Computer Workout 2.0

  • IM dpruess
  • on 3/5/10, 2:56 AM.

Today I'd like to unveil what will be an extremely valuable training tool for those who choose to make use of it: Computer Workout.

What? You've seen it before? Tried it before? No way! This is a 90% "completely" new Computer Workout, redesigned by yours truly.

So, what have I done? Well, first I deleted 90% of the workout positions we used to have. Next, I inserted rude language into the hints on the remaining positions. Then, I bought a big sledgehammer and destroyed the main database (silly Jay, hiding the server in the cupboard with the cookie jar). Finally, I changed Jay and Erik's passwords so they'll never be able to log in again and fix the site! Amazing right? If you have anything you want destroyed, just send me a private message; my fee is quite modest since I love my work.

Alright, only the first step was true. And actually before deleting the old positions, I had first designed 220 new positions to supplant them. The problem is the old computer workout positions had a lot of tactics in them, a lot of positions with very little replay value once you knew the solution. But I knew that this could be a product on par with tactics trainer, video lessons, and chess mentor. To be able to practice a position endlessly against a very strong opponent? To be able to get best moves suggested? See the evaluation if you want? The set up was fantastic. It just needed the right set of positions. And over the last two months, I have worked on that.

Curriculum/ Use Advice

I've tried to put together a set of positions that will allow you, through practice, to develop many of the practical play skills, endgame technique, calculation, and even strategic ideas that you'll need to be a strong master. There is a fairly strong emphasis on endgame play, which is what this format is best suited for. Of course in addition, you'll want to do some other things-- there are no opening principles or ideas illustrated here, and there is nobody to talk to you about your moves. If you are without friends, you may need to bit by bit draw conclusions about how to play these positions by dint of trying, and noticing patterns of what works and what doesn't work as well. If you do have friends, let me recommend another important practice for you:

Take a position and play it against your friend, trading off colors. Every couple games share your ideas with each other, and argue a little about what you disagree on. And keep playing, until you both start to reach some clear conclusions about the game in progress. It can certainly be valuable to play these positions with our computer or with your friends: the computer will produce stronger moves, and test your level very strictly, but your friends will give you someone to talk about ideas with.

Many of the positions included here I have entered along with "mirror" positions, where you play essentially the other side against the computer (usually with a couple adjustments to make things mildly more pleasant for you and unpleasant for the silicon beast). This is because it is quite important to see both sides of some of these positions in order to really deepen your understanding of them.

How long should you spend on these positions?

When you play a practice game, you should tend to give yourself 30 minutes on beginner problems, 1 hour on intermediate endgames, 1.5 hours on advanced endgames, 1.5 hours on intermediate middlegames, and 2 hours on advanced middlegames. These are suggested maximum times to give yourself, and you should then pace yourself (if you ever intend to play timed chess, it adds a useful practical element).

Some of the positions you will essentially solve on your first try. I have left in a couple positions with a single solution, partly so that you can neither assume they have 1 solution, nor assume they don't. Others you might want to practice for a few hours, others for a month. Some of them will definitely still be rewarding for you on the hundredth play through.


The three levels in which positions are divided can be explained as follows: Beginner is for players under 1400. If you 'master' every position rated for beginners, I am fairly confident you could get a USCF rating over 1300. If you are on the borderline of 14-1500, you might want to go back and brush up on the beginner problems before you get to work seriously on the intermediate ones, which will prove *quite* challenging. Because I would say that a player who has mastered all the positions in intermediate would surely be able to get a USCF rating over 1700 (more likely higher). If you are in the 1800-1900 range, you might want to check your mastery of the intermediate positions before going on to advanced. Remember, the positions have been selected on the basis of providing a fairly comprehensive chess curriculum; if you want to really have good fundamentals, be conservative about how far back in the progression you start, and when you decide you are ready to move on. The final category, advanced, includes positions which I personally have a very tenuous knowledge of (just enough to know that they were very good positions, which I wish I knew). If you could master all the positions in this category, frankly speaking, you'd probably be at a GM level. But I don't expect many people to do that.

One further change that will come to Computer Workout some time in the future is that I will break down the positions into grainier levels, probably 5 or 6 instead of 3. That will make it even clearer what you should work on. But it was a really huge project as it is. On to categories and what I mean by "mastering" the positions.

The positions fall into 7 categories: basic mates, converting a material advantage in the endgame, converting a material advantage in the middlegame, defending worse endgames, playing equal endgames, weaknesses/strengths, and classic middlegames. Here's what you should know about each category:

Basic Mates

These are all positions where you should be able to... give checkmate. Brilliant. This is a good place for beginners to start. Once you can do all the checkmates here, without asking for any "best moves" and without even looking at the little ticker that tells you how many moves left to mate, then you have "mastered" them. Move on to other categories at your skill level (don't jump to intermediate or advanced mates).

Converting Material in the Endgame

The first section I envisioned for Computer Workout. There are a few problems in here which do not have much replay value, but are there because they are theoretically important to the endgame, and I was trying to make the curriculum a bit comprehensive. For Beginner and Intermediate levels, "mastering" a position means being able to win it repeatedly against the computer. At the advanced level there are some positions which are not necessarily winning. "Mastery" here will mean 1) never losing and 2) feeling you have a good sense of the attempts you can make to win. This should provide very good practical experience, as you won't know 100% if the endgame is winnable, but you know you should try to win it!

Converting Material in the Middlegame

Like in the previous category, "mastery" of a position means that you can win it for beginner and intermediate, and in most cases for advanced as well. Many of these positions could stand being played a hundred times. This advice is good for most other categories as well: as you practice, consider writing down some notes whenever something about the position reveals itself to you. For example: "in this position, the d3 pawn always comes under pressure. Rather than passively defending it, I should just activate my pieces, let it fall in many cases, but I can still win with my extra piece. In fact, when my pieces are active, it turns out it's hard for the opponent to take the pawn without conceding something else." You should be able to make many discoveries like this along the way. Putting them into words for yourself can speed along the development process.

Defending Worse Endgames

Defense is often under-emphasized in chess literature and teaching modules. For example in our own chess videos and chess mentor courses. Wink For this part of your training, I can promise... a lot of pain. The computer will be unforgiving of even the slightest inaccuracies, and you'll spend a lot of time losing (even more than when practicing the winning positions from categories 2+3). But bear with it, this stuff is useful. The positions assigned for beginners are all drawable with best play. So "mastery" there is to draw every position (no wins, no losses). Also, all the intermediate ones you can draw... I think. For the advanced there are plenty that are simply losing. Mirror positions of the positions you played in "converting material advantage in endgames." In these positions, be realistic about your goal-- it's to survive as long as possible. How will you ever hold a draw with rook against queen, if you don't develop some toughness now? I have two draws in two tournament games in that situation. It's drawable even against most masters if you have developed the right mentality, and sat on the right side of it in practice a few times.

Playing Equal Endgames

At each level, "mastery" means to move back and forth for up to 50 moves in these positions without losing to the computer. Beginners only have three hurdles to clear here for now. At the advanced level, you'll have to be very careful; many of the positions still have plenty of life in them (GMs have lost these positions in practice before). Nearly all the Intermediate and Advanced positions include mirrors so you can see both sides.


"Your opponent's weakness is your strength; and your strength his weakness." So said an ancient chess sage. Ok, I made that up, and I'm no sage, and it's also not entirely true. But it's mostly true, and so this section includes space advantages, weak squares, pawn structure flaws, and active pieces. For example, your space advantage is your opponent's cramped kingside (weakness). His weak dark squares are the basis for your pieces finding strong posts.

Here there are only two positions for beginners to master-- by winning them of course. The intermediate positions are mostly winnable; but plenty are merely a clear advantage, and you'll find the computer an able defender. Again, mastery should mean that you are "pressing." Trying to win, and shaking your fist at the 'lucky' computer when it slinks off with a draw. All these positions are at least positions where you should:

- understand the nature of your advantage

- be able to come up with a few ideas to try to increase or exploit your advantage

- be able to also list the main ideas for the weaker side

To have mastered one of these should also mean being able to "explain" the position to another competent player who does not have experience with it.

Classic Middlegames

With the exception of one mirrored position for intermediates, this section is only for advanced players. And even these players should not be surprised to be outplayed time and time again by the computer. So what does mastery consist of? Well, as with the weaknesses section, you should be able to describe advantages and disadvantages of either side, relevant plans, typical tactics, and possible courses the game could take. These are definitely well suited positions to study and discuss with a group of peers for 2-4 weeks at a time.

By the way, truly mastering, getting to the bottom of some typical middlegame position may require as much work as getting "some idea" of 10 other typical positions. But you'll find that the process of coming to understand one position at a very strong master level will truly enlighten you as to how to approach any position in the future.

Future Plans for Computer Workout

* I would like to get input from some more, stronger players than myself about what changes they would make to the curriculum (some advice is already incorporated).

* I would like to increase the number of defined levels.

* I would like to occasionally add some extra positions, covering more openings, or perhaps relating to a video series (we already have a set of positions for those who followed IM Rensch's Rook Endgame series)

* I will go through and add hints to the positions. (currently lacking for most positions, because it is easier to add them after the problems are uploaded than prior)

* I would like to get user feedback, and possibly dramatically improve the product accordingly.

* Playlist to give you the precise order I suggest you work in.

* Finally, I would like to find time to practice some more of these positions myself!

A couple disclaimers to be aware of:

- the computer evaluation when a position loads (before you make a single move) is often faulty. don't take it seriously until at least one move has been made.

- the computer has programmed demarcations for evaluations like "equal," "advantage," and "winning." if it says you are winning with 3.01 and then you play a correct move, and the eval switches to 2.84 and it says "oops, you had a winning position," please don't worry about it. we don't have a better way to program that, so just use your own judgment to see if the eval actually fell much.

16674 reads 107 comments
13 votes
more Misc


  • 7 years ago


    "yeah, Narz, that's a very hard one. you'll have to work at it to get better at playing that one; the best move suggestions will not be enough. i would suggest you work on some easier positions before tackling that one."


    My issue isn't with my competency in regards to the problem, just wondering what's the point of "best move" if it's actually not good?  People are going to be clicking "best move" to see what brilliancy they are missing.  When even the computer can't figure it out they likely will be confused & consider their time on computer-workout not very well spent.

    With due respect it sounds a bit patronizing when you suggest I tackle easier problems when I'm trying to point out bugs for you.

  • 7 years ago

    IM Kallatroh

    What engine does this workout use?

  • 7 years ago

    IM dpruess

    yeah, Narz, that's a very hard one. you'll have to work at it to get better at playing that one; the best move suggestions will not be enough. i would suggest you work on some easier positions before tackling that one.

  • 7 years ago


    In this puzzle every time you click "best move" white's position gets worse & worse!

  • 7 years ago


    "show hint" does not work

  • 7 years ago


    I wonder if this might be an idea for the future, but to have a "favourites" section where we can put computer workouts that we would like to visit again and again would be very useful.  

    And this is a long shot - but if we could also put in this section our favourite videos, tactic trainer puzzles, articles etc - it would be so useful to finding material again on that we want to revisit. 

    I currently use a spreadsheet to record my favourite TT puzzles so that I can revisit them, and I'll start doing the same with computer workouts.  If it would ever possible to incorporate this into my page - well that would be fantastic.

    In the meantime, thank you David for your hard work.  This is the greatest chess site ever!

  • 7 years ago


    I very much agree with the idea of opening a path toward adding user-defined positions and analyses (with or without rating) -- positions by strong players I know of are almost always worth my while, regardless of what someone else might have rated them. 

    There are so many able players on this site (many at the FIDE master-level), and although realistically few of us will ever get to the IM/GM level, learning from the analyses and positions of those very strong players on the site is a possibility I wouldn't want to miss. 

    Right now, there are two avenues for such learning:

    (a) by playing 1-on-1 with them (which not many have the chance to do), and

    (b) paying attention to what such players do/say in vote chess games.

    As a coach myself, I know that more (qualitatively not quantitatively) is certainly possible, and I advocate the ability to open up such avenues, without polluting the main curriculum, as David Pruess said.

    I would personally be more than happy to share my collection of 200+ positions (and growing) that I've worked through over a long period of time, and I know many of those people in the first two categories (rated USCF/FIDE 1800 or below) can learn from those positions.

  • 7 years ago


    I never used this feature, but from what can be read here, it's definitively the pathway to chess Masterdom, Thank You IM Dpruess.

  • 7 years ago


    It's too bad the computer goes crazy in the drawn positions.  I have no idea how to program it to do otherwise and act more human without some brilliant, time-consuming programming, though.

  • 7 years ago


    Thank you for making this improvement.

    Is there a possibility for to either improve one of the tools like either tactics trainer or comp workout, so that we can learn openings better?

    I find the game explorer to show me a good idea on how to do an opening, but I don't understand the openings themselves (the ideas behind them strategically rather than immediate results). I don't understand much about positional play except to make my pieces have more space and access to the board.

    It would be nice if one of the tools went over openings, with their name included, and much like tactics trainer, except maybe say something about each move, like one little sentence such as "Bishop should usually go out before knight," or something of that nature to explain why the move is good.

  • 7 years ago

    IM dpruess

    yeah, the mate itself is quite rare. but having a knight and/or bishop is very common, so i think it's a very useful learning tool.

  • 7 years ago


    I think they take rarity of a situation into account as well when considering where to place some endgames.  At least I can do the two bishops, I need the book for knight+bishop.  It's gotta be a downer to draw that situation on the 50 move rule.

  • 7 years ago

    IM dpruess

    the computer runs on servers, not your comp.

    i know a lot of people don't learn two bishops or bishop and knight  mate (which i also put in beginners) until much later. it is my own opinion that this is good material for a beginner to work on, so that they will have better fundamental understanding of how to move and coordinate their minor pieces in all situations.

  • 7 years ago


    I like the endgame situations.  I am reading through 100 endgames I must know and it's nice to practice without being in a rated game :-)

    P.S. IM Silman has the two bsihops mate in Class C not beginner, it is a lot harder than the two rooks.  Even my five year old can do two rooks Smile

  • 7 years ago


    Does the computer run its calculation on computer, or on my laptop?

  • 7 years ago


    Hmmm...thinking of upgrading

  • 7 years ago

    IM dpruess

    slopapa, also thanks for your suggestions, which i like. the goals are described in this guide. in fact in no case should you stop because you won material. converting is the most important part of this learning tool, so keep playing until you'd bet you can win 100%. everyone should know on their own when a position has become a "gimme" for them. and then move on.

  • 7 years ago

    IM dpruess

    Great advice, naughtybishop, I agree.

    tiktaalik, once you have repeated positions three times, declare the game a draw on your own, and move on. i'm very sorry if you moved back and forth fifty times due to my imprecise directions. when you have proven to yourself that you successfully defended a position, you may stop right there. for example, if you are 1800 and trade down to opposite colored bishops and one pawn for the opponent, you do not need to play that out. you know you could draw the position, so once you successfully liquidate like that, count it as a draw, and try the position another time, or move on to a new position.

    ialbrekht is right, there may be one or two positions where the castling permissions got messed up. thanks for the link to that one, i'll fix them as they come up.

    Kallatroh and bondiggity have both correctly pointed out that the computer simply doesn't test you in some of the positions. there's actually nothing to do but to delete those problems, so that's what i will do. i couldn't test my positions with this version of the computer engine until they were uploaded. so i'll be pruning out those problems in the next few days.

    Anlam, thanks for the offer. no offense intended to anyone, but i really wanted to create a comprehensive curriculum, and i thought that was best achieved by one or two people, with a few consultations. in the end i did it myself, consulting with two other stringer players about what they thought of my choices. frankly i think that user submitted problems would dilute the curriculum, and create noise/confusion making the product less usable. however, i had an idea, i could add a separate category "user created," so that users could submit positions. then people could follow "my" curriculum or check out the positions their friends have added. the possibility of rating the problems would be cool, but it's not feasible to have it built anytime soon, I can tell you that without even consulting others.

    Thanks also for the tactics trainer idea. it sounds good, and since i love TT i may advocate for something like that at some point.

  • 7 years ago


    Hey David, 

    Instead of trying to all the work yourself, I think you should let members also edit and add positions. The resulting positions could be rated and classified by the community - this way, poorly prepared practice positions would be rated low and good ones would be rated higher. Sort of a wikipedia approach. 

    By the way, while we are on the subject of community generated content, you should I think also let members add tags to tactics trainer problems that indicate the theme - i.e. decoy, discovered attack, fork etc... 


    Anlam K.

  • 7 years ago

    IM Kallatroh

    Hmm the computer repeated moves even though he believed to have a one pawn advantage. Weird.

Back to Top

Post your reply: