm that he was the player of the white pieces:

"The game was played between myself and Jonathan Penrose in a London League match between Mushrooms and Hampstead on 3rd April 1973. It actually ended 31 Rh6+ which was my sealed move. Penrose resigned without resuming. As I recall I felt confident enough to tell him what I had sealed.
My 6 Qd3 was a TN - probably my only contribution to opening theory. The point was to play h4 and Bh3 without allowing a b5 sacrifice. In this case it also had the fortunate though unintended consequence of tempting my opponent into a very unhelpful excursion with his QN."
So, mystery solved. Here is the game, for your enjoyment, and to allow you to fix the game on the database*. My thanks to Ray and Rory for helping me to solve this mini-mystery.

* To capture the PGN data, click on the c8 square on the board, then cut and paste from the pop-up window