It doesn't matter for after 40... Qxd4+ simply 41. g4 there (major threat is 42. Qf8#). Black only has two serious tries which both fail: 1) 41... Qb4 retaining the pin on the pawn 'g4' and protecting the square 'f8' fails to 42. Qf8+!! Qxf8 43. g5# or 2) 41... Qxf6+ 42. Qxf6 Rxe3 43. g5# :-)

Also... <you vs yourself: <alekhine101> I think after 38..h5, white has 39.fxg6>

Yep 39. fxg6 is decisive! The threat is mate in two and there is nothing to stop 39... fxg 40. Qf8+ Kh7 41. Rf7# as well! ;-)

<jhoro: Mal, under 3 min is not much, but i'm a bit disappointed he got that far and missed a nice finish.>

Yes, it would have appeared in every combination book in the future! (now it's just a diagram with side note)... Not sure about <but also any of <38.Kh4 and 38.Kf4> were winning as well.> Clearly 38. Kh4 wins but on 38. Kf4 it appears messy - but in retrospect as you've pointed out <38. Kf4 h6!> However not 38... Rf1+ 39. Kg5! ~lol~ Also not 1) 38... Qc1 since 39. Qd8+ Kg7 40. f6+ Kh6 41. Qf8+ Kh5 42. g4+ Kh4 43. Qh6+ mates next, or 2) 38... Rxe3 39. Qd8+ Kg7 40. f6+ Kh6 41. Qf8+ Kh5 42. g4+ Kh4 42. Qh6+ and mate next.

However <it is funny that the idea <Qd8+, f6+, Kh4> was working at move 39 as well.> In this line after 39. Qd8+ Kg7 40.f6+ Kh5 41. Kh4 - Yes as only one line fails black from not saving this with the desperate try 41...Qe1+ and now 42. g3! wins for if 42... Qe4+ 43. g4! Qe1+ 44. Rg3 and wins as there is no way to simultaneously stop both 45. Qf8# or 45. g5# (since after 44... Rxf6 then 45. g5#). ~lol~ Nothing can be saved... After several *edits* I'm finally done - well done! ;-) - Mal

Sorry <jhoro> I was trying fix things! The first post was all set for <you vs yourself> then you posted. So, I started editing again! Must admit I'm use to Forums where we can diddle and get everything set right for an hour correcting junk. Here you have to delete to fix... ~Argh~ I'm done editing! If I made a mistake - I bloody made a mistake! ~lol~

click for larger view

34...Qa1 35.f5 Rc8?

<Better but still losing was: 35...Re8 36.fxg6 fxg6 (36...hxg6? 37.Rf3) 37.h4!>

36.Re3 Rh1+ 37.Kg3 Re1 38.Rf3

<A mistake, White could have won with 37.Qd8+ Kg7 38.f6+ Kh6 39.Kh4!! and mate in 13, however White is still winning.>

38...Rf1?

<Better was: <38...Re8> with the following long forced line: 39.fxg6 Qe1 40.Rf2 fxg6 41.Qf7 Kh8 42.Ne2 Re3 43.Kh2 Qb4 44.Qf6 Kg8 45.Nf4! Qe7 46.Qxb6 Re5 47.Qxa5 and White should win this. Also <38...Rh1> loses to 39.Kg4!! mate in 11.>

39.fxg6??

<This is the move that changes the evaluation from won for White to a draw White still could have won with 39.Qd8 Kg7 40.f6 Kh6 41.Kh4 and here it is mate in 8.>

<mack> Another way to look at it told to me by my old chess friend IM Julio Kaplan (WJrCC 1967-68) after I placed 2nd twice in ICCF Master class postal competition (1971-1975)...

"It takes luck to be first, but it takes skill to be second!"

35.h4!

If 35...(a) 35...h5 36.f5 37.fxg6 Qg1+ 38.Kxg1 Rc1+ 39.Kf2 Rf1+ 40.Kxf1 fxg6 41. Rxg6+ Kh7 42.Qg7#

If (b) 35...a4 36.h5 Qd1 37.hxg6 Qh5+ 38.Rh3 Qxh3+ 39.Kxh3 ;

If (c) 35...h6 36.f5 Rc6 37.Qxc6 Qxd4 38.Qe8+ Kg7 39.Qe7 Qf4 40.fxg6

If after 35.h4! the game might continue:

35...h5 36.f5 (threatening: 37.fxg6 Qg1+ 38.Kxg1 Rc1+ 39.Kf2 Rf1+ 40.Kxf1 a4 41.gxf7+ Kh7 42.f8N#)

36...Rc6

( 36...Qxd4 37.Qxd4 Rc4 38.Qxd5 ) 37.Qxc6 Qxd4 38.Qe8+]