Fritz calculation training - part two

Why do I say "nearly" any position? Some positions aren't well-suited to this sort of thing; for example, opening positions which we know from rote memorization and positions which are one or two plies away from mate are certainly usable but sort of violate the spirit of the thing. The "Calculation training" feature is best suited to positions in which clear visualization is required but which don't necessarily have a single clear-cut "solution" (as would be the case with classic tactics or "mate in..." problems).

Let's look at an example, using a position chosen at random from a large database of games:

If you're interested in looking this position up yourself, it's from the game Szumilo-Mikrut, Podhale-ch, 2002. Black has just played 27...f6.

After deciding on a position in which to train, the next step is to activate the Calculation training feature. Go to the Tools menu, select Training, and then "Calculation training". You'll see the normal Fritz Menu bar and toolbar be replaced with a new toolbar which looks like this:

We'll return to this shortly. You'll also see the normal Notation pane change to display only the side to move:

The chessboard in this view appears just as it does in the normal Fritz gameplaying view. But there's an important difference: when you "move" a piece in the Calculation training display, it's not actually moved on the screen's chessboard. In other words, you'll need to mentally calculate a variation by moving the pieces in your mind's eye (just as you do in a real chess game), and then record that variation on the screen without physically moving the pieces on the screen's board.

Here's an example using our chosen sample position. White's to move, so let's say I'd like to play Rd1. I click on White's f1-Rook and hold down the left mouse button. I then move the cursor to d1 and release the left mouse button. I've just played Rf1-d1 (as shown in the Notation pane):

...but the on-screen chessboard looks like this:

Compare it to the first chessboard shown above. Right -- there's no difference. That's why this exercise is called "calculation training": you must calculate a variation, making moves for both sides in your head (just as in a real chess game played against a human or computer opponent), and then make those moves on the board simply as a means of telling the software the moves of your calculated variation. The board won't change position, so you'll need to hold those moves in your mind (just as in an actual game) without any "help" provided by seeing the pieces change position on the chessboard.

Note, too, that you'll need to calculate moves for both players, just as in a real chess game. Enter moves for both sides as far ahead as you're able to calculate. For our example, let's say I enter the following moves:

28.Rd1 Be3+ 29.Kh1 Bb6 30.Bf4 Rc8 31.Qd2

...and though the moves are recorded by the Fritz software (as shown in the illustration immediately above), the pieces on the board haven't moved from the original position. In essence, this is a form of blindfold chess in which you're tasked to find the best possible continuation for as many moves as you can without actually changing the board position. (Speaking bluntly, everytime we analyze a position over the board, we're playing a truncated form of blindfold chess).

How do we move pieces on the board that have already moved? For example, Black's Bishop went to e3 and then to b6, so how did we make the latter move? Very simply, we clicked on e3, held down the mouse button, moved the cursor to b6, and released the button.

While we're talking about recording our intended variation, it is possible to make illegal moves while in Calculation training mode; we'll discuss that aspect of the training a bit later. First, though, we'll simply discuss the "meat and potatoes" of this training feature -- namely, Fritz checking your work.

After you've calculated as far ahead as you can (or as far ahead as you choose), you can have the Fritz software "grade" the variation you entered. Click on the toolbar button which looks like a paper and pencil superimposed over a computer monitor:

...and you'll see the following dialogue appear:

This dialogue controls a couple of output settings for Fritz. If you select "Annotations", Fritz will insert short verbal commentary describing how well you calculated the variation. The other setting controls the number of seconds Fritz will spend on each move of your variation. The higher the value, the more precise the software will be in its assessment, but there's also a trade-off here -- the higher the value, the longer the software will take to complete the evaluation. Low values take less time but will also be very cursory analysis. The exact amount will depend in large measure on your hardware, so experiment until you find a value which suits your available time and requirements; I recommend a value between "30" and "90", but your proverbial mileage may vary.

Ultimately, after Fritz completes its analysis, you'll wind up with something which looks like this:

You'll receive an overall score for the variation you've calculated, as well as specific scores for components of said variation. In the above illustration, Fritz grades us poorly on the choice of initial candidate move (in the software's opinion there were four candidate moves which were better), but gives us passable marks for the early moves of our calculated line of play. Overall, though, Fritz doesn't think much of the moves we've chosen and gives us a negative score overall.

I have a pretty good idea about why we received a poor score overall. The Fritz software provides us with tools for double-checking those possibilities, and we'll look at those in our next 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.

© 2008, Steven A. Lopez. All rights reserved.