|
|
| From | Message | ||
|---|---|---|---|
|
markies 25-Aug-10, 11:57 |
 Tactics exercise #345551. Ra4 , of course N x d2 2. Rf2 , Qb5 ( GameKnot) I thought and still think that 2. , Qd5 is even better ! What do you think ? |
||
|
kingdawar 25-Aug-10, 12:25 |
My answer: this position is simply unsuitable to require users to find a "best" move. It confuses me ten(!) out of twenty-two correctly estimated the three required moves. |
||
|
markies 25-Aug-10, 12:26 |
#34555 (2) |
||
|
easy19 25-Aug-10, 13:33 |
but i know why GK prefers Qb5 over your other suggestions.. it is because it leads to a forced mate in 13/14 moves and your suggestions lead to a longer mate 14/15 moves but wen you just look for the best moves just 4/5 moves ahead. it is kind of hard to pick the very best move by margin out of the moves available.. not a 1221 puzzle |
||
|
easy19 25-Aug-10, 13:43 |
i hate this puzzle. :) 2. Rf2 Qb5 3. Raf4 Ne4 << i moved Qxe5 and worked out a mate in 12 moves I can not find a shorter mate with the GK move Ne4 Well i faile dthis damn puzzle also.. Sorry, you failed to solve this problem. Good try! Your new rating: 1985 (-19.6) Elapsed time: 04:15 Problem's rating: 1359 Avg. time to solve: 00:49 Total attempts: 24 Solved percentage: 41.7% 4. a4 Qxe5 5. Rf1 g5 6. Kg1 gxf4 7. Rd1 Nc3 8. Kh1 Nxd1 9. Kh2 Nf2 10. a5 Qxa5 |
||
|
If this puzzle ... I begin more clearly to see what heinskat has been beefing about. On the whole I've been on the side of GK in these discussions (I'm still missing the quick kills at the end of the sequence - very dismal). But there is a difference, I think, between finding the immediately decisive move - the quick mate if you will, and an indifferent alternative that prolongs the game indefinitely. There has to be a clear cut, obvious-once-you-see-it, differential between best and second best moves, I think. (Mind you, in one of these exercises yesterday, I missed a mate in one in favour of a mate in 3. Ouch-ah!) In the puzzle under discussion, I think (I don't know for sure) that I would have played 2...Qb5 [a] because it carries an immediate threat, and [b] it offers the prospect of a quick attack upon the King: 1.Ra4?? Nxd2 2.Rf2 Qb5 3.Rb4 ... The rook has no other safe square (3.Rf4 is unsafe after ...Qe5 in my view) 3... Qe5+ 4.g3 ... There is no safe retreat for the King, and Rf4 is met by ...g5. 4... Rxb5 5.axb5 Qe1 ... and White has no defence left. What more ought one to ask? |
||
|
... More to this puzzle. 1.Ra4?? Nxd2 2.Rf2 Qb5 ... Now, I will have more to say about markies his 2...Qd5 suggestion, in a moment, concerning whether or not it deserves a 'fail' mark. 3.Rb4 ... I regarded 3.Raf4 as unsafe, on account of the quick attack Black got by 3...Qxe5 4.g3 (avoiding ...g5, though Black has more serious threats than that. Although in my last posting I had not seen any specific mates, the position is obviously ripe for a mating attack. As it transpires, it doesn't make much difference which of the two rook moves White makes. The attack will take much the same form either way. 3...Qxe5+ 4.g3 ... Here 4.Rbf4 would merely transpose into the line that would have ensued had the rook moved there at move 3: 4.Rbf4 Rb1 4.h4 (say) Rf1!! - and here 4...Nf1+ is obvious and good enough to clean up the game shortly. It's just that this rook move is about 2 moves better. 5.g3 Qe1! 6.Rxf1 ... (White's game is so deperate, he has to shed further material to preserve his King a little longer) 6...Nxf1+ 7.Kh1 Qxg3 (7...Ne3+ is not as quick: 8.Kh2 Qe2+ 9.Rf2 Qxf2+ 10.Kh1 Be4#) 8.Rxf1 Be4+ 9.Rf3 Bxf3#. 4...Rxb4 An alternative seems to be 4...Qe1 at once, but I haven't explored that line. 5.axb4 Qe1 6.Rxf5 ... Depriving Black of an attacker. There's nothing better, but mate must be imminent. 6... gxf5 7.g4 Qf2+ 8.Kh1 Nf3 9... Qg1# or 9...Qh2# The mates seem to be equidistant if White plays 3.Rb4 or 3.Raf4. Now, what of 2...Qd5, then? Not as (overtly) threatening as ...Qb5, it does protect the knight at d2, and attacks the e5-pawn. That pawn can not directly be defended: 2...Qd5 3.Re2?? Nf1+ 4.Kg1 (to avoid the knight fork at g3) 4...Qc5+ 5.Kxe1 Rb1+ 6.Re1 Bd3#. It seems that White really has nothing better than to play 3.Raf4 Qxe5, and transpose into the 2...Qb5 line. Which implies to me that 2...Qd5 is just as good as 2...Qb5 - slight less forcing, but perhaps the more subtle. All this, of course, comes after a considerable amount of analysis. I've spent an hour at least on this (including typing it up). For a 'pass' mark at 1221-level, Black ought to have found the Queen capyure (fine), and then a reasonable attacking follow-up to exploit the material gained: 2...Qb5; 2...Qd5 and I think 2...Qb2 ought to come into consideration, as it, too, combines attack and defence. For the rest, I would accept 2...Ne4. All these moves in my view advance Black's cause, any of which is reasonable to expect a 1221-level player to try. So maybe there needs to be some kind of gradation within the exercise, depending upon where it finishes. Several puzzles simply involve spotting a two or three move forking sequence, say. Having found that, the puzzle ends, solved. But some have a little sting in the tail: the same sort of short range gig, but then has a further forcing sequence to follow - often a short mate. Now, I don't know how often I've missed those - quite a few - simply because I've felt it was 'problem solved'. This is especially true when the position is clearly winning. So? I'm not sure what to make of it. I've failed a lot of puzzles now (several of which I mentally kicked myself afterwards: I ought not to have missed them). Most of the time I get the first move correct, then cark out by failing to find the most effecient 'rounding off'. Occasionally I have felt sure that my method would have been no worse than the given answer (but I haven't in general followed these up). But the counter-argument goes to alertness. Just because one has made a gain, don't imagine the work is finished. A good lesson to learn. But in the case of puzzles of the type shown here, either it just doesn't 'fit' the criteria GK is hoping to achieve, or else it could be a candidate for some future system of awarding for how much of the solution you find. |
||
|
kingdawar 25-Aug-10, 16:04 |
By the way... gm=board-number — link to a game; for example:  game (the board number is shown in the title bar of the game's page) fen="FEN-notation" — display chess diagram (quotation marks are required); for example: pz=puzzle-number — link to a puzzle; for example: puzzle #2090 tc=tactics-number — link to a tactical exercise; for example: tactics #17364 very nice!! |
||
|
baronderkilt 25-Aug-10, 18:15 |
It sounds like everyone markies , You are definately ahead of me on this ... I have not even figured out yet what is wrong with "1.Qd5" . |
||
|
Baronderkilt - In my earlier posting I come to the conclusion that 2...Qd5 is as good as 2...Qb5. They lead to the same mate. I'm also finding that several exercises have, for mine, more than one element. For instance, a 1250-type puzzle might involve a two-move combination to extablish a K-ng-Queen fork. One solves this in quick time. Then one comes across another 1250-type puzzle: same thing - King-Queen fork end of st- Oh. Hang on. There's more. Turns out there's a mate to be found (or something) on top of what has just happened. In my view, that exercise has two elements, and each element ought to be rewarded separately. OK then, ought failing an element be 'punished' according? Possibly. Suppose I, with a rating of 1200 attempt to solve a 1200-level puzzle that has a two-move forking combination, followed by a two-move mate. I get the fork (+10.0 - I'm not sure how the points work) and get the two move mate as well (another +10.0). If I miss the mate after getting the fork, I get +10.0 for the former and drop 10.0 for the latter - net score zero. What if I miss the fork, then? Not sure. You don't get to see the mate, so you haven't tried it. But you have failed the whole thing. But there is a possibility in multi-element puzzles to scaor the first element as a fail, and begin again a move or two later to see if the second element might be picked up. Any thoughts on this? You know: every day I manage to improve my CT rating, but man, do I take the switchback route! Today I got from 1660-odd to 1670-odd by way of 1588 - and a harrowing ride it was. ... |
||
|
lighttotheright 25-Aug-10, 23:26 |
I believe you are correct. There is nothing wrong with 1.Qd5. In fact, I believe it is best and definitely better than the rook move. The rook move leads to a long mating slug fest. In my eyes it is not a productive line for white. Either I'm missing something major or the diagram of the problem posted here is inaccurate. Something must be missing for 1.Ra4 to have any relevance. As is, 1.Qd5 leads to Nxa5 (which is better than losing the queen) 2.Qxa5 and White is down a bishop but has stymied any immediate Black threat. Arguably, White still has chances with the passed pawn on the a file. Black is definitely stronger but that is true no matter what White plays. Black should win regardless of how White responds. |
||
|
Also, we do not manually assign ratings to any of the puzzles, so the only reason this puzzle is rated this low is because of how many players solved it and how many failed to solve it, and because of their respective ratings at the time. |
||
|
I misunderstood Craig's point... There is nothing wrong with 1.Qd5. But that's not the point. Your response is asked as Black, supposing White were to play the move he did (1.Ra4). That really is the only relevance to the move: it is the one played, and hence the one to which you must respond. I like this approach. It has the feel of being in an actual game. It also I think makes for accurate time keeping on each exercise. For this reason I have tried from the beginning, to solve the exercises 'on screen', though I find I'm beginning lately to check out my chessboard when I'm not sure. For some reason I still find a physical chessboard easier to read, and, of course, I can shuffle pieces about. At that, I am beginning to think the exercises are more designed for that kind of analysis rather than trying to solve 'OTB style'. |
||
|
kingdawar 26-Aug-10, 08:48 |
gameknot_com I will take the tactics #34555 example because it is possibly quite "deep": the computer's score at the end of the line is +27.27 vs. +22.22. Let's take a look at the given continuations... to see what moves the computer saw and on which these these numbers were based... 3. Raf4 Ne4 4. a4?! Qxe5 5. Rf1 g5 6. Kg1 gxf4?! 7. Rd1 Nc3?! 8. Kh1?! Nxd1 9. Kh2 Nf2?! 10. a5?! Qxa5?! Huh? This semi-random line - with possibility to do better for both sides all around - amounts to +27.27 apparently. Now let's take the +22.22 line 3. Raf4 Qxe5 4. a4?! Ne4?! 5. Rf2f3?! g5 6. a5?! gxf4 7. h4 Ra8?! 8. Kh1 Rxa5?! 9. Rb3 Nf2+ 10. Kg1?! ...and 10. ... Qe1+ 11. Kh2 Qh1#. Mate in two. The other line does not end in mate as quickly. So which line is better? Can you see I can be suspicious of these numbers if the supporting evidence is somewhat... feeble? If in the +22.22 line, 4. ... Ne4 is changed into 4. ... g5 it should be more than 22.22. And with enough improvements in the line the objective verdict should turn into +27 as well - or just, what has been decided long before... black's position is won and he will capture the white king eventually. But the bottom line is that these numbers added to chess positions are somewhat coincidental and I can imagine other programs with other preferences picking different moves from these positions. |
||
|
hortstu 26-Aug-10, 09:08 |
Also, we do not manually assign ratings to any of the puzzles, so the only reason this puzzle is rated this low is because of how many players solved it and how many failed to solve it, and because of their respective ratings at the time. " Maybe temporarily you should allow free members access to more exercises? I assume more members, with a wide range of ratings, performing more exercises will lead to a more accurate rating for the exercises? I'd like to do my part. Mike |
||
|
kingdawar 26-Aug-10, 09:09 |
hortstu |
||
|
supporting gameknot_com |
