Definitions: Annihilation of Defense

Annihilation of defense: The removal of any pieces that stand in the way of delivering a coup de grace.

I think that's a simple enough definition for something that is often so spectacular...

Lebedev vs Valdaev, USSR, 1930 (ECM position #1746)

Black to move

The pawns in white's castle wall are in the way, so let's just be rid of them! 1. Nxg3 2. hxg3 Qxg3+ 3. Kh1 Qg2+ 4. Bxg2 hxg2+ 5. Kg1 Rh1#

Quite a number of positional factors and tactical features contributed to this combination; the a8-h1 diagonal, the pinned f2 pawn, immobilising the King, and clearance of the h-file making White's back rank weak.

Definitions: Removal of Defender

It's past time I got on with these definitions! Removal of defender is more direct than decoying - which usually doesn't (or only incidentally) involve a capture - and can perhaps be considered a simple form of annihilation of defense which I'll look at next.

Removal of defender: A defending piece is captured. The time lost recapturing leaves an unprotected target which is also captured and lost.

Reinfeld position 475

White to move

The target is the Knight at f4. After 1. dxe5 Black must play 1. ... Qxe5 to maintain two protectors. Then by removing a defender with 2. Nxd6, after either 2. ... cxd6 or 2. ... Qxd6 the Knight is lost to 3. Qxf4 or 3. Bxf4.

FICSing some tactics

Here's a fantastic position I saw in a FICS game last night. We'll skip the early part of the game and get to the interesting bit - far be it from me to admonish anyone's opening play when they can get to this sort of position!

hangbuik (2074) vs Kashniganbaruda (2100)
White to move

White's pieces are clearly better placed, and while both Queens are unprotected, Black's has a defensive role... 19. Rxd5 Bh6+ 20. Kb1 Rd8 this seems like a reasonable idea, surely the Rook must move now? 21. Nf5.

hangbuik (2074) vs Kashniganbaruda (2100)

Black to move

Quite an amazing position! White's Rook and Knight can both be captured by a pawn or the Queen, but his back rank is weak...

If the Knight is captured, taking the Queen or Rxd8+ are both good. After Qxd5 there's a fork with Ne7+, and after 21. ... cxd5 22. Qxd7 Rxd7 23. Re8+ with mate to follow. Meanwhile White threatens both Rxd7 and Nxh6+.

Black thought for a few minutes before wisely choosing 21. ... resigns.

Automatic

I'm about 25 days into my '100 Days to Chess Improvement' now, so it's time to draw a clear picture of what exactly I'm trying to achieve.

Identifying the positional features on the board is relatively straight forward, they tend to be static. It's easy to say "that's an isolated Queen's pawn", because there is no c- or e- pawn. Doubled pawns, closed or open positions, even weak squares suitable for a Knight; once you know what these are, they're just sitting there on the board in plain view.

Not so with tactical motifs. While the presence of a Bishop pinning a Knight to it's Queen is simple enough, the dynamic nature of tactics means you have to visualise the pieces moving round a bit to identify them. Even a simple barrage requires you to count up the attackers and defenders, and play through the different capture sequences to see if something might come of it. So the purpose of attempting to absorb Fred Reinfeld's '1001 Winning Chess Sacrifices and Combinations' is to be able to look at a board and see it's tactical 'parts and actions' at a glance.

Here's a random pretty picture:


You might look at it and instantly see "two animals, 4 legs each, big teeth and claws, orange with black stripes, tigers! (Siberian) day-time, snow on the ground - cold, aggressive poses...". Looking at a Chessboard and seeing it's tactical features should be like this too - automatic.

The first step toward this is to repeat '1001' until I am able to enumerate by name every tactical feature of the position for both sides, even those not featuring in the solution, then visualise all variations of the solution effortlessly. At this point it's not imprtant to find the perfect solution myself at the first attempt, although some assessment of the position, an idea where the solution should lie, is.

Working through the solution, take note of the motifs that were missed - even if the correct solution was found - and when doing the review positions note those that don't spark instant recognition and work through the solution more carefully.

Where do I hope this will lead me?

Firstly a solid foundation for finding tactics in previously unseen positions.

Most of these problems have variations which are either a single long but fairly straight branch, or a bush of many short branches. Solving them develops the visualisation and analysis skills that will help to calculate quickly and precisely in complex positions with more and longer branches at each turn.

When analysing master games, they usually don't bother - and why should they? - mentioning the 'simple' tactical features underlying the positional justification for moves. Tactics are the cog-wheels that whir silently unseen to animate the machina of the game. Tactics are the cornerstone for studying positional play. Tactics must be automatic.

What If?

Reinfeld position #562
White to move

This position looks so crushing for white that you might think it barely needs analysing, look how the major pieces control two open files against Black's naked King! This is the point of studying tactics - we want to find a clean surgical finish that will secure the win now, before Black makes something of his open c-file and barrage along h8-a1.

If the Queen at f6 wasn't there, we could play Qh8+ then Qxg7+ winning a piece on the spot. Likewise, if the Queen couldn't recapture at g7 and the Bishop at f5 could be deflected from defending h7, we could play Rxg7+ Kxg7, Qh7#, and if that's not enough, we'd still have another Rook to throw at him. What we need is control of f7 to complete the immobilisation of the King. Rxg7 Kxg7, Rg1+ Kf7, and the King is running with our loot.

Let's ask the Queen to move with 1. Nd5. The non-capture moves are 1. ... Qf7 2. Qh8# or 1. ... Qd8 losing the Bishop so 1. ... exd5 is forced. 2. Bd5+ controls f7 2. ... Be6 and the Bishop is deflected from it's defense of h7, so now we can stop the King running via f7-e8 with a check at h7. 3. Bxe6+ Qxe6 and we can force the King out into the open 4. Rxg7+ Kxg7 5. Qh7+ Kf6 6. Rh6#.

Solving this position was for me a problem of playing "what if?". What if the Bishop could pass through the e6 pawn controlling f7? What it the Queen could be overloaded and deflected or decoyed from it's defense of g7? What if those almost-mates on the h-file could be brought to bear? What if the Knight which doesn't have a role in those mates could be thrown into the fray? It costs the Knight to prove the Black Queen is immobilised, but before giving up on the notion, let's look at a few variations... if you must take, I start realising some of my "what if?" objectives.

Game Viewer, Diagrams, and Opposite Bishops

I found a nice looking game viewer at Chess Publisher to put on the blog. The move list is clickable, and I should be able to insert annotations into the PGN. To try it out, here's Magnus Carlsen's first round game from Biel 2008 which has just started.

There seems to be a large hole in Carlsen's theoretical knowledge... every patzer knows that opposite coloured Bishop endings are a dead draw. Here he gets into just such an ending, and then proceeds to win it with a nice Zugzwang to finish it off. Ah well, he's still only 17, he can't be expected to know everything just yet!



Just for the record, I've been making the static diagrams with a nice little tool called DiagTransfer v3.0.1.

Chess for Nephews

One of my young nephews has developed an early fascination with our beloved game. Although he does not yet even know the castling and en passant rules, he has an innate instinct for moving pieces to good squares. His Grandma has recently ordered a couple of children's beginner books by GM Murray Chandler from Amazon, and while they are an excellent introduction to tactics, even these are a little advanced for him just yet.

So, I want to formulate a set of simple rules that will take him from 'knowing how the pieces move' to the point where he can start to appreciate some theory.

1. Knight = 3 pawns. Bishop = 3 pawns. Rook = 5 pawns. Queen = 9 pawns.
   
2. Don't move more than 2 pawns before both Knights and at least 1 Bishop have moved.
3. Look for pieces that are unprotected and attack them.
4. Try to occupy the 4 squares in the centre of the board with pawns and Knights.

If he thinks about these things at each move, we can later give some more rules when none of these can be applied. Hopefully this will lead to questions such as "but how do I win the game?" when we can start to demonstrate basic checkmating positions, and from there simple tactics - how many times is a piece attacked? how many times is it defended?

It truly is an ocean from which a gnat may drink and in which an elephant can bathe; the trick is to find a path into it which will maintain his interest without it overwhelming or daunting him. I hope these rules are simple enough without ingraining too many bad habits, the key at this stage is to make pushing wood enjoyable while fostering an understanding of why each move is being made.

Praxis Practice 1

I'm about half-way through Fred Reinfeld's "1001 Winng Chess Sacrifices and Combinations" for the first time now, so about 100 problems attempted once, 100 attempted and reviewed, 100 attempted and reviewed twice...

In practice (by which I mean unrated 2 12 blitz games... hmmm...) it's starting to work; I see the board better, I find some tactics, and when I have winning positions I can find 'finishing' moves. I'm not really winning games so much as being able to capitalise on my opponent's mistakes - in part due to the policy of setting my hair on fire - but the results are starting to appear on the scoreboard.

It's time to move the practical side of this excercise to the next level. The next step is to start playing some rated blitz games, the longish ones that real blitzers hate, so i'll start seeking with 5 12 time controls which is about equivalent to a 15 minute rapid game. This sort of game gives you the opportunity to spend a couple of minutes looking at the critical positions, then blitz on your increments with some understanding of what's going on and a general plan to follow.

One chronic weakness I still have is that while I'm seeing more things I can do, I still miss even simple tactics my opponents can play. One technique I used to overcome this a few years ago was to 'flip' the board - this is not the same thing as 'looking for the penny under the board' when you're losing!

Playing Internet Chess, we become conditioned to play up the board with our pieces starting at the bottom of the screen and theirs starting at the top. Flipping the board and playing top-to-bottom is surprisingly disconcerting, it becomes much easier to get into your opponent's position but proportionally harder to find your own moves. I'll use this technique in 2 12 unrated games to break that psychological barrier.

I'll need to start playing some shortish standard games soon too - minimum 30 30, and at least a couple of those a week. I'll start imposing on some of my Current Affairs team mates soon, when I don't feel like I'm completely wasting their time blundering around.

Reinfeld Position #308
White to move

Oh no! Not again! This one's a bit cleaner - this time Fred isn't even being offered the princess's hand :).

Targets

I have previously described an 'ideal target' as a piece which is both unprotected and immobilised.

Spotting simple cases of immobilised pieces is quite straight forward - they have no free squares to move to. Identifying them and unprotected pieces is becoming part of my routine when looking at the board.

As I begin the forth pass of Reinfeld, I am stopping to deliberately name each feature of the position by tactical motif when doing the review problems - "This piece is pinned, there's a barrage, the King is exposed...", and as I work through the solution "X-ray on the pinned Queen, Black plays Queen takes Queen, White plays the Zwischenzug Rook takes Rook check, then recaptures the Queen.".

Now I want to start relating more complex forms of 'unpotectedness' and 'immobilisation' to the tactical motifs - ultimately they are one and the same thing.

Reinfeld position #20
White to move

The Knight at d5 is unprotected because the Rook at d6 immobilises the Pawn at c6 by a pin, and the Bishop at e6 by an X-ray - if it moves the Queen forks the King and the Rook at f6. This makes it easy to find 1. Bxd5 capturing the piece which has been left unprotected despite having two 'defenders'.

Reinfeld position #880
White to move

Here we see the King is exposed and has limited mobility. 'Exposed King' is another way of saying 'unprotected' since the King requires blockers (ie a castle), 'defenders' are pointless since it cannot be exchanged! If White can control f8 forcing the King to e8, it will also be immobilised whereupon we need only find a suitable check and it will be mate. 1. Rxf4 annihilating the defence (Bd6) 1. ... Qxf4 2. Bb4+ Ke8 and the first goal is achieved. Now if only we can make that check, the cost is immaterial. 3. Qxe6+ decoying the f7 pawn from protection of g6 (clearance of the h5-e8 diagonal) 3. ... fxe6 4. Bg6#.

Pinning is an obvious form of immobilisation. A piece being overloaded takes a little more work to identify, but it unprotects one of it's defensive tasks by being called to act on a different one (the f7 pawn in the second example). Interference is a simple form of blocking the line of a defender so unprotecting its charge, whereas Zugzwang unprotects by forcing a piece to use its mobility against its will.

This relationship between targets, unprotectedness, immobilisation, and tactical motifs needs to be clearly defined and incorporated into the method of looking at the board.

Oh, and I rather stupidly copied something from GM Averbakh without thinking it through for myself and adapting it to the context of this discussion... there's quite an important tactical idea missing.

5. CAPTURE THE ATTACKING PIECE!

Duh.

Definitions: Pawn Promotion and Pawn Breakthrough

These two are almost closely enough related to be the same thing but I want to make some distinctions.

With pawn promotion, we can generally say that a pawn that has securely reached the 7th rank already has the value of the opponent's lowest value remaining piece - the piece that must be sacrificed to prevent the promotion. Considerable material can be sacrificed if it ensures the newly promoted Queen cannot be captured.

Scholtz vs Lorenz, 1964, correpondence (ECM position #798)
Black to play

Black makes a pawn avalanche by 1. ... Nxd5 2. exd5 e4 3. Nc4 ef3 4. Nd2 f2 5. Kd3 f3 promoting.

Pawn breakthrough I want to look at from the point of view of what pawns can and can't do when unassisted by other pieces.

Crippled Majority (Queen Side) and Tactical Breakthrough (King Side)

On the King side White can force promotion with 1. g6 hxg6 2. f6 a decoy, for if Black doesn't capture, 3. fxg7 followed by promotion, and after 2. ... gxf6, the h-pawn is free to run home.

The Queen side of the board illustrates what can't be done. Black's majority is crippled. Even if the Black pawns get to a5, b5, c5, c6 any attempt to break through will fail, provided White is not forced into a capture as this will allow Black to create a candidate passed pawn which is promotable. Notice the a3, b2, c3 formation. If Black ultimately plays ... bxa3 or ... bxc3, White recaptures and Black's a-pawn and two c-pawns are blockaded. This formation occurs from the Ruy Lopez Exchange Variation, and represents White's compensation for Black's two Bishops; White tries to press with his King side majority, and can hold a draw in hand even if he loses his e-pawn.

Definitions: Clearance

Usually referring to single squares, but lines are also cleared so pieces can penetrate the enemy position.

Clearance: Moving a piece so the square on which it stood or the line it blocked can be used by other pieces.

Reinfeld position #548
White to move

White plays 1. Qxe5 clearing the f6 square for the 'Family Fork' 2. Nf6+.

Here's a very famous and entertaining example from Pillsbury.

Pillsbury vs Em. Lasker, Nuremberg, 1896
White to move

21. f5 clearing the f4 square for the Knight 21. ... gxf5 22. Nf4 h4 preventing the Queen's penetration by Qg3-g7 23. Ra1 Be7 24. Rxa4 decoying the Bishop at d7 from the protection of the King side 24. ... Bxa4 25. Ndxe6 fxe6 26. Nxe6 now if say 26. ... Qc8 27. Qxf5 is crushing. Lasker actually played 26. ... Bd7 to prolong the game but couldn't save it (based on Reti's notes from 'Masters of the Chessboard').

Definitions: Discovered Attack

This might be called 'removing a friendly pin'.

Discovered Attack: A piece is moved to unblock a line of attack against a target.

Euwe vs Thomas, Hastings, 1934 (ECM Position #51)
White to move

White plays 1. Bd5 discovering an attack against the Rook at f8 and simultaneously attacking the Queen. If 1. ... Rxf2, 2. Qg8#, and if the Bishop is captured, 2. Rxf8+ followed by 3. Qxg8#.

Definitions: Zugzwang

It really means something like "move-bound" in the sense that the player whose turn it is, is required to move. Chess is a "move bound" game unlike for example Go in which players have the option to 'pass' their move. In any position the side to move is technically 'Zugzwang', but that is not what we usually mean when we use the term in Chess.

Zugzwang: The player to move does not face a direct threat, but any move he does make will result in a loss of material.

"Trebuchet"
Whichever side is to move, loses

Zugzwang can occur in the middle game too.

Reinfeld Position #860
White to move

The Knight is pinned, but it is both attacked and defended 4 times. After 1. h4 Black can play a few pawn moves but then he is Zugzwang - playing either 1. ... Ke8 or 1. ... Qe8 leaves the Rook at c7 inadequately guarded, and White will play 2. b5 winning a piece.

Mortal Kombat

I've never understood why many of the citizens of FICS so despise disconnectors. It's my favourite way for a game to conclude, and it happens more frequently when you find a killer tactic to finish them off. One of my as yet unrealised FICS ambitions is to have a game history full of wins by disconnection, the best I've managed to date is 5/10.

Resignation is one of Chess' more or less unique features as a game. It's a way of saying "Ok, you bested me in this game, but at least I'm good enough to know when I'm beaten". They can go away to lick their wounds, and come back to fight again another day.

As for disconnection... Oh! you can feel their anguish! You just know that somewhere on the other side of the globe, a mouse just got hurled across the room, a boot just got thrust through a monitor. "I like the moment when I break a man's ego." - Bobby Fischer. Resignation? that's a sober self-reflection. Disconnection? Well, this is Chess, this is Mortal Kombat.

Flawless victory! FATALITY!!!

Setting your Hair on Fire

While my results aren't improving yet, something about my game is, and starting to see tactics again I have rediscovered a joy in the game I haven't experienced for a long time.

When things start looking grim from your side of the board, one thing that can be done is to play anything that maximises the activity of your army, create wild complications, disregard conventional material evaluations, and give your opponent the best possible opportunity to go wrong. This strategy is known as 'setting your hair on fire', you light your curly locks then run around screaming and madly waving your arms in the air, hoping the other guy will catch fire and burn to death before you do.

My present approach is to apply this strategy preemptively. While sacrificing a whole Rook for a taste of the initiative may sound a little expensive, why wait for things to turn sour before starting the fun?

In his classic work on conflict management (the title of which is often translated as "The Art of War", I prefer something more like "The Tao of Conflict"), Sun Tzu advises that when encircling an opponent, you should always leave them a single escape route. The reason for this comes down to animal nature; when there are thoughts of leaving the battle and fleeing to safety, they won't put so much energy into the fight. When there is no escape but through victory in the battle, there is no option but to fight 'tooth and nail'. Kill or die.

So it is that I am starting each game with the idea of forcing myself into a desperate situation. "No quarter will be asked, and none shall be given!". My results are on a 'J' curve (he hoped), but it makes for some fun and interesting games - and it compels me to seek out and exploit the tactical possibilities of the position each time I play.

Progress Report 1

I've worked through a couple of hundred of Reinfeld's problems now, so some observations and a look at whether we're getting anywhere.

I'm spending far too much time on some problems and not enough on others. At this stage, it's not necessary to find the perfect solution before going to the back of the book, it is enough to have at least found the key ideas in the difficult positions. In some of the simpler problems, I have found the correct solution but missed a defensive idea for being too certain I recognise that pattern. It is necessary to take a little more time looking around the whole board in those positions. The objective at the moment is to acquire a database of those 1001 combinations, and be able to visualise their solution at a glance. The next step will be to apply that learning to positions I haven't seen before and produce perfect results every time :)

I'm still mostly playing unrated blitz games (bad!) but I am looking at the board and thinking "what are the targets?", "which pieces are unprotected?", "identify the tactical motifs... pin there, barrage there, fork possible here..." (good). The internal dialogue is still messy, I need to more clearly define the method.

This was a blitz game played at 5 min + 5 sec increments. One of those Lopez Exchanges, where Black plays Bg4 and dares White to open the h-file against his own King. I left the offer open for 11 moves before it was finally accepted - there's a time when you can do it, but the details can be complex to calculate in a blitz game!

RandomGuest vs SimianChatter (blitz)
Black to move

White's back rank is weak. Nc4 prevents the queen penetrating at d2 to cause death and mayhem. I'm going to send in a decoy! 1. ... Rd2 2. Re2 Rh1# if 2. Kg1 I get to sac the other rook 2. ... Rh1+ 3. Kxh1 Qh4+ 4. Kg1 Qh2+ 5. Kf1 Qh1# 2. Nxd2 Qxd2 3. Resigns.

The first noticeable result from what little tactics practice I've done so far is that I'm starting to find clean, clinical finishes. There's a long way to go yet, but just maybe I'm beginning to see the board like a Chess player again.

The Secret Computer Tournament

Time out from studying tactics for a moment for a quick thought about computer Chess. Until the early 1990's computer Chess was seen as an interesting field of research, but the amused wisdom of the experts was that they'd never compete with human GM's. Then they started beating... Garry Kasparov.

Now they're generally regarded as all seeing all knowing iron monsters - which they are not. It is said a player's reputation can destroy his opponent before he even reaches the board, and computers do enjoy a large psychological advantage over their human opposition (for a start, they don't have a psychology to upset!). Humans play noticeably differently against machines, simple example - Kramnik 'forgetting' he'd left mate-in-1 on the board in his match against Fritz, a mistake he'd make against no human player.

So how do we determine a computer's true playing strength with respect to human GM's, that is, have them play humans without their 'unfair' psychological edge? Here is an idea for a tournament I thought about to address that problem. There are some obvious flaws, but it's just an idea...

Arrange a tournament with players of the same supposed strength as the computer. Before the tournament begins, one of the players is selected to be the computer - the other participants know this will happen, but only the selected player is told. It is the task of this player to act out playing their games while they are secretly fed their moves from the Chess engine.

At the end of each round each player will be required to guess who the computer is (this could make an interesting competition for the spectators too). At the end of the tournament additional points can be awarded for making the most correct guesses, and also for players who beat the machine in their individual game. This adds an element of poker to the event, for example a player would be allowed to nominate them self as the computer. It might also contribute to the discouragement of uncontested draws with the players being unsure who they need to beat for bonus points.

Shootout

I like watching FICSers do battle... it's 'everyman' Chess. Here's one I just saw that nicely illustrates compounding tactical themes one upon another.

smallblackcat(1967) vs Juffi(2007E)
Position after 19. ... Nc4

20. Be4 pinning the unprotected Knight to the Rook. 20. ... Ncb6 21. Ba5 aiming to remove the defender. The Knight is not quite overloaded yet, but Black is teetering on the brink. 21. ... Rab8 the target of the pin removes itself from the line of fire to support the defense. 22. Bxb6 Nxb6 23. Bc6 simultaneously discovering an attack on Be7 while taking aim at it's defender. Black can only defend through counter-attack 23. ... Rec8 the mutually unprotected Bishops allow White to harvest his first fruit. 24. Bxb5 now Rc2 immediately, or axb5, Rxe7 Rc2 might be Black's best, he decides to try and shoot it out instead. 24. ... Bf6 25. Bxa6

Position after 25. Bxa6

25. ... Bxb2 26. Rab1 with a two-fold attack against Bb2 and Rc8. 26. ... Rc2 27. Bd3 renewing the two-fold attack. 27. ... Rc3 again attempting a defense by counter-attack 28. Rxb2 Resigns. After 28. ... Rxd3, 29. Rxb6 wins the Knight because the Rook at b8 is overloaded defending both it and the weak back rank.

Trying the Method

The solution to this problem didn't jump out at me, so let's try working through the method on it.

Reinfeld Position #550
White to move

What are the targets?

• Rook at d8 is unprotected
• Bishop at f8 is immobile
• King is exposed

What are the immediate threats?

• hmmm... nothing obvious

What are the potential threats?

• d6 is advanced and might be promotable
• Nxf6 further exposing the King
• somehow annihilating f6 and g5 then playing Qg5+ forking the Rook
• something like Rc2-c7 aiming the Rook at the King

Still no obvious clear win. Let's list the positional features.

• Black is already a pawn up
  
• Rook barrage on the d-file
• Black King insecure
• Open c-file
• Advanced and passed but isolated d-pawn
• Black controls half-open a-file

Can we extract some candidate moves from all this?

• Nxf6 exposing the King, maybe in conjunction with Ne4+ aiming at Nxg5 or Qxg5
  
• Rc2 aiming at the Queen and to control c-file and 7th rank, opens c1-h6 for the Queen
  
• Nxg5 exposing the King and looking for Qxg5
• d7 makes way for Nd6, the Rook needs to be moved to promote it though
• Qb1 or Qc2 might put the Queen on a better diagonal
  

Ok, a few ideas there... to the analysis of variations then.

1. Nf6 the most violent idea because I know I'm looking for a combo... so method-wise probably a bad idea. 1. ... Kxf6 we'll also have to consider whether this capture is forced. 2. Ne4+ K?? 3. Rc2 Q?? 4. Q or Nxg5. Need to replace the '??' with actual moves.

1. Rc2 not forcing a reply... 1. ... Qa4 threatens to pin with Bb3 black has other moves too.

1. Nxg5 fxg5 2. Rc2 Qf4 looks reasonable for Black, Bb3 coming again.

1. d7 lets Bf8 move again. 1. ... ??? 2. Nd6 Bxd6 3. Rxd6 ??? 4. Ne4 Qxc1 5. Rxc1 might be onto the f6 pawn... not very concrete.

In fact the whole thing's a bit flaky... all those ?? null moves. Kotov in "Think like a Grandmaster" says you should look at each variation once and once only, and be able to analyse it quickly and accurately. Lot's of work to do there!

Still, Nxf6 looks most promising, let's try to fill in some blanks.

1. Nxf6 Kxf6 2. Ne4+ the King has e6, g6, g7, f5

2. ... Ke6 Qxc4+ winning
2. ... Kg6 3. Rc2 hitting the Queen and discovering a barrage mate threat against g5, winning
2. ... Kf5 3. Rc2 Qa4 (Qe6 4. Qxg5#) 4. Qxg5+ Ke6 5. Qg4#
2. ... Kg7 3. Rc2 the Queen has e6 or a4 (or b3)

3. ... Qe6 4. Qxg5+ and Qxd8 winning
3. ... Qa4 4. Qxg5+ Kh8 5. Qxd8 threatening Qxf8 Bg8, Qf6# so Black has no time for Qxc2... winning.

There were two lines I couldn't see all the way through for myself:

3. ... Qa4 4. Qxg5+ Bg6 5. Rc7+ Kg8 6. Qxg6+ hxg6 7. Nf6+ Kh8 8. Rh7#
3. ... Qa4 4. Qxg5+ Bg6 5. Rc7+ Rd7 6. Nc3 and the overloaded Queen has to leave the protection of the Rook.

Solving these problems is one good way of practicing analysis skills, but without the skills you can't solve the problem and you end up running in circles again - knowing you're close to a solution but missing a key point in one line (ie. 6. Nc3) you have to reject the whole idea, especially in a real game.

Conclusion. Practice raw analysis of complicated positions too!

Oh yes, and I still didn't consider alternatives to 1. ... Kxf6. Neither does Fred in the solution... and really the whole thing's still a mess anyway; in the second last variation, we'd have to look at 5. ...Kh8 when there's no check with 6. Qxg6. My drawing board is very busy at the moment!

States and Exploits

All these definitions will never fit neatly. Probably because there is no clear dividing line between one motif and another. Maybe I should start looking at which are states and which are exploitations.

So for example being overloaded is a state of trying to defend too many points with too few pieces, and a decoy is what we call the piece that exploits that.

The pin is another state that is exploited by various means, such as a barrage against the pinned piece... but a barrage can be a state too. Back to the drawing board...

Definitions: Overloading

Overloaded pieces can be difficult to spot, sometimes it's a fine line between a piece operating energetically and it becoming overworked. The overloaded piece will usually be somewhat immobilised by it's juggling act.

Overloading: Applying pressure to a piece that is performing two or more simultaneous defensive tasks.

Reinfeld Position #426
White to move

The Black Queen is overloaded defending the Queen + Rook + Knight barrage against g7, while also defending the weak back rank at c8. White uses the second threat to decoy the queen away from defending the first.
1. Rc8+ Qxc8 2. Qxg7+ Rxg7 3. Rxg7#

Definitions: Decoy

A bit like overloading, but here there might be only a single defensive task which cannot be maintained.

Decoy: A defending piece is distracted from it's task by the decoying piece.

Averbakh Position #144
Black to move

White's Queen is tied to the defense of the mating threat at f2. 1. Ba6 using a pin by the Bishop as a decoy, Black forces the Queen away from protecting f2.

Check, Check - Check, Check! Check! Check!

I didn't include Double Check in the list of motifs, it's a special case of Discovered Attack, but here's a pretty position on that theme:

A. White 1919
Mate in 12

Definitions: X-Ray Attack

Looks similar to, but is distinct from pinning (see previous post).

X-Ray Attack: A higher value piece is attacked. If it moves, a lower value piece on the same line is exposed to attack.

Reinfeld Position #757
White to move

White plays 1. Rh8 a pawn promotion threat, 1. ... Rxa7 2. Rh7+ when the King moves, the X-Rayed Rook is captured.

Definitions: Pinning

The idea is easy, exploiting them apparently gives me more trouble than i should like.

Pinning: A lower value piece is immobilised by being in front of a higher value piece on the same line of attack.

The pinned piece might also be referred to as a 'blocker', see point 3 of 'What can be Done'.

Note also the difference between a Pin and an X-Ray; with an X-Ray the higher value piece is in front, if it moves, a lower value piece behind it is exposed to attack.

Reinfeld Position #1
White to move

Black's Queen is pinned to the King. White exploits this with 1.Rd8 an X-Ray attack, 1. ... Qxc4, 2. Rxd8+ a Zwischenzug.

A Pin is more a positional feature than a tactic in it's own right, since an additional motif is usually required to exploit the pin.

Definitions: Barrage

I think we will call it this... this is the most basic type of Chess tactic.

Barrage: An attack on a piece by more pieces than are defending it.

Reinfeld Position#188
White to move

By 1. Qe4 White combines the mating threat Bishop + Queen barrage at h7 with a barrage against the Bishop at d4 by the Queen and Knight.

Barrage is frequently seen as a positional idea on a particular line or against a particular square. Alekhine and Reti both had a type of barrage named after them, known as 'guns'.

Alekhine's Gun (White) and
Reti's Gun (Black)

Definitions: Forking

Generally this is thought of as something Knights and Pawns do, but all pieces can fork.

Forking: The simultaneous attack on two (or more) pieces by one piece.

Averbakh Position #38

White to move

White plays 1.g8/Q+, Kxg8 2.Bd5+ forking the King and the Rook.

List of Tactical Motifs v0.1

Here is my list of tactical motifs; any and every combination should be able to fit into one of the categories (I'm not asking for much am I?). I've used Informant's "Encyclopaedia of Chess Middlegames" as the basis of the list; each one will need a clear definition and illustrative example.

1. Double Attack
2. Forking
   
3. Discovered Attack
4. Pinning
5. Deflection
6. Decoying
7. Interference / Interception
   
8. Annihilation / Removal of Defender
9. Clearance
10. Blockade
11. X-Ray Attack
12. Overloading
13. Zwischenzug / Intermediate Move
14. Pawn Breakthrough
15. Passed Pawn
16. Demolition of Pawn Structure
17. Pursuit
18. Zugzwang
    

Of course there are problems with this list before I even begin... the redoubtable Mr. Averbakh again points out that all tactics are a form of (1) Double Attack. Indeed there are even 'positional tactics' where by threatening to gain one type of positional advantage (say a doubling of the opponent's pawns) the defense allows you to gain a different positional advantage (say control of an open file). Perhaps we will rename this one 'Barrage' - attacking a piece with more pieces than defend it, as this most simple idea isn't explicitly covered in the other categories.

Zugzwang and clearance have an interesting inter-relation, we might call them both 'tempo tactics' at opposite ends of the scale; with Zugzwang you compel the opponent to gain a move to his own disadvantage, with clearance the piece that moves to clear a square or line usually does so compelling the opponent to defend a threat that effectively loses a move with respect to defending a secondary threat. Clearance and discoveries are also closely related, clearance usually referring to position (squares and lines) whereas discoveries usually refer to material.

The list needs a little work, it could perhaps be more concise, but it's a start... now to the definitions!

What can be Done

Some more wisdom from GM Yuri Averbakh, who manages to state simple things succinctly in his "Chess Tactics for Advanced Players". What to do when a piece is attacked?

There are 4 and only 4 possibilities:

1. Move it
2. Defend it
3. Block the line of attack
4. Counterattack something of equal or higher value

You might think 5. Ignore it! is also an option, but this is really a special case of point 4. This is the one that usually gives me most trouble; from my earliest days as a Chess player I was warned "cross-attacking is always dangerous!". Usually it is, but good tactics, as with most things, comes from understanding when 'the exception proves the rule'.

The Method v0.1

The idea is to go through a controlled internal dialogue during a game. This is the one I'm developing to hone tactics, and when it has been so often repeated that the script plays automatically, a new one can be developed for other aspects of the game.

1. What are the targets?
   - Unprotected pieces
   - Immobile pieces
   - Overloaded pieces
   
   
2. What are the immediate threats?
   - Checkmates
   - Immediate captures
   - Forks, Skewers, X-Rays
   
   
3. What are the potential threats?
   - Possible checkmating positions
   - Pins, Discoveries, Double Checks, Overloading, Interference, Clearance, Trapped Pieces, Decoys, Pawn Promotion
   - Exposing the King
   

The exposed King is a special case of an unprotected piece. You have to defend / block (ideally with pawns) the squares around the King, not directly protect the King's square itself.
'Overloaded pieces' are those that have more than one defensive task, I'll try to elucidate that when I do the definition of overloading.
The purpose of the definitions is to ensure clarity about what is being looked for ("find an overloaded piece"... "sure, what does overloaded piece really mean?") and the tactics practice reinforces what it 'feels' like. Theory and praxis.

The list of motifs needs to be formalised; ie forks are really skewers by a Knight, discovered check is really just a special case of discovered attack.

Definitions: Interference

Let's start with interference, since this is a relatively non-trivial concept compared to say pinning and forking.

Interference: The interfering piece moves with it's own threat forcing it's own capture. The capturing piece blocks the line of one of it's own pieces which is defending a different threat.

Reinfeld Position #583

White to move

White plays 1. Rd5 threatening Rxd8 mating.
1. ... exd5 interfering with the Queen's protection of the Rook, 2. Qxd8#
1. ... Rxd5 interfering with the Rook's protection of the back rank, 2. Qe8#
1. ... Qxd5 interfering with the Queen's protection of a1-h8 diagonal, 2. Qf6#

Establishing the Method

Before considering variations and moves, some things should be considered at every position.

What are the targets? Targets are:

• Unprotected pieces
• Immobile pieces
• Overloaded pieces

Enumerate potential mating positions!

Internal dialogue should proceed "which pieces are unprotected? which pieces are immobilised? what threats are latent for each side?" then run through the list of tactical motifs. Proactive. Not "omg, what should I try next?", followed by "what was I thinking in that game?".

Some definitions are in order. Basic checkmate patterns are assumed. From them grow tactics, from tactics comes strategy.

What are strategy and tactics? I like GM Yuri Averbakh's definition for conciseness and simplicity.

Strategy is what to do.
Tactics are how to do it.

I'm leaving the 'what' to one side for the time being... learning 'how', and hoping to be promoted to management later. As to definitions, I'm working on clearly defining and illustrating the different motifs.

Fred Marries the Princess

At the end of the last post I gave position #296 from Reinfeld to illustrate a few points I've been making. I spent rather longer than I would have hoped on this one, and here's how I approached it...

It's a 'discovered attack', so obviously Nd4 moves somewhere. Nf5 looks like a good start, protecting my Be3 and attacking Bg7 and Qb6 (the discovery). So, 1. Nf5 Queen moves, 2. Nxg7+ is a piece. Easy. Hmmm... Nxg7+ Kf8, better get the Knight out... but all his escape squares are covered, Black gets the piece back. And where did the Queen go when she moved? Oh, Nf5 Qxb2 attacking my Nc3. Nf5 has holes in it, let's look for something more clear-cut.

So... Nb3 covers the b2 pawn... but then just Qxe3. Ndb5, same.

Ne6 maybe? Threatens the fork so Qxe3 2. Nxc7+ Kd8 3. Nxa8 gets the exchange... but something like Qxf4 immediately, and the Knight's got Bukley's chance of getting out of the corner surely? nope, can't see a way, don't like that.

Ok, something else... save the discovery and play something else first... 1. Nd5 hit the Queen with the other Knight, looking at 2. Nc7+. Just Qd8 might not be best, but it defends the threat and then where's this 'discovered attack'?

Back to the Nf5 idea then...

And so young Sim ran in circles until conceding defeat and looked up the solution (translated from English to algebraic notation):

"White's discovered attack wins the Queen or forces checkmate: 1. Ne6! Qxe3 2. Nd5! Qe4 3. Ndc7#" - Fred Reinfeld.

Brilliant! 1. Ne6 immobilises the King, then we just put the other one on c7 by hitting the Queen again via d5, GAME OVER! There was a fault in my thought process in that I didn't notice the King's immobilisation by Ne6, and I didn't combine the Ne6 and Nd5 ideas.

Wait a minute! Did I say vague, Fred? You didn't give the solution to 1. Ne6 Qxb2. That must so obviously win the Queen that the line doesn't rate a mention, right? Let's check. 1. Ne6 Qxb2 now both my Knights are en prise... where's the killer? 2.Bd4 re-setting the discovery? 2. ... fxe6 3. Nb5 Qb4+... maybe I get the exchange at the expense of trapping my Knight in the corner again, whatever Black seems to be punching harder than I am. Where's this Queen win?

I'm sure I'll have more to say about using computer Chess engines in the future, but for now suffice to say that if this is a simple "wins the queen" for a human, Crafty will find it in no time, right?

From the diagram, Crafty whips off a search of a million positions in a few seconds.
Evaluation +0.90, depth 8 ply (4 moves)
1. Nf5 Qxb2 2. Nxg7+ Kf8 3. Nd5 Kxg7 4. exd6 Qa3 5. Nc7 Qxe3 6. Nxa8 exd6 7. Qxd6

Not awe inspiring, no Queen win in sight, but Crafty is a 'quick and dumb' engine, it's being thorough. Give it a minute at least...

Evaluation +0.99, depth 9 ply, 5.171MN (about 5 million positions searched)
1. O-O dxe5 2. Nf5 Qxb2 3. Nxg7+ Kf8 4. Ne6+ fxe6 5. Ne4 exf4 6. Bxf4

Erm, yes, +1 is certainly winnable, but this is the solution to a problem combo? Time to get ready for work, and 15 minutes later Crafty reports

Evaluation +2.03, depth 12 ply (6 moves), 55.657 million positions searched
1. Nf5 Qxb2 2. Nxg7+ Kf8 3. Nb5 Kxg7 4. Bd2 Nd4 5. Bc3

Let's leave Crafty running while we're away and ponder an old Chess story which no one seems to know, but everyone seems to know the ending of. By back analysis, we can determine the player of the white pieces to be the King, and the player of black to be his daughter's suitor. The King applies a tactic which befits a man of his status.

"You may marry the Princess my son, provided you swear an oath never to take the Queen Knight's Pawn with your Queen!".

But our man Fred was right all along as we return to discover that after about 6 hours Crafty found

Evaluation +2.61, 11.912GN, 16 ply
1. Ne6 Qxb2 2. Nxg7+ Kf8 3. Nb5 Kxg7 4. Bd2 a6 5. Bc3 Qxa1 6. Qxa1 axb5 7. exd6+ f6 8. dxe7 b4 9. Bd2 Nxe7 10. Bxb4

Yes, that's nearly 12 thousand million positions searched, and if the Prince takes the Queen Knight Pawn, there is a... erm... 'simple' win of the queen. But it's Fred who marries the princess after all. 13 hours and the Queen Knights Pawn didn't tempt him once...

Evaluation +2.28, 24.232GN, depth 17 ply
1. Ne6 Qxe3 2. Nd5 fxe6 3. Nxe3 dxe5 4. fxe5 Ndxe5 5. Bb5 Bd7 6. Qd2 O-O-O 7. O-O-O Kb8 8. Bxc6 Nxc6

Position after 8. ... Nxc6

We all saw that, didn't we?

Tactics!

I've completed the first, and am about half way through the second pass of Reinfeld's 1001. You wouldn't think exploiting a simple pin would give me so much trouble!

There are pros and cons to having the tactics organised by theme. When you look at a position, you're inclined to start thinking "so, how do I find a winning move that's a discovered attack?" - then you either find the solution instantly because you were given a clue, without considering the whole of the position, or you run round in circles trying to find the discovered attack, only to realise it was simple, but you had to 'remove the guard' first. All but the most trivial of combos involve several motifs combined. Makes for lazy thinking.

When you are presented with "White to play and win (theme: discovered check)", you know there is a way for white to win... this of course is not so in a real game. A search method needs to be developed to determine the likelihood of there being a tactic, and if there is, of finding it. Breaking down tactics into themes provides a checklist of ideas to run through in analysing the position.

So, before we madly start throwing pieces around the board looking at variations, we need to make some assessments. What are the potential targets?

Target: an ideal target is a piece which is both immobilised and unprotected.

An immobilised piece is a weak piece, an unprotected piece stands on a weak square. Ennumerate the targets!

But a target is only truly 'weak' if it can be attacked. In the starting position of Chess, all four rooks qualify as 'weak pieces on weak squares', but there is no combo from that position! Next we might start looking at what attacks can be made... "Nd5 attacks the Q at b6", "If I could maneuver a Rook or Queen to h8, I have a basic mating pattern"... drawing up a pool of candidate moves.

Then themes might come into play... "if I push e5, my B at g2 makes a discovered attack on Ra8", "the f6 Knight is pinned to the Queen, can I add pressure to the Knight or deflect the Queen?".

Only then, when the ideas contained in the position have been fleshed out, and a list of candidate moves established, should the analysis of variations begin.

Some such method needs to be formalised, and so ingrained into the thought process that it becomes automatic... 'sight of board'.

Reinfeld Position #296
White to play and win
"Discovered Attack"

Back to Basics

What is the basic toolkit of the amateur Chess player? A bit about the openings to be sure - 'sortez les pieces' is enough... get the piece out! Shove the King towards the centre in the endgame, this is generally a good idea.

But 95% of all games below Master level are won by statistics!

Let me repeat that. 95% of all games below Master level are won by tactics. Tactics! Tactics! Tactics!

Until basic tactics are second nature, that combination, the 'crowning glory of good positional play' will not materialise while it remains unseen.

Fred Reinfeld's "1001 Winning Chess Sacrifices and Combinations" is divided into 20 chapters by tactical motif, with 6 diagrams to a page. A week or so ago I began on the grand plan.

• Work through the first page of each chapter, then compare your analysis to the solutions.
• At the second pass, review the first six problems, then solve the next six.
• Continue the cycle reviewing all previously attempted problems, until all 1001 problems have been completed.
• Start at problem 1. Review all problems in the book through to 1001.
• Do it again.
• And again.
• And again.
• DO IT AGAIN.

This should breed a familiarity with each problem, each time revealing a few more of it's secrets (and let's face it Fred, sometimes your solutions can be a little... well.. vague!) until it should be possible to pass through the whole book in around an hour. This is will now become a once or twice weekly tactics training session, and a quick tune-up to get into the right frame of mind before a game.

Backward - If you don't Use It, you'll Lose It

Some men in white lab coats did some stuff with Chess players involving lots of electrodes and clipboards. What they found is that Chess players process Chess patterns with the same part of the brain that the rest of us use for recognising people's faces.

A few years ago I'd studied combinations daily for six months and they would just leap out of the board at me. I once heard a Master (who's name I regret escapes me) say "Tactics must be automatic, you spend your time looking at strategy." and that is how it was; Chess patterns came alive of their own accord while I thought on other things.

So you will understand what I mean when I say the 'rust' has set in.

"I remember you, didn't we meet at that party a few years ago? It's Tom isn't it? Oh no, Ted? Ah! That's right, nice to meet you again Mary!".

That's how it now feels looking at the Chess board. There's something familiar about that piece configuration, but I can't quite put my finger on it... does the Knight go here and then the Rook takes that pawn or...