ChessBase Light 2007 - part 8

We've been considering various ways to run engines and display the outputs. But we haven't yet looked at why you'd want to use an engine to analyze a position.

I'm often asked for advice on the "best" engine settings to use in order to get "perfect" analysis of a position. Unfortunately that question misses a fact about chess engines: they won't give you "perfect" analysis. In the case of 99% of chessplayers, they will give you better analysis than what you're liable to find on your own. Chess engines aren't "perfect" chessplayers, but they're likely better players than you are (unless you're a top-level GM). So what you're after isn't a "perfect" suggestion, just the opinion of a better player (and this is why you'll frequently see different chess engines give different recommendations in their analysis).

Let's consider a sample position from one of our Linares '91 games:

White has just played 21.Qa4, backing the Queen up from b5 in response to Black's 20...a6. It's an interesting position. Black has various possibilities here. So do we: for example, we can fire up an engine right away to see it's opinion on what Black's next move should be, we can examine our own ideas as to Black's best course of play (double-checking our own ideas by using an engine), and we can jump ahead to Black's next move and have the engine analyze the new position. (There's also the possibility of launching an engine and displaying multiple variations to see where a particular idea ranks in the hierarchy, but we've discussed the multi-variation display at length previously, so we'll omit it at this time).

Let's start off by launching an engine to see what it thinks Black should do in this position. We'll go to the Engine menu and click "Add kibitzer: Fritz". As we've seen in previous columns, Fritz will start up and begin chewing on the position. It's a matter of personal preference (as well as hardware speed) as to how long you should let an engine analyze, but I'm a firm believer that you should let an engine analyze to a ply depth at least somewhere in the teens -- anything less should be considered very cursory analysis. The maximum is up to you.

Once we're satisfied with the depth, we can play through the variation using the Variation Board (see previous columns for more information on this feature). We can also right-click in the engine analysis pane and select "Copy to notation"; the top variation at that point will now appear in the Notation pane as a replayable variation directly in the gamescore:

As I said, this is a replayable variation. In the example you could click directly on the first move of the variation (21...h6) and play through the entire variation just as we would the main line moves of the game. The illustration above bears a bit of explanation. You'll see the variation begin with the name of the engine which generated the analysis. The variation ends with a numerical evaluation (in this case -0.01 meaning that Black is 1/100th of a pawn ahead; for all intents and purposes we can consider this position dead even). The number after the slash (in this case "18") refers to the ply depth achieved by the engine.

Another path we can take is to check out our own ideas in a position. I'm thinking that 21...b5 (to kick the Queen again while grabbing some space) is a possibility. Let's push that pawn on the chessboard (grab it with the mouse and move it) to see what happens.

The first thing we'll notice is that we get a popup dialogue:

I won't go into detail here on this dialogue (it's been covered in previous ChessBase Workshops); obviously, though, we want to click the "New variation" button at this time:

...and we see that the move's been added to the gamescore. Now we can start Fritz, let it run, and see what it thinks of this idea (copying the analysis to the gamescore as we did previously):

We see that Fritz' analysis of the position after 21...b5 has been added to the gamescore as an extension of the variation we added manually. After a seventeen ply search we see that 21...b5 leads to a position that's a little better for White, but not by a huge amount (only 8/100ths of a pawn separate the two variations). So Fritz doesn't think that 21...b5 is the best move, but there's certainly nothing terribly wrong with it.

You've doubtless noticed in the above illustrations that Black actually played 21...Bf8 in the game. Let's click on that move in the gamescore to jump to that position; then we'll start Fritz and see what the chess engine thinks:

Since after 21...Bf8 it's White's turn to move, you'll note that Fritz' analysis has been added as an alternative to White's 22nd move.

So let's say we want to save this analysis as a permanent part of this gamescore. We go to the File menu, select "Replace game", and we see this:

Whoops.

We'll look at the activation procedure in the 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.
 

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