In Thread 93472, Dragonfire gave some interesting examples of games where a player in a vastly inferior position managed to achieve a draw. I played over Game 2983433 and have concluded that, though Dragonfire found a resourceful way to draw, it required his opponent to make a serious error which threw away the win. It seemed to me that posting the following puzzle on that thread could be construed as hijacking, so I have started a new thread. So, for those of you who wish to participate in this, play over the game and then answer the following questions.
1. What is the last (highest numbered) move where Black still had a winning position, but erred, giving White the opportunity to draw.
2. What move should Black have played instead?
p.s. Dragonfire: As I am sure you know the answer, please refrain from posting it (at least for a few days).
Well, I guess if I'd known a 2000+ player was going to answer (correctly) so quickly, then maybe I should have restricted the puzzle to lower ranked players. Erekose is absolutely correct that after 44... Kb3? 45 e5! Black no longer has a win and that 44 ... e5 preserves the win.
So, (because perhaps I am a glutton for punishment) let's try to extend the puzzle one more time and this time, for the first couple of days, the puzzle is restricted to players below 2000.
The question: Besides 44 ... e5, does Black have any other moves which preserve the win? Give analysis to support your answer.
Originally posted by AlboMalapropFoozerLooks like moving the h pawn or any move keeping the bishop on that diagonal down to g1 would work. Mainly, it follows the rule - when up material, trade pieces not pawns, and when down material trade prawns or resign. ;-)
Well, I guess if I'd known a 2000+ player was going to answer (correctly) so quickly, then maybe I should have restricted the puzzle to lower ranked players. [b]Erekose is absolutely correct that after 44... Kb3? 45 e5! Black no longer has a win and that 44 ... e5 preserves the win.
So, (because perhaps I am a glutton for punishment) let's ...[text shortened]... does Black have any other moves which preserve the win? Give analysis to support your answer.[/b]
Originally posted by ErekoseClearly I had tried for a "cheapo" for some time as black had a clear win here (incindently I know Scott from OTB play as he played for the same club as me and his grade is significantly higher than mine, i.e. about 2200+, and is only less here because the nature of his employment causes him to time out too often). I had realised that I could afford to give up all my remaining pawns and still leave him with a B & RP and draw so tried to swindle that and it was only due to "complacency" (with apologies to Scott) that I was allowed to do it with my 45th move.
Lets see 44 ... e5 looks like an easy win for black, so I guess the question is whether black can win from some position past move 44. If so, I haven't found it.
What I think is interesting although it is some time since I last looked at this game and perhaps someone else can confirm it but I don't think most (all?) engines find my 45th move and only decide its a draw some time later in the game.
P.S. These swindled draws (or even wins) are usually the result of serious errors (as here with blacks 44th move) but the important thing is to try and create the opportunity for error).
Originally posted by Dragon FireHIARCS 11.2 on a very elderly Mac (500 Mhz G4 from 2000) finds 45. e5 almost instantly. It gives black a slight plus but that declines with time to almost zero. Shredder 11 on the same machine also finds e5 very quickly and gives it as a draw straight away. I think that may be because it has endgame tablebases built in and is coded to explore lines which transpose to favourable tablebase positions if possible. In this case Shredder also gives another drawing line but I'll leave that for others to find. Interestingly Fruit 2.3.1 on the same machine also finds e5 instantly and has decided it is draw 22 seconds later without the benefit of tablebases. Sigma, the engine built into the GI used to run HIARCS and Fruit, also finds e5 instantly but then scores the position as a win for black although the score does decline slowly over time. That's it, all the engines I have access to. I don't think there are that many more that run on Macs.
Clearly I had tried for a "cheapo" for some time as black had a clear win here (incindently I know Scott from OTB play as he played for the same club as me and his grade is significantly higher than mine, i.e. about 2200+, and is only less here because the nature of his employment causes him to time out too often). I had realised that I could afford to g ...[text shortened]... cks 44th move) but the important thing is to try and create the opportunity for error).[/b]
Well you did ask!
Originally posted by KeplerIts not 45. e5 that I meant but my response giving up my pawns for the cheapo draw.
HIARCS 11.2 on a very elderly Mac (500 Mhz G4 from 2000) finds 45. e5 almost instantly. It gives black a slight plus but that declines with time to almost zero. Shredder 11 on the same machine also finds e5 very quickly and gives it as a draw straight away. I think that may be because it has endgame tablebases built in and is coded to explore lines which tran ...[text shortened]... have access to. I don't think there are that many more that run on Macs.
Well you did ask!
45. e5 seems pretty obvious but as Scott didn't immediately play it, it can't be that easy.
Originally posted by Dragon FireBut you are white in the game so 45. e5 is your move rather than Scott's. Ah, I see, 44. ... e5 is the move in question. I'll have another look later.
Its not 45. e5 that I meant but my response giving up my pawns for the cheapo draw.
45. e5 seems pretty obvious but as Scott didn't immediately play it, it can't be that easy.
Originally posted by KeplerActually I think you have answered the correct move. Just a little confusion reading over this (from me whilst trying to work) without having a board in front of me as both black and white can play e5.
HIARCS 11.2 on a very elderly Mac (500 Mhz G4 from 2000) finds 45. e5 almost instantly. It gives black a slight plus but that declines with time to almost zero. Shredder 11 on the same machine also finds e5 very quickly and gives it as a draw straight away. I think that may be because it has endgame tablebases built in and is coded to explore lines which tran have access to. I don't think there are that many more that run on Macs.
Well you did ask!
If black plays e5! rather than Kb3?? it appears an obvious win. I have now put this into Fritz 10 which immediately finds that e5 draws for white after Kb3. I might be wrong as its a while since I played this game but I think I looked over it afterwards with Fritz 5 or 6 and have a feeling it did not find e5 although perhaps it would have given time. It certainly looks as if engines are getting better at playing some of these endings.
Exuma (and anybody else who is interested): Against most Black moves, White can draw with 45. e5. There are only two moves which preserve the win for Black: The previously mentioned 44 ... e5 and the other move I discuss below. It's your choice if you want to continue to analyze or decide to view the solution in the next paragraph.
Dragon Fire: I started a new thread specifically because I did not want to hijack the topic of that thread (achieving a draw from a theoretically lost position). Let me be clear that in no way am I trying to demean your find of 45. e5 nor am I criticizing your opponent for overlooking the possibility. What I am trying to do is demonstrate in this thread (and in particular with the question in my second post) is a principle which is often helpful when choosing between two "similar" moves, which is "if move 'A' has all of the advantages of move 'B' and has an addtional advantage and no additional disadvantages, then move 'B' should no longer be under consideration." In our position, what this means is that even if your opponent doesn't see the threat of 45. e5, he should not have played 44 ... Kb3 because 44 ... Kb4 covers all of the important squares that 44 ... Kb3 does plus it covers the important c5 square. And as it turns out, that additional advantage is sufficient to win (one example line is 44 ... Kb4 45. e5 Kc5 46. Ke4 h6 47. Kf4 d5 48. cxd5 Kxd5 etc. ) Of course, I would agree that 44 ... e5 is the best move and leads to a relatively simple win, but a complicated win is still better than a draw.
To sum up, while this prinicple may not directly help us find the best move, if it eliminates one or more inferior moves from consideration, that is useful.
Originally posted by AlboMalapropFoozerInteresting post.
If I remember correctly, I think Nunn talks about something similar in 'Secrets of Practical Chess'. i.e. the concept of comparing two lines and noting that one does at least as much as the other, and maybe more