GameKnot: GameKnot related, Bug in puzzle-maker ?

GameKnot related: Bug in puzzle-maker ?

From	Message
baronderkilt 22-Oct-09, 16:04	Bug in puzzle-maker ? Has anyone else seen this? I just made this puzzle today #16238. And was originally working with mate in 8 solutions, but during the course of it I did find the mate in 7 which is now listed as an Improvement. During creation of the solution it even showed the Seven moves and the last of those ending "#". But for some reason the puzzle did not update from showing "mate in 8" on that very page which showed mate in 7 gamescore. And so I pondered it at the time, then thought it must be waiting to leave the page before it updates to making it a mate in 7. But it did not. So I shrugged and thought, must have been an inferior defense?! But now I see it not so as the very line I saw then, I just played as being the new solution to the Improved puzzle?! ****
baronderkilt 22-Oct-09, 16:13	A PS ... Earlier today I did another, #16235 which someone improved from mate in 4 to mate in 3. In that one I simply missed a pawn push that did shorten it to three. And so the improvement there was handled legit by the system, for that one. *** No complaint there ... except that missing the first myself makes the BUG Bug-Me twice as much~! ... as you can see, by the following, "Bugged-Smiley Wearing Dunce Cap!!" <}B-p
kingdawar 22-Oct-09, 16:45	It's not a bug in the puzzle-maker
baronderkilt 23-Oct-09, 11:50	Deleted by baronderkilt on 23-Oct-09, 11:52.
baronderkilt 23-Oct-09, 11:52	Henizkat, if so then I think that it needs upgraded. I am under the impression that it is self-correcting to the shortest sequence as I work thru it, as it has seemed to do in the past. But if not we should have some instructions on how it does work. Also it could be upgraded to better make use of real game situations, like people actually see in their play. Not refuse them as too many alternative solutions. * Otherwise it cannot be avoided that there will be places where a King takes a rook check, then can move to a few different squares, then the Rook will sac itself in a final check to lengthen the mate. But each of those K squares becomes an Alternate. It would be cleaner to have that recognized by the system and return to there and prompt for alternative king moves. * Or if we do not want real game situations then that should be stated. But to me that is what differentiates a Puzzle from a "Problem", the latter being expected to be ultra-clean and not use extra pieces, no "cooks" aka Alternate Solutions, etc. As it is, it seems to be most amenable to "Puzblems". It will take a bit of Puzzle, but not too much. * What do others think? Are we better to limit to Problem compositions? Real life situations? Both or blend, just as it is? Personally I enjoy real life puzzle types best. Problems tend to provide too many "clues" for my taste, from knowing the "rules" by which they are made. Tho I do appreciate many as being very brilliantly done. And still good to study to get the numinous of the tactical situation which can be transferred to ones real game analysis. But imo, the main difference in analytical ability between an Expert and a GM is not to do with length of analysis possible, Is largely due to recognition of more tactical motifs, but also largely due to ability to analyse more BROADLY in terms of Candidates and Analysis Tree (ala Kotov). * While Problems address the Second of those, but not often the third. Whereas Puzzles are ideal from practicing the Latter ability. No?!
kingdawar 23-Oct-09, 11:55	Sorry but... CCMCA you have no idea what you are talking about I agree though that it might be handy to have some sort of tutorial on how to use the puzzle application.
baronderkilt 23-Oct-09, 13:25	HK ... Thanks, if I don't that tends to prove the need for good, clear instructions. What was it You are talking about !? }8-/ (this is "Huh, did he really say anything" smiley)
kingdawar 23-Oct-09, 13:28	"Enter the solution: by making the moves on the board. The virtual opponent's moves will be made for you automatically, one by one, covering all possible moves. To verify the solution, you need to provide a sequence of responses to all possible moves by your virtual opponent that always lead to a checkmate." The thing just works through every possible move until it has covered all possible replies for the "virtual opponent". Once you have checkmated in all variations it says hey, you can save it. That doesn't mean you have given the shortest solution though, as can be seen in the puzzle section ...
baronderkilt 23-Oct-09, 13:39	That I know ... What I do not know is why it did not Take the mate-in-7 when I entered it after seeing the better move in the process of entering solution. AND also used it when setting up the Qe3+ Alternate Solution line. Prior to that Rxf1 was the mainline. In the past it has always changed the "mate in" number when a shorter mate or longer defense is found during the process. As it DID do when the Solver Improved it after it was saved. When I did make the Alternate Solution I forgot to use the best move a couple times but did use it many times. And would think that when I did use the mate in 7 the system should have erased the Mate in 8 lines and redirected me to them for the mate in 7 correction ... just as it Did do when Improved by the Solver. ?!
kingdawar 23-Oct-09, 14:28	Simple, you did not find a mate in 7 in all variations.

« Back to forum