drunknknite
He was winning,
but he didn't see it
and I escaped - as usual.

-Levon Aronian

Speaking Up For Those Who Can't

By drunknknite
Recently wang berated chess engines and databases. So I'm going to go ahead and play Devil's Advocate. First of all because I am a former debater and future lawyer and I love writing arguments. Secondly because I am someone who uses chess engines and databases very frequently and I believe they are very valuable to a chess player when used responsibly. So I feel it is my duty to defend these helpless machines...

I'm going to take a very structured approach to supporting the opposing side so it would probably be helpful to familiarize yourself with wang's post before continuing.

To give you a road map I'm going to start by answering wang's points on learning openings, then I'm going to move on to his points on game analysis.

1. This has nothing to do with computers. I am currently looking at a 280 page book on the Sicilian Sozin (1.e4 c5 2.Nf3 d6 3.d4 cd 4.Nd4 Nf6 5.Nc3 is the start then White plays 6.Bc4 against either 5...a6 or 5...Nc6). I am reading this book with a board, no computer, just going through it. It's still the same thing, copious diagrams and annotations and GM games and evaluations at the end of lines.

2. Turn: Computers actually make the learning process easier in the opening. The process of learning an opening is discussed in many sources but the overwhelming opinion is that the best way to learn an opening is to familiarize yourself with as many games in that opening as possible and then to play 'training' games where you practice the ideas you learned in friendly competition. I have many databases that start from a position and contain roughly 30-50 games from that position and by going through them I learn where to put my pieces, typical attacks and defenses, typical pawn structures, typical endgames, these things come in handy when you are ready to actually try the line out because you know what you are looking at. As far as the practicing component, this is something I usually do in blitz games on ICC, but a lot of people use chess engines for this component showing that the database engine tandem can be very helpful in learning openings.

I have been where you are, overwhelmed by GM evaluations and not understanding where they get their notations from or why they choose specific lines, but by familiarizing yourself with the games and ideas in a particular variation you will start to see why certain evaluations are given and why the GM takes things for granted. Whether you use a computer or a book and a board you still need to put in the time.

On to game analysis:

I concede the point that allowing Fritz to annotate your game without attempting to annotate it yourself is a lost cause and a waste of time, but allow me to salvage some dignity for the machine.

1. Fritz is a BRILLIANT resource. It can and will find ideas that you overlooked in your game, every game for that matter. The fact that you are not able to play certain variations because they are too hard does not mean that you should not familiarize yourself with the ideas of those variations. If it is a tournament game and you spent 4-5 hours thinking about the position, you should be familiar with the nuances in the game and Fritz will show you how they get exploited.

2. Just because you could mess up a position is not a reason to reject the evaluation. We all make mistakes. The evaluation of the game inevitably fluctuates during a game. However, the goal should be to achieve positions that are easier for you and more difficult for your opponent. If you are winning, you have to do less to win than your opponent, that's the definition of a +/=. They are not instant wins, your opponent may even be able to hold on with 20 moves of Fritz play, but you have an easier game.

3. Fritz helps. In the commentary to your post BDK mentions that he is tired of seeing 'Fritz says' in the analysis to a game. I include this line constantly in my analysis so let me explain why. First of all, I do not want to take credit for the ideas that Fritz comes up with, so that's the only reason it is included. I include these ideas because they are important to the understanding of the overall game. I don't add all the lines Fritz spits out or irrelevant moves just because Fritz says it's better. I add lines that COMPLEMENT my analysis. A lot of times there are variations that I avoided or my opponent avoided that are too complex for me to solve on my own. But Fritz can help give me some guidance and 'together', 'we' can work it out. I'll give it a line that it didn't 'see' and after going through some of the variations it will fill in the gaps and confirm my feeling about the position. And sometimes it will give me lines I don't see and help me reject continuations that I found promising. If Fritz hasn't found a one pawn blunder in your games in 6 months you are simply not giving it enough time.

4. It all comes down to work. After Fritz spits out moves I try to put it's ideas into words. You will often see long explanations of positions in my analysis. This is me taking Fritz's += or =+ and trying to give an explanation of the key elements of the position. Working to understand why the computer feels the way it does about this position. Attempting to further my chess understanding by using Fritz as a guide. And if I disagree with Fritz, I have the opportunity to prove it by following it's analysis and offering my own suggestions, which either exposes where I am flawed in my thinking, or exposes that Fritz overestimated or underestimated certain elements of the position. If I had a top-level GM that would sit with me and go through games and give me lines and analysis I would use him the same way, but I think it would cost a lot more.

I'll conclude this point by saying that I recently hired a coach and he asked for a sample of my games so I sent him my ChessBase database where I created the games before they get put into Chess Publisher. I gave him all my analysis. He said "You analyze games like a GM."

So although I understand your frustration with people's reliance on computers, it seems to stem from people who use computers as a shortcut. There are no shortcuts to learning, you must put in the work to see results. And all I can say is that computers allow you not only to keep your work organized, but also to check your work. This is helpful indeed.
 

8 comments so far.

  1. transformation March 7, 2008 at 12:47 AM
    100% amen brother. i use chessbase with Fritz 2 to 3 hours per day most days. great post, thx, dk
  2. Chessaholic March 7, 2008 at 9:26 AM
    cool post man. really neat to learn a little more about your approach to analysis, opening study etc, I'm taking away a couple of good pointers. So you hired a coach? You're gonna kick some serious ass soon...
  3. Blue Devil Knight March 7, 2008 at 9:27 AM
    I think this is very reasonable. Bookup made opening study fantastic.

    Also, when I complained about people saying 'Fritz says' I was specifically referring to when a bunch of kibitzers is watching a big match at ICC or something, these IMs and GMs offering their analysis, and someone chimes in with, "Well, Fritz suggests this..." It's just annoying.

    In practice, when I analyzed my games right, I would spend about an hour going over it on my own, and then quickly with Fritz to see any fluctuations in the analysis, and then, just like you, I would have to put into words why Fritz liked it better. At the time, I had a coach, and if I didn't understand it I would then go and ask him, and if he (an IM) didn't understand it, I figured it wasn't something meant to be seen by a human.

    It is an unhealthy dependence on Fritz that I rail against.No matter what level you are, go over your own game at least once before consulting a computer. Even if you suck, and are leaving your queen en prise, go over it, think about why you left her out hanging in that position, what features of the position tricked your visual system into missing such an obvious blunder, etc.. There is so much to be learned from personal explanation and analysis that it is just kicking Caissa in the nuts to do otherwise.

    We had lenghty discussions of related issues once, quite contentious, over at J'adoube's blog (one Knight Errant got so mad he left the Knights over it!). The issue there was somewhat different, though (there it was whether we should try to emulate computers).

    On that topic, you can find the main thread here, here, and here. I said a little bit here (I pointed out one feature that is very important to human play, flexibility, that would likely not help computers as they are so good at crunching through variations, while our relative suckiness at this means we should often go for the less sharp, even if objectively worse, line). Jim was overall wrong in spirit I think, correct in some details, but it is fun to see a different opinion.
  4. Blue Devil Knight March 7, 2008 at 9:33 AM
    Plus, obviously the better someone is the more they will likely get from Fritz. GMs use computers extensively.
  5. drunknknite March 7, 2008 at 10:37 AM
    transformation - gotta represent ya know :) thanks.

    chessaholic - yeah, i've been skeptical about it but i really wanted some guidance and motivation. i've felt 'lost' lately.

    bdk - that was a cool thread. i like your point about flexibility and 'human' moves vs 'computer' moves. Ironically last night I came up with a solution that was highly inflexible teetering on the edge of defeat but... Fritz says it's ok so I can't complain (there was a better idea that I missed though). We are at an age that you certainly cannot reject computers, they are here to stay, so I say get as much out of them as you can. I agree with you that some people have an unhealthy dependence on Fritz and accept Rybka's word as truth (comme J'adoube) and this is not helpful to chess progress. Just because you can see the moves on the board doesn't matter if you can't understand that and recreate that.
  6. Sir Nemo March 7, 2008 at 2:16 PM
    This post was actually very inspiring. I have always been a firm believer of analyzing your games BEFORE letting Fritz analyze them, but I also believe that premarital sex is probably a pretty big sin, so I don't exactly follow/do what I believe. Perhaps I will start analyzing my own games now. It's really impressive that you were told that you analyze like a GM. I hope that somebody will tell me that someday.
  7. drunknknite March 7, 2008 at 2:49 PM
    nemo - i usually have most of the alternatives picked out before i let fritz at my game, as you get stronger it becomes less a matter of overlooking things and more a matter of how to differentiate between lines and actually make a choice. i would still suggest infusing your own ideas into the game before letting fritz rip you apart. as far as the quote, the fact that it comes from an IM made my day. i was so nervous about giving him all that analysis because i didn't know what he would say.
  8. wang March 11, 2008 at 8:00 PM
    Ok,

    I wouln't say that I berrated them...

    I have a rebuttal to your rebuttal, well, sort of. I'll post it over the weekend. It will be my last post for part 1.

Something to say?