White is a Pawn down and has two tempting tries: (A) 24 Nxf6 gxf6 25 Rc6 Qb8 26 Qg8 Rxh2 27 Rxc8 (B) 24 Rc6 Qxc6 25 Nxf6 Qc4 (if 25...Qxg2 26 Qxg2 gxf6 27 Qg8 Rxa2 28 Qxc8 should win on material) 26 Nxh7 Qg4 Material is level.
So I'll go for continuation (A) and hope for the best (5/5 this week not counting Friday's seen before):
=====
I found the right first move,but did not find the best Black defence: 25...Qb7! is much better than 25...Qb8. Winning the Q for R+B+N is not a serious option, so Black goes into the ending a piece up. White has a draw by perpetual check: is there more?

Crafty link to the puzzle position:

click for larger view

(Smirnov vs Rublevsky 2004, 24?) http://www.chessvideos.tv/endgame-t...

Crafty follows the game continuation until 30 Be4 and then counter-attacks with 30...b3 e.g. 31.axb3 Ra5 32.Kb1 Re5 33.Qh7+ Ke8 34.Bg6+ Kd8 35.h5 Rd5 36.Rxd5 exd5 37.Bf5 (White has to be careful here: 37 h6!? Qb6! with a mate threat e.g. 38.Qh8+? Kc7 39.Qxf6 Qg1+ 40.Ka2 Nb4+ 41.Ka3 a5 abd Black wins) Ne7 38.Qh8+ Kc7 39.Bxd7 Kxd7 40.Qxf6 d4 41.Qxd4+ Ke6 with a draw. Of course there may be better plans for White.

<Pawns:> White has three pawns protecting his king and one pawn in the h file. Black's queenside pawns are supposed to be helping in the attack against the white king, yet they are not. His kingside pawns are more effective in their defensive role, but, without the collaboration of his own pieces, they cannot do much.

<Knights:> Both are centralized. However, the white knight is supporting his advanced rook and could take a lead role later on. The black knight, instead, is pinned against his queen, itself undefended, so it is not effectively protecting c6.

<Bishops:> Neither bishop is doing much at the moment. However, the white bishop is beautifully placed at g2, where, if it were not for white's own knight, it would control every square in the long diagonal. The black bishop is still in its initial square.

<Rooks:> Here is where the difference is the most abysmal. White's rooks control the central d file, creating a barrier that prevents black from coordinating his forces. Black's queenside rook is helping sustain the position, but his kingside rook is in an awkward square defending a lone pawn. (Not without a reason. If that pawn fell, black's position would fall to pieces.)

<Queens:> The white queen does not lead the white attack, but she directs it from afar. The black queen is barely sustaining the position alone.

So the conclusion is white is clearly better. However, black has his pieces positioned in such a way that, if white is to give life to his pieces, one of them is to die.

The power of a pin against the queen! Now <both> rooks are effectively attacking black's rear.

<25...♕b7>

25...♕b8 also meets 26.♕g8!, and the white rook is attacked by one piece less.

<26.♕g8!>

This decisive penetration of black's backrank signals the beginning of the end for black, who cannot play 26...♖f7?? in view of 27.♕d8#.

<26...♘xc6>

But now the knight is pinned again and black's pieces only prevent each other from collaborating in the defense of the king.

<27.♕xh7+ ♔f8 28.♕h7+ ♔e7>

Another possibility is 28...♔f7 29.♖d8! ♘xd8 30.♗xb7 ♕xd8.

<29.♕g7+ ♔e8 30.♕xf6>

The simplest. White threatens 31.♗xc6, so black must give back an exchange.

<30...♖d5 31.♗xd5 exd5 32.♖e1+ ♔d7>

Not 32...♘e7? 33.h4, and the h pawn will be unhindered in its advance.

<33.♖e6 ♔c7>

Parrying the threat 34.♕f7#.

<34.♕f4+ ♔b6 35.♕xb4+ ♔c7>

Not 35...♔a7?? 36.♕c5+.

<36.♕d6+ ♔b6 37.♖f6 ♕d7>

Parrying the threat 38.♕b4+ ♔a7 (38...♔c7 39.♖f7+) 39.♕c5+.

<38.♕b4+ ♔c7>

Slightly worse would be 38...♔a7 39.♕c5+ ♔b7 40.♕f8, because d6 would not be defended by the king.

<39.♕f8 ♔b6>

24.Rc6 Qxc6 25.Nxf6 Qxg2 26.Qxg2 gxf6 27.Qg8 is very strong but black could play 26...Kxf6 instead.

There's much more to look at here such as 24.Nxf6 Kxf6 and what happens after that.

Black's, 22...Rh7, + 1.29, swung the game in White's favour. Rf8 +0.53, would have been a better move. 23...Rb5 +1.93, worsened the situation, better was b3 +1.26. Black was also a pawn up so was not doing too good at all. Black went a Knight up with 24...gxf6 but still trailing +1.93 in reality but +5.18 if allowing for the Knight advantage (-3.25).

Anyway White stumbled with
29.h4 +1.04. better to hold the advantage was Rd8, +1.93.

30.Be4. +0.01 wasn't too clever,
(Black, by the way is still a Knight up) much better was Qh7+, +0.58.

Black ruined things with
30...Qc7 + (#6). better, Qa7, = 0.01.
and Black is still a Knight up.

B has an extra pawn but W, overall, is clearly better. He has sound control of the d-file, the possibility of an attack against the BQ at c6 when the e4-N leaves, a pin against the BN reinforced by a powerful discovery after the d6-R plays and a potential skewer on the 7th rank. Plus, the BK is opened. Of course, one could argue that B too has his own discovery against the WQ but there's a subtil point in W's favor though: W's madame is defended and not the B's madame. The outcome will depend whether W can coordinate these factors.

I studied several candidates but I paid special attention on 2 more particularly, that is, 24.Nxf6 and 24.Ng5. I found that 24.Ng5 suffers major drawbacks: It cuts off the WQ on the g-file and B is not obliged to take it, after all. He just removes the R.

24.Nxf6 is quite an other story. If the R merely moves, it has one safe square from which to prevent the skewer on the 7th rank: f7. But then,W can keep things going with 25.Ng8+ and 25...Ke8 is forced: W threatens mate by means of 26.Rd8+. But here's the problem: after 25...Ke8 W has 26.Bc6+! making god use of the mess on the 2 central files. B will either lose the Q or be mated after 26...Kf8 27.Rd8+.

Black can take at f6: 24...gxf6. Then there's no point continuing 25.Be4 Rf7 26.Bg6? Rg7.

But what else? 25.Rc6 Qb8. Playing the Q on a light square doesn't make sense to me. On b8 at least, the BQ is protected.

There is nothing obvious here. Maybe since the the pin is not effective any more W should eliminate the defender on the K-side: 26.Qg8 Nxc6 27.Qxh7+ Ke8.

Fritz picked holes in white's 29. h4, black's 30. Qc7 (30. Qa7!) and my 24. Ng5. So it seemed to me that he was in a grumpy mood. And that got me thinking about why he might be in a mood. And that eventually got me onto Romeo and Juliet, which goes something like this:

Romeo is in a mood because he loves Rosaline, but she doesn't want to know him. So his two friends Benvolio and Mercutio take him to a masked ball to look at other beauties and take his mind off Rosaline. He sees Juliet and instantly falls in love with her. Unfortunately, she is from the Capulet family who are at war with Romeo's family, the Montagues. They marry in secret because their parents would be furious.

Romeo gets into a duel with Tybalt, a Capulet. But because he now loves Juliet, he refuses to fight Tybalt. Mercutio doesn't know why Romeo is acting all girly, steps into the fight and is killed by Tybalt.

Romeo is banished to a neighbouring state. So that she can be with him, Juliet takes a secret potion which makes it look as if she has died (as you do) but she doesn't tell Romeo. He thinks she has died, comes to her crypt and kills himself. Then she wakes up, sees that Romeo has killed himself and kills herself. The war between the Capulets and the Montagues is over, everyone shakes hands.

The moral (#1) - if you have a friend who is lovesick, think long and hard before you take him to a masked ball to see other beauties. You never know how many people might get killed as a result.

I saw 24.NxP/f6, gxf6; 25.Rc6, and then 26.Qg8, but I had no real clear idea if it was winning or not.

I get a little bit bored sometimes when every post in a POTD just repeats the same analysis. After all, when the solution is a a two move combination (eg on a Monday or Tuesday) then do we all need to post saying exactly the same thing as each other?

If I need to explain one of my analogies, then it patently hasn't worked, at least for you. And for that I am sorry. But here is the thinking I was trying convey:

<serious chess point> Fritz finds three questionable moves - one for me, one for white and one for black.

It's a little unusual for Fritz to find three incorrect moves in a POTD, so I put forward a fantasy theory that Fritz is feeling grumpy. Now, of course this is silly, absurd, ridiculous. Software doesn't have feelings. But let's play with the idea and see where it takes us.

So then we think about what might have made Fritz feel moody. And that leads to the equally silly idea that he might be unlucky in love. Who might he be in love with? Why, another piece of software on the hard disk, of course. Fritz is a male name, so we need to think of a female piece of software, so we pick Paint as a suitably feminine target for his affections.

Cue a little three-handed scene where we demonstrate Fritz's doomed love for Paint.

We can't stop there, so let's think of a way to cheer Fritz up. And that brings us to the opening scenes of Romeo and Juliet, where Romeo is in the same predicament as Fritz. Unfortunately, things don't end too well for Romeo, so let's continue the fantasy by assuming that the same tragedies and faintly ridiculous plot twists happens to Fritz.

I suppose you might call it anthropomorphism - putting human charactertistics onto a non-human subject, in this case a piece of software.

I fully accept that not everyone will "get" the stuff that I write. Some will find it too abstract, flippant, overlong. And, because I try to use a different theme each time, it can be hard to find material that people will recognise. I am always careful, for example, not to use source material that is too British or too specialised.

And as I am playing with the format, I occasionally try new ideas and approaches. Some work better than others. I suppose that's inevitable.

But all the time the idea is to try something new, take a risk, invent, with the overall intention of amusing, entertaining and ... sometimes ... instructing.

My Initial try had been 24. Ng5!? as mentioned by <Obit>. If Black takes the bait with 24. Ng5!? fxg4, then 25. Bc6! gives White a winning attack. Unfortunately, after 24. Ng5!? Rh6! =, Black can hold the position.

So if White is to make any progress, he must go in for the demolition sacrifice 24. Nxf6!!

Here's some analysis, based on a move-by-move look with Fritz 10:

<24. Nxf6!! gxf6>

If 24... Kxf6, then White's strong and clear winning move is 25. Rd8!, when play might continue 25...Rh5 (25... Nd3+ 26. Qxd3 Rh4 27. Rf8+ Ke7 28. Qf1 Rf5 29. Rxf5 exf5 30. Qe1+ Re4 31. Bxe4 Qf4+ 32. Kb1 Qxe4 33. Qxe4+ fxe4 34. Rd4 Bb7 35. Rxb4 ) 26. Rf8+ Ke7 27. Qxg7+ Nf7 28. Qxf7#.

<25. Rc6!> This precise deflection is the only clear way to maintain a winning White advantage.

<25...Qb7 26. Qg8!> This is the point of White's initial puzzle-solving combination. Black is forced to go into a middle game where he has two minor pieces for a rook. This might, under other circumstances, be a slight advantage for Black (i.e. six points for Bishop and Knight versus only five points for a Rook).

However, the weakened position of the Black King, the White open attacking lines for all the pieces, a more secure White King position and the poorly placed Black major pieces give White both a positional and a tactical edge.

<25...Nxc6 27. Qxh7+ Kf8 28. Qh8+! Kf7 29. h4?!> This is not White's strongest followup, as it allows Black practical chances to hold the draw.

White has two better winning tries -- 29. Bxc6!? and 29. Rd8! Of the two, 29. Rd8! is the clearest and strongest winning move.

After 29.Rd8! Nxd8 (29... f5 30. Rf8+ Ke7 31. Rxc8 ; 29... Nxd8 30. Bxb7 Nxb7 31. Qxc8 ) 30. Bxb7 Bxb7 31. Qxd8 , it appears White has just enough of an edge with his outside passed pawn and Queen for Rook and Bishop to force the win.

<29... Bd7 30. Be4 Qc7??> This is a mistake, as the absence of a flight square enables White to find a quick mate.

Instead, Black could have provided an escape square for the King and put up more resistance (with perhaps some practical drawing chances) with 30... Qc8!, when play might have continued 31. Qh7+ Ke8 32. Rg1 Kd8 33. Rg8+ Be8 34. Bg6 Qd7 35. Rxe8+ Kc7 36. Qxd7+ Kxd7 37. Ra8 Ra5 38. h5 Rg5 39. Be8+ Kc7 40. Rxa6 Ne5 41. c3 .

<31. Qh7+!> White sees the mate, and gives Black no second chance to recover from his mistake.

<31...Kf8 32. Qh6+ Kf7 33. Bg6+ Ke7 34. Qh7+> 1-0