ChessBase 8.0 statistics - part one

CHESSBASE STATISTICS -- PART ONE

by Steve Lopez

Let's give it a whirl using Mega Database 2003 in ChessBase 8. Start CB8 and right-click on the Mega DB 2003 icon. Choose "Search" from the popup menu to bring up the Search mask. Uncheck "Ignore colors" (since we'll want Capa's games as Black). In the first box to the right of "Black" type Capablanca and in the second box to the right of Black type Jose (there is another player named "J. Capablanca", so we'll need to use the Capa's full first name for this search). Then click "OK".

The "Search results" window will appear and begin to display game headers. After a few moments, the progress bar will reach 100% and (if you've done this correctly) your search should yield 409 games. Right-click on the first game in the list to get a popup menu. In this menu, go to "Edit" and choose "Select all" from the submenu. This will highlight all of the games. Then right-click on the first game again and this time pick "Statistics" from the popup menu. You'll get a display that looks like this:

This is a graphical representation of how Capablanca fared with the Black pieces over those 409 games we've found in which he played Black. We get a pie graph showing a visual representation as well as some straight statistics.

In the pie graph, the Black area represents the number of Black wins, the White area is the relative proportion of White wins, and the gray slice of the pie represents drawn games. The first thing we notice is the whopping great size of the Black shading. The books are right -- Capa was a terror when playing Black!

Let's have a look at the actual numbers. To the left of the pie graph is a section which provides the precise figures. Out of 409 games played as Black, Capa won an astounding 200 (48%) of them! This is a very high percentage for wins with the Black pieces. We also see the numbers for draws and losses (162 games/41% and 47 games/11% respectively).

The percentages given in bold type (after "Total") for White and Black have the draws factored in: the "Total" figure shows the percentage of the available points which were earned by White and Black. Here's how it works. If you're a tournament chessplayer, you know that you get a full point for a win and a half point for a draw in a tournament's standings. It's counted that same way for the purposes of the "Total" percentage in this display: each win counts as a point while each draw is a half point. So in the graph above, Capa gets 200 points for his wins as Black and 81 points (half of 162) for his draws. Add them together and you get 281 points earned of the possible 409 points (had he won every game as Black). Use the formula 281/409 and you get 68.7, which is rounded up to 69 -- thus Capa got 69% of the total points available in his 409 games. I find this to be a very useful statistic when looking at a graph of my own performance -- if I'm scoring badly with a particular color, I know which color I should be playing when I'm practicing against Fritz.

Of course, there's no "Rating" information given for Capa because he was active long before the Elo system came into use. If you use this display for rated players, you get the average rating for both White and Black, as well as a performance rating based on the success of the player(s) of each color.

Click on the radio button next to "Length" and you'll get this display:

This provides a bar graph displaying the length of Capa's games as Black. The numbers along the bottom of the graph (the X axis for you mathheads) shows the number of moves, while the lefthand side of the graph (the Y axis) shows the number of games. So, for example, the really tall bar means that Capa played 18 games as Black that lasted for exactly 30 moves. Note the top of the display: it shows 409 games total and then below it has a weird symbol with a number. This shows the overall average number of moves: Capa's games as Black averaged 38 moves.

There are some interesting things you can do with this graph if you perform a slightly different search. Let's say that you do a search for your losses in which you played Black (this gives you a different starting point than the "Capa" search we're doing -- in this case, you're looking at only your losses as Black, not all of your games in which you played that color). In this case, the pie graph won't do you any good: you'll see an all-white "pie". But click on "Length" and you'll see some potentially useful information. For example, if most of your Black losses last for a relatively long number of moves, it means that you're getting hammered in the endgame and that's the phase of the game you should be studying. Likewise if your Black losses tend to be short, it's time to take a hard look at your opening repertoire as Black.

Next click on the "Years" radio button:

Of course, the numbers along the bottom of the graph display the years while the lefthand numbers refer to the number of games. This gives you a nice overview of the time when Capa was most active (1909 to the late teens and again from the mid-1920s through the early 1930s).

Note, too, that there are some arrow and "plus" and "minus" buttons at the bottom of this display. They don't do anything in the "Result" (pie) graph, but they do some useful things to the "Length" and "Year" graphs. The arrow buttons scroll the display to the left and right. The "plus" and "minus" buttons do a number of different things depending on how many times you click them. I'll leave it to you to experiment with this, but in general they "zoom in and out" on the display and will also change the display to show a number at the top of each bar (referring to the exact number of games which that bar represents).

There's a lot more that you can do with this Statistics display in CB8 and we'll look at some of those next week. In the meantime, I'll leave you with a major tip for using the graph in conjunction with player searches (as opposed to opening stats or some other form of search): do color-specific searches (i.e. all the games of a particular player in which he played one color or the other, not both). Why? Because if you do a search for all the games of a player regardless of color, you'll have no reference point for the numbers -- you'll be looking at statistics for that player and his opponents combined, which is essentially meaningless information.

Until next week, have fun!