Saturday, November 21, 2009

Part 3: Finding training patterns in your Repertoire

In part 1 I proclaimed my new training regimen which basically consists of creating a personalized set of positions from my own games, repertoire and study material.

In Part 2 I showed you how I used chess base to create my own training positions to have a set of tactical and positional puzzles themed from my own games.

In this entry I will show you how I develop a repertoire data base and how I use this to create study positions.

But first a brief update on the growing A.C.I.S of Caissa movement:

Following an action from the playbook of Loomis, I plan on keeping the ACIS of Caaissa updates limited to THIS blog and not on my mirrored site at the greater community at I like the smaller closer circle of friends here as it tends to promote a more supportive environment. The "how to" stuff will get forward to

Unlike the brutish Knight’s errant DLM movement of the mid ‘00’s, A.C.I.S. of Caissa is proving to be more accessible to the “common folk” as the final circles of the MDLM method was just unreachable if you work, have a family and want to practice good hygiene. If the truth be known, most of the knights errant had modified the MDLM method to smaller circles and approached it in a more realistic manner.

Loomis has joined in this universalistic approach and proclaims to be Baaaack.

Steve (learn’s chess) Eddins is firing up the blog to declare his mission and thoughts.

Chunky Rook has fired off a series of gif patterns worth checking out on his blog

Linuxguy reviews a game he played on FICS and shows appreciation to having studied Zurich 1953

And finally, Chess tiger was lulled into this quest with this line:

What pulled my attention is that one may choose his or her own study plan. So
one isn't pushed towards Rapid Chess Improvement of Michael De La Maza or How to Reassess Your Chess from Jeremy Silman or Novice Nook written by Dan Heisman orInternational Chess School (ICS) or Lev Albert's Chess Course or ... . This is a good thing because for all we know, all combined may bring a bigger outcome chess wise then following only one of these courses.

Building a repertoire database.

I use chess base for a lot of reasons. One of the things I’ve done was create a blunder-rep database with games centered around my openings I play. First, I pull in games that I have studied from the classic tournaments that are in any shape or form close to what I play in my chess games.

Hastings 1895, London 1924 and Zurich 1953 is not enough resources for what I am looking for. There are several ways to approach this. You can set up a position using chessbase and use the search online tool to pull games from their huge inventory. I find this tedious as I haven’t found a clean way to import the large volume of games as a result of this method. The best I was able to achieve was dumping them all into one huge game file or saving each one individually. I will use this method to find key players ( grandmasters) who play this variation but not as a means to build the volume I seek for the purpose of building a training database.

I wanted a quicker method to build the base up. I use google to search for PGN or CBH data bases of specific variations. There are several websites that fill this gap. will allow a search for the position and provide a collection of games to download as PGN. Chessopolis ( is another resource I use frequently and they actually have CBH files that can import directly to Chess base. There are plenty more if you search.

The trouble with “canned” data sets from some of these places is the quality of games are littered with amateur games. But my philosophy at this stage in my improvement path is that I can still learn from these amateurs.

Panning for gold.

Once you have a repertoire database built up, the next step is to use the search capability of chess base to find positions to study.

Finding Traps in the opening to avoid or inflict:

The first thing I do is to find the opening traps I want to avoid. I will set the search to find the games that end in 15 moves or less where the side I would most likely play loses.

I will create training positions described in my previous post for each of the unique wins. Some are duplicates and worth skipping over. What you get is a clear pattern of what not to play in certain lines. Optionally, you could run the engine on each of these to get some annotations and ideas what to play. I merely reference my books and make a quick note where to improve and what not to play. The opening tactical trap becomes the positional study that I solve for the aggressor. Then I look at the notes in the game centered around the failing position. This is where having an amateur database comes in handy as you will more likely have a lot of examples to chose from.

On the flip side, from the same repertoire base I will change the search to games where my side wins and repeat the process. The result will build up tactical positions found in the openings of my games that I can inflict if my opponent doesn’t play exactly in this line. Positional themes start to come about from these and I get a better understanding of the opening.

Finding Mating themes:

Another search I will conduct in the bluder-rep is to find those games that have ended definitively with a check mate. To weed out the previous search I set the move order to a range greater than 15 to include the long games. I go through the same process of looking at wins for both sides to see the kind of attacks typical from both perspectives. I then create training positions from these making notes of the type of attack as a memory marker for the pattern.

For added measure I use the same filter but instead of definitive mates, I search for results being my side to win. This will include winning endgame positions to come about in my games with higher probability.

Middle game positions:

There’s no way around this but to review games against masters who play the same openings in your repertoire. I am building on this with my tournament games studies and include several positions from each of the highlighted games.

So far I have just over 100 positions as I build upon this. I think it’s a good start. I’d like to build this to at least 500 by Spring, but I don’t want to get stuck in the process before using it. 100 problems to start with will be a good litmus for the upcoming Pillsbury Memorial here at the end of the month.


Will said...

Another interesting post. This reminds me of Jesper Hall's Chess training for budding champions where he talks about studying position types that arise from your openings. Then he searches for typical structures that arise and then finds the typical plans, combinations etc.

Steve Wollkind said...

BP, I'm loving this series. One of my big problems with chess study is that it has typically been too undirected. I noodle around instead of moving with direction.

As my and my opponents' play improve, more often I reach the end of my opening knowledge and am unsure about the best plan. I've known for some time that the best way to combat this would be looking at, even rapdily, games in these lines, just to get an idea of how they go. I've never gotten motivated before, but your comments about finding traps to inflict or ignore really got me going. I spent the last hour or so looking at games white lost in the closed sicilian in short order and have already learned some valuable lessons about what not to do. I'm even seeing white fall for the same traps over and over to the point where I now recognize them.

If only they had read blunderprone and studied the miniatures!

BlunderProne said...

@will: thanks, I'll have to look up Jesper Hall's stuff.

@Steve W ( the former Mr. Zweiblumen): Glad to have lit a spark. Having direction is key in self study. Does this mean you are joining the ranks of ACIS of Caissa?

Grandpatzer said...

I've done a couple exercises using ChessBase that have returned interesting results.

One is to take my large database of personal games (>8000), construct an opening tree, and determine which lines I have a >1% chance of seeing in my own games. The result boiled down to 5 pages when I told ChessBase to "Print Repertoire". I'm trying to make sure that I know the main lines for each of these variations stone cold.

The next level that I'm dabbling with is constructing a "Grandpatzer ECO": using an opening tree constructed from my own games to record every "book line" (either an actual book, or a common line acording to my Mega Database) that's ever occured in one of my games. This is much more than 5 pages, but still it's interesting to see how minimally "book knowledge" applies to amateur games.

This sort or analysis does help focus attention to lines that you're more likely to encounter in practice. For example, at my level the Steinitz and Steinitz Deferred in the Ruy Lopez are major lines.

Unknown said...

BP, please add my blog "Quest of the Chess Novice" to your ACIS crew:

My latest blog entry commits to learning your method, and reporting back.

I look forward to learning and tailoring your method!

Loomis said...

Thanks for the link, your site is far and away the number one provider of visitors to my blog in the last 3 days. :-)

This is a great How-To series. Have you considered using labels on your posts? It's a little work at first to go back and retro label, but easy to do as you go along. Then people can find info by topic. I picked up that idea from likesforests who categorized different endings that appeared on his blog. I've used it for making a CT-Art errata at my blog as well as tagging progress report, tournament report, tactics puzzles posts, etc.

I'll second the positive comment on the supportive group of friends you get from having a blog in the chess improvement blog-o-sphere. Happy to be a part of it.

Howard Goldowsky said...

I have to ask: What percentage of time here is spent preparing to learn, and what percentage is actually spent learning?

I like Jesper Hall's book. It's got a nice outline of how and what to train, from a general perspective.

Typical positions arising from ones opening systems are nice to know; however, I find in my own training that learning the minutia of specific move orders and specific moves, even "traps," only helps if it contributes to learning the typical ideas and structures associated with a family of openings. Too many openings transpose, drift into each other, have similar pawn structures, etc.

Which leads to another question: If the tactics you're looking at are still close enough to the beginning of the game that you can classify them as "a typical tactics for my opening system," then are you training your general tactical ability that will benefit you at all stages of the game or are you merely training your ability to recognize "traps" in your opening?

By moving a single pawn, one square in any direction, you can change a tactical puzzle from one motif to a completely different motif. So is there value in spending time sorting games and tactical patterns by specific openings? maybe. A single pawn (or piece) move, one square, might completely change the nature of the position.

I present these questions simply as an advocate for the devil, not in ill will. :)

BlunderProne said...

@GP. Sounds like you do a couple of useful "how to" posts. I'm only scratching the surface in comparison.

@Harvey: done

@loomis: Your welcome. This seems to be growing.

@Howard: But I thought I was the "duval's" advocate? This post is a mere glimpse of my entire training database which focused on opening positions to include. I do mix in all aspects of the game in my training. Typical endgames that come from this repertoire, middle game positions, and my horrendous games.

The Mascot said...

I see. Trying to make your blog all "cool" and "Matrix-esque" with that picture of all those numbers.

But until you show us a video of you dodging a Bishop launched at your nose (in slow-motion), I'm not buying it.

From the patzer said...

Looks like lots of work. Cant you not just play a game, look opening up in NCO or another big opening book? See where you left theory and then just add one move extra to your 'brain'knowlegde.

That way you dont spent hours just preparing a database but you actually spent time learning an opening in a quick and correct way.

Dont get me wrong, you system is fine aswell but i am affraid that you have to many things to look at and the chance that you forgot 80 percent of all this knowlegde when it comes on the board.

With other words, i am not attacking you method, i am just worried that the turnout of all this hard work will be minimal.

BlunderProne said...

@Mascot: I'm getting ready to film that scene with the bishop.

@CT: I understand your concern. I've gotten to the point where that method was not effective enough for retention. I can actually create these positional exercises rather quickly. I was able to add 100 positions described in this post within one week. Now I have a tool I can come back too and run through training exercises to help my retention.

Steve Wollkind said...

I would like to join the ACIS, but I am not sure I have a good enough plan of study. I actually was planning to talk to you about this last night at the club but traffic was awful and I arrived with little time to spare.

My chess study time has often been very poorly planned and undirected. I have many books that would have good things to teach me if I sat down and really worked through them, but I find it more fun to just leaf through and read the text. I've never made it through a game collection and have been pretty lazy about really looking through games in the openings I play.

Are you in the business of helping wannabe ACIS members craft their plans of study?

Robert Pearson said...

I'm getting Chessbase for Christmas--I am like the last person in the whole chess world to do so? Anyway, I look forward to seing how I can use it to focus my study, because like you, Steve Wollkind, I have spent a lot of time reading chess books (and enjoying it) over the years, much less time on a focused, written improvement plan!