ELECTRONIC T-NOTES


CHESSBASE USA'S WEEKLY ON-LINE NEWSLETTER


FOR THE WEEK OF DECEMBER 29, 2002


A QUICK LOOK AT SOME NEW FRITZ8 FEATURES

by Steve Lopez

Why a "quick" look? There are several reasons. I've had the program for about a week (as I write this) and everytime I launch Fritz8 I find something new. I don't want to give anyone the mistaken impression that this is a complete rundown of all of the new goodies. Also, these are the holidays and, while I'm normally a hard worker and love to be busy, I do like to kick back a bit at this time of the year and play with the kids' toys when they're not looking (not to mention my own toys -- the new Battlebots board game is trés cool, by the way). But the big reason for the word "quick" is that a lot of these new features are deserving of complete ETN articles of their own and we'll be looking at them in the weeks to come.

So let's tear off the shrik wrap and get right to it...

1. The new "photo-realistic" 3D board

I've always hated 3D boards in chessplaying programs. OK, maybe "hated" is too strong a term. I've never found any that I really liked all that well (with the possible exception of the one in Virtual Chess, which was pretty decent). But after looking at the new "photorealistic" 3D board in Fritz8, all I can say is "Wheeeeeeeeeeeeeeeeee!":

Man, this sucker is sharp! (And this is a screen shot off of my home PC, which has a crappy 8 MB video card. I got a new 64 MB card from S. Claus, so the 3D board might look even better after I've installed it -- I'll let you know). What you're seeing is the "Classic Marble" board (there's also a "Wood" board included). I don't think this graphic really does it true justice -- when Fritz is running on my PC, you can actually see the reflections of the pieces in the board's marble surface.

And, by the way, for the people that hate 3D boards no matter how good they look, don't worry -- the 2D board is still available in the program (as well as the original 3D board for folks running older systems that might not handle the new graphics).

In a future ETN, we'll look more closely at the 3D options, which include the ability to zoom, pan out, rotate the board, and a host of other tweaks.

2. Opening training

Over the years that I've been talking to customers, this is probably the single most-requested feature for future Fritz versions. Relax, it's here:

This dialogue allows you to select an opening from the menu and drill yourself on the opening's moves. If you don't see your favorite opening, don't worry. This feature is configurable (and you guessed it -- that'll be yet another future ETN topic). You can add openings or alter the existing ones in the same way that you can save games and add variations to existing ones.

3. Endgame training

Just as with opening training, you can now drill yourself on numerous basic endings. Unlike the opening training, this isn't a "right answer/wrong answer" thing. This dialogue acts as a shortcut that loads a position for a particular endgame onto the board and lets you play it out against Fritz. So it's a quick way to get a standard endgame position without taking the time to set one up using "Position setup". This, too, is configurable and we'll revisit this feature in a future ETN.

4. "Theme blitz" training

This is a fancy term for a neat new shortcut that lets you play a game (or multiple games) against Fritz from the current board position. If you're playing through a database game and come across a position that looks interesting, you can play from that position against the program. In fact, you always could do this (as far back as Fritz1). So what makes this feature new? The dialogue for this feature gives you some neat options. One is to return the board to this position with a couple of clicks (you used to have to do this manually, meaning that you had to note the move number of the position where you started). With the old method you always had to play from the position with the same side to move (in other words, if it was Black's turn to move in the game from the database, you had to start with a Black move); this new feature just loads the position and the program doesn't care which side moves first -- it's strictly position-driven. So you can play a position with White to move in one game, then switch sides to make it Black's turn to move in the next (a cool way to explore this "with the idea of" annotations in chess books). And you can also use this feature for opening training -- click a special toggle and it sets up the board for the start of a game; your goal is to exactly reproduce the moves from the game's start to the position at which you activate the training feature.

5. The new notation display

Prior versions of Fritz displayed the game notation in a "block paragraph" form that some users found difficult to view (despite the fact that it allows for easier navigation of variations in annotated games). Fritz8 introduces a new notation display called "Scoresheet", which displays the game score in two-column format similar to the tournament scoresheets that we've all come to know and love:

And if you widen the Notation pane, you can display even more moves in the pane (very similar to the standard scoresheet):

If you prefer the old "paragraph block" style, don't sweat it -- it hasn't been replaced. Fritz8 gives you both display options (and you can switch back and forth between them just by clicking tabs on the screen). Note, however, that game annotations and alternative variations aren't displayed in the new Scoresheet notation; only the main line (actual) moves of the game are displayed. (Sorry, you don't get scribbled notes in the Scoresheets margins. That would be cool in a really demented sort of way, but would probably be very tough to program).

6. Multiple engines in Deep Position Analysis mode

Just what it says. You can now set the program up to use multiple engines in this analysis mode. Note that, as with Compare Analysis mode, this will significantly increase the time it takes to perform the analysis (and the math's not too hard -- two engines will take about twice as long, three will triple the time it takes to complete the analysis, etc).

7. The Fritz8 engine itself

Of course, you get the new engine. Which, of course, brings with it the usual baggage and questions in all the Interrant newsgroups: "Is it stronger? How much stronger? Didn't they rename an older engine just to rip us off? Can it bring world peace and cure the common cold? Will it clean my oven while I sleep?" etc.

Well, like Boz Scaggs, I'm bringing you the lowdown. Yes, it's better. No, I don't know by exactly how much. No, it's not Deep Fritz7 in a new wrapper. If more people played chess, we'd likely have a more peaceful world, so that one's a "maybe" until I accumulate more data. My one kid has a cold right now, so just having the CD in the house didn't cure it -- I haven't had him sit down to actually play Fritz yet, though, but I'm pretty well convinced that Fritz8 isn't a faith healer on a disk. My computer's not in the kitchen, so I can't test the oven thing -- but I'm pretty sure that one's a "no".

Enough smartaleckry (and is that a word? It sounds better than it types). Here are a few points that help answer the serious questions (including one I didn't pose a couple of paragraphs back).

Fritz8 analyzes more slowly (in positions per second) than its predecessors. This is because there is more positional knowledge built into the program -- sort of an internal checklist of positional criteria that are compared to each analyzed position to see if any apply to it. This means that the engine must (by necessity) analyze fewer positions per second but will understand more of what it sees. The days of greasy-fast lightning calculator engines are dead -- programs are getting better by being smarter instead of faster (and a few brainy mugs like Mark Uniacke knew this all along; Mark was way ahead of the curve before most of us even saw the curve coming).

Contrary to a few Interrant rumors, Fritz8 is not a single-processor version of Deep Fritz7. I've tested this on a number of "quiescent" positions (positions with no checks, captures, or obvious tactical shots) and the analysis of the two engines often differs. In some cases they disagree right from the get-go, choosing variations that start with different moves. Often they agree on an initial move but vary from each other somewhere deeper in their suggested variations. So, no, they aren't the same engine.

Is it stronger? I'm not a bright enough boy to look at two bits of high-level analysis and determine which suggestion is "better", so I've tried to determine this the old-fashioned way: let the two engines sit on opposite sides of a virtual chessboard and try to rip each other's guts out. So far (on a single processor machine) Fritz8 is kicking DeepFritz 7's butt across the room, up the stairs, out the door, and around the backyard. OK, I exaggerated -- it was only across the office and halfway up the stairs. But the ominous and scary "thwacking" sounds are starting to keep the kids up at night.

So how much stronger does this make Fritz8? Some folks on the Interrant are already claiming 50 Elo points. I won't make that claim -- I haven't seen a 50 point engine strength jump in quite a while (twenty to forty points is pretty much the norm between versions). We'll have to wait and see what the numerous independent testers and enthusiastic amateurs say. But, so far anyway, I've been really pleased with Fritz8's performance against the previous versions, enough so to convince me that it is indeed demonstrably stronger than previous Fritz versions.

There are a pile of other tweaks and changes in the Fritz8 interface and some new features on the Playchess server (which seems to mutate on a near-daily basis anyway) and I'm sure we'll be gawking at those as well in the weeks to come. But for now that should be enough to whet your appetite -- I'm still bugging out over the 3D board.

Until next week, have fun!

HAPPY NEW YEAR!

You can e-mail me with your comments, suggestions, and analysis for Electronic T-Notes.