< Earlier Kibitzing · PAGE 690 OF 690 ·
|Mar-05-14|| ||Blunderdome: <Shams> Can't you test your own link by logging out and visiting the link, or opening another browser and visiting the link?|
|Mar-05-14|| ||davide2013: I found a problem in "openings" on how to search for games, based on transpositions. Here the problem, I was looking for games under this line:
1.b3,d5; 2.Bb2,Nf6; 3.e3,e6; 4.Nf3,Be7; 5.c4,c6; and for the database there are 0 games.
But if we use this move order:
1.c4,c6; 2.Nf3,d5; 3.b3,Nf6; 4.Bb2,e6; 5.e3,Be7; the openings database gives 4+4 games based on 6.Be2 and 6.Qc2, and then when I click on: "search database for this position" it comes out with only 5 games.
|Mar-05-14|| ||SwitchingQuylthulg: <davide2013> That's known behavior and isn't really a problem. It's simply the way the Opening Explorer handles transpositions. |
The Opening Explorer is based on nodes (i.e. positions) and links between them, not on moves or move orders. If at any node somebody plays a move that leads to another node, a link is created between those two nodes.
Node 2734724 (Opening Explorer) is linked to two further nodes, node number 2734725 and node number 2325972 (6.Be2 and 6.Qc2 respectively), both of which have been reached in four indexed games. Four games provide the link between nodes 2734724 and 2734725 and only one the link between 2734724 and 2325972 - the other three games at 2325972 are being transposed into - but one game is all that's needed for a link.
5...c6 isn't listed at node 404002 (Opening Explorer) because it wasn't played in any of the 153 games at that node and there is therefore no link between nodes 404002 and 2734724.
|Mar-05-14|| ||SwitchingQuylthulg: <cg> I did chance upon something that seemed strange to me, though.|
This may look suspiciously like the initial position... but it's actually node 4663975: Opening Explorer
Why is node 4663975 listed as a node separately from the starting position? Is it because the starting position is some kind of pseudo-node? (In lists of nodes like <nodes=2.417183.417458.416885.2328285.673289> the initial node is never listed.)
On a possibly related matter, in the FEN string of the initial position <Black> is given as the side to move:
rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR b KQkq -
By contrast, the FEN string for node 4663975 correctly gives White as moving:
rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq -
Indeed, if one searches by FEN and uses the correct initial FEN one will wind up at node 4663975; to get the initial node one has to give Black as the side to move.
|Mar-05-14|| ||davide2013: The problem is that I used Chess Assistant 12 to look for games using my 1.b3 line, and it found the games that began with 1.c4, and that at the 5th move would have the same position of the line which began with 1.b3.
So in the beginning I was thinking chessgames.com was missing those games, but they were not missing. I don't know if it is an index problem, but if at the 5th move they are the same positions, then it doesn't really matter where they came from. As example also in the French exchange there are a lot of transpositions, in the lines with 4.c4 which are also coming from the queen's gambit, and that is a problem for a database, if it doesn't give you what you are searching for. Bottom line: is it fixable? Because if it is not, then a database program is more reliable.|
|Mar-05-14|| ||SwitchingQuylthulg: <davide2013: The problem is that I used Chess Assistant 12 to look for games using my 1.b3 line, and it found the games that began with 1.c4, and that at the 5th move would have the same position of the line which began with 1.b3. So in the beginning I was thinking chessgames.com was missing those games, but they were not missing.> |
The Opening Explorer can handle transpositions quite well. Given an opening position, it will find all indexed games that reach that position, regardless of the move order - indeed, the five games we have that reach the position after 5...Be7 arrived there by quite different move orders:
1. c4 c6 2. Nf3 d5 3. b3 Nf6 4. Bb2 e6 5. e3 Be7 (B Sikanjic vs Vedran Mesin, 2001 and
B Tot vs L Asztalos, 1938)
1. e3 e6 2. c4 c6 3. Nf3 Nf6 4. b3 Be7 5. Bb2 d5
(R Bouhallel vs P Poslednik, 2001)
1. c4 e6 2. Nf3 d5 3. e3 c6 4. b3 Nf6 5. Bb2 Be7
(L Aroshidze vs A Palekha, 2005)
1. Nf3 d5 2. c4 c6 3. e3 e6 4. b3 Nf6 5. Bb2 Be7
(M Medina vs W Elliott, 2006)
This should show the Opening Explorer has no problems with transpositions. But it can't find transpositions that never happened because they aren't there - and if the Explorer started seeing things that aren't there, that would cause as many problems as it solves.
|Mar-05-14|| ||davide2013: Whatever you want to believe man, it is ok.|
|Mar-05-14|| ||davide2013: I tried on chesstempo, with this sequence of moves: 1.b3,d5; 2.Bb2,Nf6; 3.e3,e6; 4.Nf3,Be7; 5.c4,c6; and 8 games come out (like in chessgames.com database, with the difference that they are really 8), the first one is Tot Bora vs Asztalos, Lajos of 1938.
two things are strange:
1. Chesstempo games database didn't know it shouldn't find something which didn't happen.
2. The mistake on white's player name is the same, so who knows who copied it from whom. Since the white player correct name is: Toth Bela, not Tot Bora.
|Mar-05-14|| ||SwitchingQuylthulg: <davide2013: As example also in the French exchange there are a lot of transpositions, in the lines with 4.c4 which are also coming from the queen's gambit, and that is a problem for a database, if it doesn't give you what you are searching for.>|
In such lines you'll find the Opening Explorer will in fact quite happily give you what you are searching for. The same goes for other common (and uncommon) transpositions... it even suggests 2.e4 as the apparent most common reply to the From Gambit (1.f4 e5) - Opening Explorer - simply because that transposes into the King's Gambit.
If in the position at Opening Explorer you want the Explorer to suggest 5...c6 as a possible move - even though it's never been played in that position in this database - and call it a "problem" needing to be "fixed" that it does not... well, it <could> be "fixed", but only at an enormous increase in either the space required by the Opening Explorer or the computational time and power required by the Opening Explorer or both, and doing so would create as much confusion as it would clear up.
Finally, always keep in mind that if you want to find all games that reach a given position, you can simply search for that position - by FEN string - and the Explorer will give you just that, just as a database program would.
|Mar-05-14|| ||SwitchingQuylthulg: <davide2013: I tried on chesstempo, with this sequence of moves: 1.b3,d5; 2.Bb2,Nf6; 3.e3,e6; 4.Nf3,Be7; 5.c4,c6; and 8 games come out (like in chessgames.com database, with the difference that they are really 8), the first one is Tot Bora vs Asztalos, Lajos of 1938. two things are strange: 1. Chesstempo games database didn't know it shouldn't find something which didn't happen.>|
In their database (which is larger) the move 5...c6 did happen, providing the critical link between those two positions. (The specific game is http://chesstempo.com/gamedb/game/2....) If that game were added to the Chessgames database then the Chessgames Opening Explorer would also provide 5...c6 as an option - by the same token, if you limit your ChessTempo search to 2300+ vs. 2300+ then 5...c6 will not be offered as an option even though five of those eight games you mentioned were between two 2300+ players. The two Opening Explorers behave identically.
|Mar-05-14|| ||chessgames.com: I follow what both Switching and davide are saying. A few points:|
• As Switching rightly points out, the Opening Explorer doesn't make a link between positions unless there is actually a game where that transposition took place. There are arguments both for and against this methodology, but over the years we've come to consider that this is the best way to do it.
• Transpositions are trickier than most chess players realize. When work on the Opening Explorer started many years ago I felt that I firmly understood the concept of transpositions as well as any grandmaster; after all it's just a mathematical structure and not a specific chess talent. However, since those carefree days, I've been greatly humbled and now see how very slippery some of the nuances can be.
• We have an Opening Explorer FEN Lookup that sometimes can reveal transpositions that might be missed using a move-by-move approach.
• Switching's discovery of that odd node 4663975 is a surprising anomaly. It may be difficult to discover how that crept into the database, but I remember processing a game that started 1.e5 or something like that, so perhaps a broken game left a nonsensical stranded node in the tables. I'll look into it a little deeper and see if it's tied to a specific game.
|Mar-05-14|| ||SwitchingQuylthulg: <davide2013: 2. The mistake on white's player name is the same, so who knows who copied it from whom. Since the white player correct name is: Toth Bela, not Tot Bora> |
It would be interesting to know what you base this claim on. Quite a few games in this database from the 1930s and 1940s were once attributed to Bela Toth, but they were all reattributed to Bora Tot on the grounds that Tot was active in the right area at the right time (and is credited by independent sources such as Chessmetrics as having played in the actual tournaments involved) whereas IM Tóth was born in 1943. Béla Tóth is a common enough name there could be another player by that name, but...
|Mar-05-14|| ||chessgames.com: About the anomalous node 4663975 it seems to have something to do with weird openings such as
Stockfish vs Houdini, 2013. (There are at least 4 or 5 similar games in the database, where the players dance their knights back and forth before actually playing the game properly.) This doesn't fully explain what's going on but it's a hint in the right direction.|
Question: it is a fact that you cannot play a legal game of chess wherein the starting position appears with Black to move, is that correct? No amount of clever knight dancing can create that position?
|Mar-05-14|| ||SwitchingQuylthulg: <chessgames.com: Question: it is a fact that you cannot play a legal game of chess wherein the starting position appears with Black to move, is that correct? No amount of clever knight dancing can create that position?>|
The background story of Stockfish vs Houdini, 2013 is that the organizer decided to play one pair of games with no opening book at all, making the computers think from move one, but all other games were played from pre-decided 16-ply opening variations and that was what the computer interface had been set to expect. So the no-book games started with 16 plies of useless knight hopping :)
|Mar-05-14|| ||notyetagm: <CG.COM>
Can we please have forums for the ongoing <2014 CAPPELLE-LA-GRANDE OPEN> and the <2014 REYKJAVIK OPEN>?
TWIC REYKJAVIK PGN -> http://www.theweekinchess.com/asset...
TWIC CAPPELLE-LA-GRANDE PGN -> http://theweekinchess.com/assets/fi...
|Mar-05-14|| ||notyetagm: <CG.COM>
Also, the <2014 NOTEBOOM OPEN>.
TWIC NOTEBOOM PGN -> http://www.theweekinchess.com/asset...
|Mar-06-14|| ||Check It Out: <CG.com> I'm sure this has been brought up before, but the quality of the Game Of The Day could be vastly improved by uncoupling it from the puns. |
In addition to improving the quality of the games, it would allow you to be more flexible and topical in your choices, for example celebrating top players' birthdays, featuring an exciting game from an ongoing tournament or recently finished tournament, etc etc
Puns restrict the choices and since puns aren't technically chess related and are relatively unimportant that seems a strong argument for for making the change. Also, the GOTD is the top link on your homepage and you want to make a good first impression for new browsers.
That doesn't mean puns need to go away. You could do both.
|Mar-06-14|| ||Everyone: <chessgames.com> May I ask you to look at and take action towards the trolling / spamming /stalking of User: Moszkowski012273 against one of the most respected chessgames.com members. |
|Mar-06-14|| ||Eggman: <Chessgames> Did you overlook my post here: chessgames.com chessforum? I didn't get a response.|
|Mar-06-14|| ||RedShield: <most respected chessgames.com members.>|
Behave! Dumb Americans trolling one another accounts for about 90% of the site content. Do you want to put cg.com out of business?
This should improve site content, however.
|Mar-06-14|| ||chessgames.com: <Eggman> Sorry for the nonresponse but yes, we read your suggestion and pretty much agree. We're working on the Candidates newsletter right now and will probably make some small changes before the event starts, like a logo and links to relevant sites.|
<Check It Out> Puns aren't technically chess related? Puns are relatively unimportant? You are trodding on years of tradition there!
|Mar-06-14|| ||Check It Out: <chassgames.com> You wouldn't need to stop the pun tradition. You could still have a pun of the day. But just think of all the great games that could be GOTD once you no longer had to worry about a pun being formulated for it. It would be much easier for you guys to choose the daily games and their quality would be higher. In terms of tradition the quality of the GOTD must rank higher. This is not to say there aren't many good GOTD, but we all know that clunkers are used frequently simply for the pun. Uncouple them and have the best of both worlds!|
|Mar-08-14|| ||Blunderdome: To be honest, if not for puns, I doubt I ever would have gotten into chess in the first place.|
|Mar-08-14|| ||Shams: <cg> I'd like to second this request:|
<Can we please have forums for the ongoing <2014 CAPPELLE-LA-GRANDE OPEN> and the <2014 REYKJAVIK OPEN>?>
|Mar-08-14|| ||Shams: Thanks that was very prompt. :)|
< Earlier Kibitzing · PAGE 690 OF 690 ·
100% Cotton Chess Puzzle Shirt