Bob Durrett on specialized databases

I'd first like to detour by presenting a ChessBase Workshop column which is long overdue. About two years ago, in a column devoted to reader mail, I printed a highly intelligent missive from Madison, WI, USA's own Bob Durrett. The arrival of an e-mail from Bob hitting my box is always a treat, due to the thought and insight he brings to the use of ChessBase software -- in short, Bob thinks of things that I don't, so I'm always happy to share his ideas with the readers of CB Workshop. Part of his letter from that particular two year old column is reproduced here:

(5) I use a White Study Pointer, Black Study Pointer, White Opening Repertoire, Black Opening Repertoire, tree of recent games where each player is ELO 2600 or above, the same for 2500+, and a few other tools. If you have the wild urge to read a letter on how these tools are created, grown, and used then send me an email and I'll reply.

...The Study Pointer and Repertoire databases ... are "living documents" inasmuch that they change after every game.

I've received several e-mails concerning that tantalizing tidbit, asking me if Bob ever offered any further exposition on these special databases. I invited him to do so in that same column:

5) Fire away. I'll probably end up overwhelmed and confused ("like a duck who's been struck on the head" as Lincoln so famously worded it) but I'm game.

...but Bob never replied. It seems that he'd abandoned chess as his primary hobby in favor of a new pastime: molecular biology. (I would love to say I'm making that up, but it's the truth; Bob's one seriously smart cookie). Urged by several readers, I finally e-mailed Bob several months ago and received a reply explaining his "Study Pointer" databases in more detail. He graciously granted me permission to reproduce his comments in ChessBase Workshop, so I'll now turn the floor over to Mr. Durrett:

I am not doing chess right now, having moved on to other things. At the time I wrote those words, I was spending almost all my waking hours on chess. Being retired, this translates to more than ten hours per day since I had nothing else to do.

Whether or not you will wish to adopt my methodology will depend on how much time you want to spend on chess, how patient you are, and your objectives.

Make no mistake about it! Chess can consume ALL your time if you let it.

Chess was my passion for more than 50 years. Unfortunately, until I retired, I simply did not have enough time to do it right. It was only after retirement, 13 years ago, that I began doing chess full-time.

I decided to create an opening repertoire consistent with current practice at the GM levels. Previously, I had never taken the time to study much opening theory being preoccupied with work and college. Instead, my tournament games [last one in 1995] were all played with openings which I made up on the spur of the moment. I played mostly French as Black and almost everything as White. This approach allowed me to play at the USCF Class A level only. Advancing to higher levels without study of openings seems improbable unless one has exceptional talent and other exceptional mental attributes. Having qualified help would also make a big difference.

The "Study Pointers" were databases of opening lines created to add discipline to my opening study efforts and to prioritize the work. The underlying idea was to work first on the opening lines played most often and to postpone study of less popular [at the top GMs level] opening lines. Technically, the Study Pointers are customized ChessBase databases. Each line extended only as far as I had investigated that line at the time. I used the "Black ELO" column to enter numbers representing the number of times the line had occurred in the top GM [2600+] games. The database was sorted by the number of games played, with lines played most often listed first. It was these lines which I studied first. After a line had been investigated, the line would split into two or more lines, each with a "number of games played" less than the original number of games played. Hence, it was necessary to re-sort the study pointer database often. I alternated between White and Black, spending a month or two on one and then going over to the other.

The opening repertoires resulting from this study effort were "works in progress." As I adopted each new line, I added it to my repertoire and committed it to memory. Most people keep their White and Black repertoires in a single database but I had two, one for White and one for Black. That has both advantages and disadvantages compared to using a single repertoire database.

The 2700+, 2600+ and 2500+ databases were actually trees, constructed from ChessBase Megabase, current version. The 2700+ tree, for example, was constructed by first creating a database of all games in which both players were rated 2700 or above [ELO] and then creating a tree from it using the tools available in ChessBase. It was from these trees that I was able to determine the popularities of each line. These popularities were entered into the "Black ELO" column as number of GM games played with that line. [Technically, the number of games was the number of games in which the end position of that line had occurred. This is significant since opening line transpositions abound in GM chess.]

Not having the time or money to play in over-the-board tournaments anymore, I tested my new lines in internet chess, playing blitz. Generally, I tried to get people to play at longer time limits but most people playing internet chess want to play fast. The outcomes of the games were not significantly influenced by my choices of openings since middlegame and endgame errors usually determine blitz game outcomes. The faster you play, the more errors you will make. After playing a couple dozen games or more, I would then do exhaustive post-mortem analyses of one or two of the games. This included opening investigations, and the opening findings also fed into my Study Pointers. I did a lot of work identifying my personal "mental process" defects and working to correct them. This self-analysis did a lot more good for blitz than opening study.

Please do not assume that I was spending all my time developing opening repertoires or playing blitz chess. I spent more than half of my time analyzing GM games. That can be very productive if done right.

The above is actually an excerpt from an e-mail that Bob wrote to another ChessBase user (who'd written to him directly) and which Bob carboned to me. The ultimate result of this correspondence was that I'd inadvertantly dragged Bob back into the chess world (whether or not he was kicking and screaming while being dragged remains unknown). The imaginative and irrepressible Mr. Durrett later sent me a much longer and detailed missive, which expanded on the above material and provided what is in essence a complete repertoire study course. Again with Bob's permission, I'll be reproducing that e-mail as well here in ChessBase Workshop in another few weeks, so watch for it.

In the meantime, I've not dipped into the virtual mailbag in several months, so we'll look at some of the questions which have come my way in next week's ChessBase Workshop. Until then, have fun!

You can e-mail me with your comments on ChessBase Workshop. All responses will be read, and sending an e-mail to this address grants us permission to use it in a future column. No tech support questions, please.

© 2007, Steven A. Lopez. All rights reserved.