Retros for Dummies

It's been suggested that we need a thread for introducing n00bs to retrograde analysis. So here it goes.

What are 'retros'?

"Retro" is short for 'retrograde analysis' and it is generally any chess problem in which you need to determine what happened in the past - the moves that must have occurred before the given diagram.

How do I solve retros?

You solve them through deductive reasoning. I will show some typical tools later on.

But, but, the positions are so ridiculous! Those would never happen in a real game!

Retro fans know that, and we don't care. This is a different aspect of chess than the usual white vs. black competitive game. All we care about is whether or not the position is legal.

Now that we've got the mini-FAQ out of the way, on to...

Stupid Retro Tricks

(Basic tools needed in practically every retro.)

We solve retros by taking back moves instead of playing them normally. Going in reverse is a bit strange at first. Here are the tricks needed to do it successfully.

Trick #1: Uncapture

People are so used to the normal flow of the game where pieces are never added to the board. In retros, we can do just that!

White is in check, so Black must have moved last. But where did the Rook come from to give the check? Clearly not from the d-file, because it is blocked in both directions. But what about the 2nd rank? I can hear people screaming now "Even I know that white's King is still in check, and that is not legal with Black to move!"

This is where the uncapture comes in. We can't leave white's King in check, so why not place a piece on d2 and say that black just captured it? That solves the dilemma.

Looks so obvious when playing forward, doesn't it?

So what we really need is to train the retro-vision. Look at the move from the end and practice taking back the move until we get used to the visual effect of uncapturing.

Hit the back arrow from this diagram a few times until you get used to the idea of the white Q popping out from under the Rook as it moves away.

Stay tuned for more Stupid Retro Tricks...

Stupid Retro Tricks

Trick #2: walking into check

Like trick #1, it's not so much that this is anything earth-shattering; it's just that it looks damned weird when you do it backwards.

Trick #2 is: You may walk a King into check, so long as the other side can immediately get that K out of check.

The trick is handy in positions like this.

OK, so white must have moved last if it is stipulated that Black is to move. But all white has is a King! Surely it can't move, being pinned down by the Rb2 like that.

Except that it can during a retraction. Just fearlessly walk him into check - put the K on a2 - then black gets him out of check right away. Rb2 goes to b1.

But the K is still stuck between the 2 Rooks! - I hear you scream. No - rememeber what we just talked about. Walk through check again. Ka2 to a3 - then Rb3 to b2 to remove the check.

Let's see that first in normal forward motion:

Nothing terribly unusual about that...just a lone white K being mated in a scholastic chess game.

Now, let's look at it in Retro-Vision®:

Hit the back arrow a few times to see how it looks playing it backwards. Do this until it sinks in!

Remember -

You can walk a K into check so long as the enemy immediately un-checks!

Me likey. Would I that someone had told me this earlier; playing retromoves is something I only recently wrapped my head around.

It might also be clearer to explain the above are tricks to "rewind" moves from a given position, or to play retromoves, which is a convenient visualisation method to see how a given position arose, rather than trying to reach it from normal forward play. Of course the normal laws of chess apply in forward play (no walking into check!)

Good point. Added a paragraph about taking back moves instead of playing forward.

I wouldn't say "Stupid Retro Tricks", as it sounds really derogatory.

I'm not quite sure whether I should introduce this early or not, but in any case...

Where was the rook before the last move?

Clearly the last move is White's (Black is in check), and the king cannot discover the check. So the rook moved on the last move.

Suppose that we use Trick 1 to capture something on h8. But then the rook still comes from somewhere on rank 8, already checking Black. It's impossible!

...but it isn't, due to the strangest piece in chess...

A pawn has the most "special" rules. While almost all other pieces only define their movement pattern and that they capture identically as they move (only king and rook has extra rules on them), a pawn has a bunch of other rules stuck to it. Two-square move on the first move, capture diagonally, en passant, promotion...

Trick 3: Unpromote

A piece might be a pawn that is promoted. With this hint, can you solve the above again?

So, what if the rook was once a pawn?

If before the last move, the rook was still a rook, the same impossible situation happened; the rook came from rank 8 where it would have checked Black already. So before the last move, the rook was a pawn! But how can a pawn get to h8? By capturing from g7 to h8, of course! (h7 is occupied.) So the last move was g7xh8=R+, and so before the last move, the rook came from g7.

As you might have noticed, retro problems are simply regular chess problems, but viewed from an entirely different viewpoint, where you don't analyze the future, but the past. This skill is never used in actual games (you know exactly the past), and is rarely used in regular chess problems (probably coming as tricks later), but there are retro enthusiasts, just like there are chess grandmasters and grandmaster problemists.

Now there seems to be enough information to introduce this topic, I guess.

Trick 4: Possible and impossible checks

Now that we're familiar with some of the usual retromove oddities, we can start to investigate some tougher oddities with checking. An impossible check is one that could not possibly have been given by a legal sequence of forward moves, or equivalently one where there is no legal retromove to retract. (Remember, the checking side must immediately retract a move to eliminate the check.) The following are simple examples of impossible checks.

However, there are many instances of checks that only seem impossible, but are in fact possible. The earlier tricks, uncapturing and unpromoting come in to provide a solution to the conundrums, usually in conjunction with the discovered check. Rarely, uncastling is also seen. Can you see how the situations in the following diagrams arose?

Here it's obvious black's king came from a7 last move, but then how did white give check with the bishop?

Here white seems to be giving an impossible double check like in the second example. But it is in fact possible; how?

Yet another seemingly impossible double check. But the little white pawn on b6 is the star here. Also, don't just stop at resolving the immediate problem; always make sure the position is obviously legal after all retractions!

The above diagrams are of 3 well-known possible check patterns. Although at first glance they appear impossible, clever retractions show all 3 to be possible. Never immediately dismiss seemingly impossible checks without considering all the sly possibilities first!

Solutions:

There are many more patterns of tricky possible checks out there. Don't be fooled by appearances!

Trick 4 is basically leading to the elusive double-check. The moral of the trick is that you don't give up just because a King is in check by two pieces at once. Always consider the possibility of a double-check.

Even a Queen can be part of a double-check if she's on the promotion rank.

When retracting, look for ways to mask one checking piece with the other checking piece.

It's a play off David Letterman's "Stupid Pet Tricks". Also, it's the feeling you get when you notice one of them in a problem you've been staring at for awhile.

Seconded. Especially all the promotion shenanigans like phoenices/prenices, Pronkins, Ceriani-Frolkins, Schnoebelens etc.

Also, I think explaining the cross-capture and impostors/phoenices would be the most useful to getting people quickly retro-literate, since those are the least intuitive to see but tend to be critical manoeuvres. Plus, they don't need Retro-Vision ® to be applied.

Also, sorry for lagging the page with 8 diagrams. x.x And that was just for the more well-known patterns only.

I want to add to Trick 4, which uses another trick.

Try to figure out the solution first!

Solution for the above: The last move was by White, 0-0-0, and before that it was Black king capturing something (otherwise White has no last move).

What are these terms... I need to learn too.

On which squares did the captures occur?

Some of you will start screaming "a2 and b1". A few more that actually think for a while will say "cannot be determined". Well, that's the cue for another trick, entering the intermediate stuff...

Trick 5: Cross captures

Let me give you a hint. What move did the pawn from a7 do?

Not enough? Here's another. How did the rook at b1 come out from a8?

Yes, the pawn wall of Black makes it impossible to let out the rook! If you didn't think of this trick.

The pawn on a7 went to b6, and b7 went to a6, performing a cross capture! Be careful, what looks innocent to the eye might hold the whole key.

Now we can explain how the rook went out: a7 goes to b6, then the rook gets out, then b7 goes to a6. Or the other way around, with b7 to a6 first. This means the two captures are on a6 and b6. A possible game describing the given situation:

As a bonus exercise, figure out why the order, or the pieces captured, cannot be determined with certainty!

Is the position legal?

Trick 6: Inventory

Count White's pieces. There are 15. Now Black needs a capture axb6 and dxc6 (remember, pawns capture differently from their non-capturing movements). Add them up to get...17! But White begins with 16 pieces; impossible! So the position is illegal.

Another kind of inventory trick that is not about captures, but about our stranded riders...

Is the position legal?

It looks like it's legal! There is no pawn capture required. But here's a thing about rooks and bishops; they can be obstructed by pawns. (Queens have the same problem, but queens are more mobile and on the center, so when queens can't move then most of the other pieces tend not to be able to move too.)

Look at the rook from h8 and compare it with the missing rook from a1. How can the rook from a1 get out? It can't! The bishop on c1 is frozen in place (by b2 and d2), and the pawn at a3 doesn't give way. So the rook from h8 must come from...h1? If so, where did the rook from h1 come from? The only remaining possibility is that it comes from a promotion.

Now, we compare with the bishop on a8. Since it's a light-squared bishop, it must come from f1, which is also blocked by e2 and g2. So again, the remaining possibility is that it's also a promotion! There are two promotions here.

But count White's pawns. There are seven of them. Plus the two used for promotions, this gives White nine pawns, impossible. So the position is illegal.

Very nice thread, I wish I had found this before