ARCHIVED POSTS
< Earlier Kibitzing · PAGE 558 OF 1118 ·
Later Kibitzing> |
Sep-20-12
 | | SwitchingQuylthulg: <Robed.Bishop> <chessgames.com> I think this is just another instance of a bug I've reported about five times so far but you never bothered to fix :) If the last move is made by the losing player (meaning he resigns after his own move) you get one more opportunity to make a move, and if you grab it you get that particular error message. |
|
Sep-20-12
 | | chessgames.com: < If the last move is made by the losing player (meaning he resigns after his own move) you get one more opportunity to make a move, and if you grab it you get that particular error message.> We've investigated this even more thoroughly now and you are absolutely right. The problem stems from the program that initially sets up the GTM game, what we call the "Toga analyzer." in this specific case (Tal vs M Pasman, 1953) the last move that the user should be concerned with is 40.Rd3+ but Toga actually bothered to calculate that after 40...Qxd3 the best reply is 41.Qxd3+. Our Toga analyzer was programmed to "keep going as long as there are more moves" ... it sounds logical, but it's naive. It needs to "keep going as long as there are more moves by the player who wins." So that still needs to be fixed. Even then, we have to go back and look for other games that meet this criterion and then fix them. That will actually be helpful for other reasons. Anyhow, sorry for the annoyance, but at least we can be relieved that it doesn't deduct or add to your scores in any unfair way. |
|
| Sep-20-12 | | Robed.Bishop: Thank you for the explanation. In all honesty, I wasn't too concerned about the score because I don't use the feature as a competition, but others might be. It does keep a user from actually completing the game and receiving a final score, or at least a final score that is visible to the user. |
|
Sep-20-12
| | chessgames.com: The new status is:
New GTM submissions with this property (the loser making the final move) will be processed correctly. The old games with this property need to be reprocessed (this Tal game of course is an exception--it's been reprocessed a dozen times in the last day!) It shouldn't be too tricky to write some software to figure out which GTM games have that property and then this 7077 error will be behind us for good. Thanks for your help, especially Switchy who basically had this figured out when I was still shrugging my shoulders thinking "million-to-one anomaly." |
|
Sep-20-12
| | SwitchingQuylthulg: <Robed.Bishop> Are you sure you didn't get a final score? I've triggered that bug many many times (including twice in Morphy vs A Meek, 1857, which I used for testing purposes) and always received my final score despite it. <Two> final scores, in fact - one without my points for the phantom move that counts for the par, and another including those extra points that shows up in Guess-the-Move: Review all Your Scores! On one occasion no one else had played the game in question, thus highlighting this oddity... <Given 1 time; par: 96; your score: 99> I'm so good at Guess-the-Move, I can beat the par even when I'm the only player :) |
|
| Sep-20-12 | | Robed.Bishop: <SwitchingQuylthulg> I'm not sure how to answer your question. I put in my last move and was taken to the error page. I could return to the page that I entered the move, but that was still asking for a move, and so didn't provide a final score. As I said at the start, I was clearly at the end of the game, and the last move to be made wasn't going to alter the outcome, just give me three more points. Whether there was another move after that seems doubtful, but I cannot say that for certain. The final score isn't the most important component of the feature. For example, I have scored above par on many games, but in some of those I guessed moves that subtracted points from my score and which, when I looked at them more closely, were game-losing moves. So having a higher score doesn't necessarily mean that in a real game I would have won, only that I guessed other moves correctly. I use the feature to learn or practice systems I am not so familiar with, or to help recognize alternative approaches to defenses I use. In the game that started this, I was going through some Sicilian games as white to help recognize attacking positions. Getting good scores might make me feel warm and fuzzy inside, but really what I'm looking for is an increase in scores as I become more familiar with system patterns. This would generally indicate a trend toward fewer mistakes and more accurate attacks. |
|
| Sep-20-12 | | Robed.Bishop: <SwitchingQuylthulg> I forgot to add a thank you for your input on this issue. Thank you. |
|
| Sep-20-12 | | Calli: Problem:
Sub-rounds on scorecards. Some of the old tournaments require a number like "12.2", but four characters are not visible in the "rnd" box. See, for instance, Alapin vs J Mason, 1901Suggestion:
Make the "Score" box one character narrower and the Rnd box one char wider. Since it is really a result info, replace the word "Score" with "R" or "Rsl" or "Rslt". All are valid abbreviations for "Result". |
|
Sep-20-12
| | SwitchingQuylthulg: <Robed.Bishop> You can find your final score on either the Guess-the-Move: Review all Your Scores page or the Tal vs M Pasman, 1953 page :) |
|
| Sep-20-12 | | Calli: A bug:
As an "ADMINISTRATIVE GAME EDITOR", I made a mistake and entered a Feb 29 for a non-leap year. The system accepted it and reset the year to "0000". |
|
| Sep-20-12 | | Robed.Bishop: <SW> Yes, I see it there on the game page. I was not aware of that link. Thank you again. |
|
Sep-20-12
| | chessgames.com: <Calli> Good work on that leap-year problem. Now that I think about it there might have been similar problems with dates like "September 31st" of any year. We just experimented with Winawer vs Von Scheve, 1901 (I assume that's where you found it?) to see this for ourselves. And the Librarian fixed those two games you flagged. We just put a fix in place. That means there is a new kind of error you can get that will look like this: <Sorry, the date 1901-02-29 seems ot be invalid. Please doublecheck.> About the round-number box, that's a PGN viewer issue (that is to say, not our software, and may be changed or at least become changeable in the future). You could suggest your idea to Nikolay (the CVD author) over at Viewer Deluxe chessforum but frankly there isn't a lot that can be done--there will always be data that don't visually fit in those fields. Maybe a few pixels could be earned by dropping the word "Score" but even then it's debatable if those pixels should make the round-field bigger or help out the event-field. I have a feeling Nikolay pondered these issues long ago and made a pretty good compromise. |
|
Sep-20-12
| | chessgames.com: <SwitchingQuylthulg: <Robed.Bishop> You can find your final score on either the Guess-the-Move: Review all Your Scores page or the Tal vs M Pasman, 1953 page :)> Exactly, Robed.Bishop was stuck on an error page but behind-the-scenes he did in fact get a final score for the game, which can be found out later with a bit of effort. And it's even the right score (no surprise penalty for his last move or anything crazy like that.) We're almost rid of this 7077 error, we just now have to compile a list of all GTM games with that property of the loser resigning while it was their opponent's move. |
|
| Sep-20-12 | | Calli: <chessgames.com> okay, I'll bring it up to Nikolay. IMHO, score means the moves of a game and not the result of the game. To me, it's been mislabeled from the start. |
|
Sep-20-12
 | | LoveThatJoker: <CG> Glad to see that Guess-the-Move is in the forefront here on this page. Any word on the upcoming <CG> Guess-the-Move Forum? LTJ |
|
| Sep-21-12 | | Robed.Bishop: <chessgames> Thank you for your assistance with this feature. <LTJ> A forum? You've got to be kidding. |
|
Sep-21-12
| | chessgames.com: <A forum? You've got to be kidding.> Not at all, we've thought it over and we think a forum for discussing Guess-the-Move is a good idea. It could encompass bug reports, challenges, boasts, whatever. It might even be linked to some GTM features that otherwise hard to know, like "newest GTM additions" or a "GTM of the Day" type challenge (to help get some of those zeros off the scoreboards). Look for it soon. First we have to worry about the London Grand Prix, the newly rediscovered "7077 bug", new historical tournament discussion pages, the official Playing Zone launch, and a few other important features. It's going to be a busy September-October. |
|
Sep-21-12
| | chessgames.com: <Score means the moves of a game and not the result of the game. To me, it's been mislabeled from the start.> That's a very good point. "Res" for "Result" might be the best. In fact it can almost be unlabeled--if you see "1-0" or "0-1" it can only mean one thing. |
|
Sep-21-12
| | LoveThatJoker: <CG> Great stuff!
LTJ |
|
Sep-21-12
| | SwitchingQuylthulg: <chessgames> An update on the phantom move bug ;-) It doesn't look to me as if "new GTM submissions with this property (the loser making the final move) [are now] processed correctly" - or if they are, then Guess-the-Move itself is mishandling them somehow, allowing us to guess that extra move even though it has not been processed. I just uploaded two games with this property to GTM, namely Guess-the-Move: Lasker vs R Illa, 1910 (yesterday evening) and Guess-the-Move: Lasker vs Janowski, 1909 (this morning), and in both cases I was able to trigger the 7077 error message. I didn't receive any points for the phantom move, so that at least seems to have been fixed, but those games were still either misprocessed or mishandled. Another (rather minor) GTM issue - the new #1 Hardest Game is Alekhine vs A Romashkevich, 1906, which I don't think is appropriate for GTM at all: <Some games are not apppropriate [sic] for Guess-the-Move. Games in which the winner commits atrocious blunders, or wins through some abnormal means, and some composed games, are examples [of] what would not be desired. > This isn't one of Alekhine's best games at all; he both commits an atrocious blunder on move 13 (failing to play the obvious 13.cxd4, which simply wins a piece) <and> wins through abnormal means (he doesn't seem to be winning in the final position - according to <Calli> in the kibitzing section, his opponent withdrew from this CC event), so I'd say this game doesn't belong. |
|
| Sep-21-12 | | benjinathan: <cg> you may want to know that there is some suggestion among team black in the Albin game that the game could have been advertised better. Not sure if that is true or not but it is something to think about for the next time. |
|
Sep-21-12
| | chessgames.com: <SwitchingQuylthulg> Thanks for the report. A quick check reveals that your new games actually were processed correctly (no extraneous moves were computed at the end) except now you've brought to light that there seems to be an analogous last-move-bug in a different program. These results are a little confusing (I tested that Winawer game yesterday with success, so what's the difference?) but we'll walk through it again and figure out what the deal is. Thanks for your keen scrutiny. |
|
Sep-21-12
| | chessgames.com: <benjinathan> Thanks, we hadn't seen that. |
|
| Sep-21-12 | | shivasuri4: <chessgames.com>, someone apparently gifted me a premium membership yesterday, and I don't know who it is. Could you please tell me who it is (via mail, if not here), or would that be a violation of the user's privacy? Also, whoever it is that has gifted me the membership, thanks a lot, although I have no idea what I have (or haven't) done to deserve it! |
|
| Sep-21-12 | | rapidcitychess: <shivasuri4>
I once asked, when I was gifted with a membership. It is against their policy, apparently. If your donor wants you to know, (s)he'll let you know. |
|
 |
|
ARCHIVED POSTS
< Earlier Kibitzing · PAGE 558 OF 1118 ·
Later Kibitzing> |
Kibitzing
