< Earlier Kibitzing · PAGE 64 OF 64 ·
|Jan-02-13|| ||jepflast: Happy New Year, all!
<YouRang> Those are very good suggestions. During the last game I think I did figure out how to detect transpositions. When a node is added that is identical to an existing position, I was going to have the code add a comment to that node saying that it might be a transposition. Further, I could add a button to "swap" all of the continuing analysis from one of those nodes to the other. (If there's more than one identical position, well, I'm not sure what to do yet!)
Idea #4 is really cool. I'm sure it's possible, but it may not be with the technology I'm using. There would have to be a database lookup triggered by the mouse-over. I'll have to investigate this.
|Jan-05-13|| ||Tiggler: Hi <jepflast> and Happy New Year.|
I reviewed <YouRang>'s post and also like (1) and (4).
With regard to transpositions: this is clearly the thorniest issue, and the reason is that a chess analysis is actually not a tree at all but a directed graph:
A singly connected directed graph is a special case, which may be displayed as a tree. A k-connected graph is a graph in which at least k nodes must be deleted to disconnect the graph into disjoint parts. A tree has k=1, but if there are transpositions k>1. Now we run into the problem that if k>1 the graph cannot be displayed in 2 dimensions without crossing connections. Pretty significant issue.
There are work-arounds, that I think you are considering. Best wishes for success, and even without, the AT is still a great tool in practice.
|Jan-13-13|| ||cro777: <The Caissars: The 2012 Caissar for Most Constructive ... second place yet quite deserving jepflast.>|
Congratulations, <jepflast>! You deserve more than an honorable mention.
|Jan-13-13|| ||hms123: <jepflast>
<hms123>, my friend who is very constructive>
Thanks. That means a lot.
|Feb-06-13|| ||karpkasp: <jepflast> As a former member of Team White during the Thematic Challenge, I can confirm that you are very contructive. I really enjoyed this game and I have devoted almost all my free time to play it.|
Many thanks for your double (!) nomination in the Caissar Awards. This was a big surprise for me:)
Unfortunately, we couldn't win this Albin game :(
|Feb-08-13|| ||jepflast: <karpkasp: I really enjoyed this game and I have devoted almost all my free time to play it.>|
Yeah, me too. Remind me not to do that again! :)
|Sep-09-13|| ||Boomie: <jepflast>
Long time no see.
Can you set up the AT for Chessgames Challenge: S Williams vs The World, 2013?
|Sep-25-13|| ||Tiggler: <jepflast> I was looking at the AT for the Chessgames Challenge today, immediately after our fourth move was posted, when something magic happened.|
The transposition moves that were going to disappear got transferred instantly to the tree nodes after the "last move Played". Did you do it by hand, or have you implemented some smart code?
I had started out marking the transpositions with ! in the comment boxes. Was this unnecessary?
Anyhow, whichever way it works, thanks!
|Oct-02-13|| ||jepflast: <Tiggler> I really haven't been following the game. I can only guess that someone performed the transposition by hand. There is nothing automatic in place.|
I still like using the ! to indicate transpositions, and I can still perform transpositions of nodes by request.
|Oct-03-13|| ||hoodrobin: <there is no password required right now>|
|Oct-06-13|| ||jepflast: <Tiggler>
Wait, I think it's supposed to be * to indicate transpositions.
|Oct-22-13|| ||mistermac: <jepfast>.
What a great utility!
Now I know better how to use it to good effect.
|Oct-30-13|| ||Tiggler: <jepflast: <Tiggler>
Wait, I think it's supposed to be * to indicate transpositions.>|
Yes. I think I introduced that convention, after noticing that * was not used for anything else. But I forgot.
|Nov-08-13|| ||Boomie: <jepflast>
I noticed a problem with the tree.
After 15. Qf3 Nb4 16. 0-0-0 Nxa2+ 17. Kb1 Nb4 18. Qf5 d3 19. c4 Qxf5 20. gxf5 f6 21. Rhf1 Kf7 22. f4 exf4 23. Rxf4 h5 24. Rf2 a5 25. Rg2 h4 26. Rdg1 Rag8 27. Rg4 Rh5 28. Rf1 Nc6 29. Nc3 Bb4, one of the offered moves is Nd5+. However Nd5 is not check.
|Nov-09-13|| ||jepflast: <Booms> Excellent bug! I found the problem and repaired it. I even made you a new Nd5 node with your old comment.|
|Nov-10-13|| ||Boomie: <jepflast>
I expected to see distress calls here but no.
In the slight chance that you're not aware, the AT web site is down today. I get a can't connect message.
|Nov-10-13|| ||WinKing: Hi <jepflast>. The AT appears to be down right now. Has been down for at least a few hours(maybe 4 or 5) that I know of.
Not sure what the problem is. Thought you might want to check it out. Might be connected with what you & <Boomie> were talking about(or not?).|
|Nov-10-13|| ||YouRang: <"Oops! Google Chrome could not connect to www.jepflast.com"> :-(|
|Nov-11-13|| ||jepflast: Dear Team: It's 1:09 AM EST, and I have just confirmed that the AT is down. The other sites hosted on the same server that I know of are also down. At the moment, I don't know the problem, but I hope to be able to provide another update in the morning or afternoon. Hopefully it will be up and running again by then, but who knows.|
I'm posting this on the main game page as well.
|Nov-11-13|| ||YouRang: Back up! Thanks <jepflast> :-)|
|Nov-11-13|| ||jepflast: The server has been repaired.|
|Nov-27-13|| ||mistermac: I have suggested that transpositions could be noted by using FEN as a sort key, with the other field as node number. |
You would find duplicates with the node numbers noted.
|Nov-27-13|| ||YouRang: <mistermac> I suppose that <jepflast> has considered ideas on how to address transpositions, but I'm sure that there's nothing trivial about it.|
I think the <ideal> solution would be for the AT to be position-based rather than move-based. That is, each node represents a position rather than the specific ordered series of moves by which one arrives at that node.
That way, different analysts working on different lines might add a new move, only to discover that it leads to a node that was already added by someone who arrived at that same position some other way. In fact, this could lead to tremendous insights into the game and even resolve a number of potential disagreements.
A node-position would be defined as all attributes of the position (position of pieces, whose turn to move, castling rights, en passant options) *except* for move number. That way, one person could arrive at a position on move 23, and another could get there (some very different way) on move 27.
While, ideal, it would create its own little hornets nest of issues. One being that comments will contain analysis lines based on a given move number, which might not be correct for someone else. Also, when at a node that was adding in two or more different lines, there would need to be a way to look at all the different lines that got there (at least until some of them got pruned by the actual game progression).
BTW, I'm not really offering this as a suggestion. I'm sure it would be an enormous job, and there may be other problems beyond the ones I mentioned. As always, the AT as it stands is a fantastic tool for storing analysis.
|Nov-27-13|| ||mistermac: Of course! I am a newbie to the wonders of NewChess, and FEN is not identical with Position, but is "pretty close to it". In other words, a FEN contains most of the elements to define position by its very structure.|
This leads to my "new" idea of a "verticle" FEN as against the traditional "horizontal" FEN.
A verticle FEN would make for the identification od things like Open Files, and Backward Pawns.
The creative juices *are* flowing today!
|Dec-23-13|| ||YouRang: <jepflast> Another thought on the subject of transpositions.|
Something that *might* be doable (hopefully not too much work) in this regard is to add a button to the main "position" page that says: "Search for this position".
Clicking this would just scan the tree looking for nodes with the same position as the current one, where "same position" means everything is the same except not necessarily the same move number.
If it finds any matching positions, it just displays a list of node numbers where the match was found. An extra bonus would be to show how much analysis exists at each matching node (perhaps the number of nodes further down that branch?).
This alone would probably suffice to handle the transposition problem. In our current game with GMSW, we have seen a few cases where the same position could be reached a multitude of different ways, and unfortunately, such positions are rather burdensome to analyze within the AT.
< Earlier Kibitzing · PAGE 64 OF 64 ·