{"content":{"sharePage":{"page":0,"digests":[{"id":"65656628","dateCreated":"1389649546","smartDate":"Jan 13, 2014","userCreated":{"username":"Bretwa","url":"https:\/\/www.wikispaces.com\/user\/view\/Bretwa","imageUrl":"https:\/\/www.wikispaces.com\/i\/user_none_lg.jpg"},"monitored":false,"locked":false,"links":{"self":"https:\/\/chessprogramming.wikispaces.com\/share\/view\/65656628"},"dateDigested":1531474005,"startDate":null,"sharedType":"discussion","title":"Complete all magics ?","description":"Could we complete all the magic numbers on this page ? Even if thesesare not necesseraly the best, why shouldn't we put all the magics ? After we could improve them is we find better ones.","replyPages":[{"page":0,"digests":[{"id":"68781056","body":"Sure we could, I don't have them handy. The page was intended for "used" < "bits", since "used" == "bits" magics are simple to determine anyway as suggested in Looking for Magics<\/a>.","dateCreated":"1389655962","smartDate":"Jan 13, 2014","userCreated":{"username":"GerdIsenberg","url":"https:\/\/www.wikispaces.com\/user\/view\/GerdIsenberg","imageUrl":"https:\/\/www.wikispaces.com\/user\/pic\/1202793136\/GerdIsenberg-lg.jpg"}}],"more":0}]},{"id":"4296069","dateCreated":"1211017789","smartDate":"May 17, 2008","userCreated":{"username":"LasseHansen","url":"https:\/\/www.wikispaces.com\/user\/view\/LasseHansen","imageUrl":"https:\/\/www.wikispaces.com\/i\/user_none_lg.jpg"},"monitored":false,"locked":false,"links":{"self":"https:\/\/chessprogramming.wikispaces.com\/share\/view\/4296069"},"dateDigested":1531474005,"startDate":null,"sharedType":"discussion","title":"Magics and square mapping","description":"It seems like Onno's magics are based on another mapping than yours, Gerd. May I suggest to use bit number instead of square to make all magics independant of mapping.
\n
\nRegards, Lasse","replyPages":[{"page":0,"digests":[{"id":"4407139","body":"Hi Lasse,
\noups, sorry I missed your questions until now.
\nGrant posted the magics in WBF shortly before it was locked.
\nI guess the coordinates need to vertically flipped because Grant A8 == 0 mapping, but I did not try. What do you mean with bit number instead of square? Isn't that just another mapping?
\n
\nSkool,
\nGerd","dateCreated":"1211617132","smartDate":"May 24, 2008","userCreated":{"username":"GerdIsenberg","url":"https:\/\/www.wikispaces.com\/user\/view\/GerdIsenberg","imageUrl":"https:\/\/www.wikispaces.com\/user\/pic\/1202793136\/GerdIsenberg-lg.jpg"}},{"id":"4414587","body":"Using the bitnumber shouldn't be ambiguous anymore: For some b2 might be bit#9, for others g2 would be #9 - however, for magicbitboards if we know that #10 is blocked, it doesn't matter whether it is c2 or f2.
\n
\nRegards Edmund","dateCreated":"1211718067","smartDate":"May 25, 2008","userCreated":{"username":"Edmund_Moshammer","url":"https:\/\/www.wikispaces.com\/user\/view\/Edmund_Moshammer","imageUrl":"https:\/\/www.wikispaces.com\/i\/user_none_lg.jpg"}},{"id":"4421599","body":"Is that Ok now?","dateCreated":"1211798061","smartDate":"May 26, 2008","userCreated":{"username":"GerdIsenberg","url":"https:\/\/www.wikispaces.com\/user\/view\/GerdIsenberg","imageUrl":"https:\/\/www.wikispaces.com\/user\/pic\/1202793136\/GerdIsenberg-lg.jpg"}},{"id":"4423757","body":"looks good now
\n
\nsomething else: is there a list of previously generated magic-keys and its bit-demand? eg has anyone generated a sq#9 key with 10 bit before?","dateCreated":"1211819329","smartDate":"May 26, 2008","userCreated":{"username":"Edmund_Moshammer","url":"https:\/\/www.wikispaces.com\/user\/view\/Edmund_Moshammer","imageUrl":"https:\/\/www.wikispaces.com\/i\/user_none_lg.jpg"}},{"id":"4424609","body":"Sure, 10 bit should be simple - we are looking for one bit less ;-)","dateCreated":"1211822039","smartDate":"May 26, 2008","userCreated":{"username":"GerdIsenberg","url":"https:\/\/www.wikispaces.com\/user\/view\/GerdIsenberg","imageUrl":"https:\/\/www.wikispaces.com\/user\/pic\/1202793136\/GerdIsenberg-lg.jpg"}},{"id":"4424859","body":"Indeed, 10 bits was simple, I found quite many of those. But I doubt it can be improved for this position. The best keys I found use 656 index numbers, which is bigger than 2^9","dateCreated":"1211823766","smartDate":"May 26, 2008","userCreated":{"username":"Edmund_Moshammer","url":"https:\/\/www.wikispaces.com\/user\/view\/Edmund_Moshammer","imageUrl":"https:\/\/www.wikispaces.com\/i\/user_none_lg.jpg"}},{"id":"4436455","body":"Just as another idea: wouldn't it as well be important to optimize the number of the highest index? Especially for squares that cant be reduced by another factor 2 we might still be able to squeeze out maybe 30%-50%","dateCreated":"1211896644","smartDate":"May 27, 2008","userCreated":{"username":"Edmund_Moshammer","url":"https:\/\/www.wikispaces.com\/user\/view\/Edmund_Moshammer","imageUrl":"https:\/\/www.wikispaces.com\/i\/user_none_lg.jpg"}},{"id":"4440215","body":"You can, but if you like to use Lasse's latest postmask trick, you have to consider indexrange = max(indexrange(sq1), indexrange(sq2)).","dateCreated":"1211911781","smartDate":"May 27, 2008","userCreated":{"username":"GerdIsenberg","url":"https:\/\/www.wikispaces.com\/user\/view\/GerdIsenberg","imageUrl":"https:\/\/www.wikispaces.com\/user\/pic\/1202793136\/GerdIsenberg-lg.jpg"}},{"id":"4441413","body":"thats true - so we should aim at optimizing sq1 and sq2 :)","dateCreated":"1211916575","smartDate":"May 27, 2008","userCreated":{"username":"Edmund_Moshammer","url":"https:\/\/www.wikispaces.com\/user\/view\/Edmund_Moshammer","imageUrl":"https:\/\/www.wikispaces.com\/i\/user_none_lg.jpg"}},{"id":"30045423","body":"There is one thing I don't quite understand: The numbers shown can't be improved?
\nor is as the title of the pages: best so far.
\nmy point is, what about those squares with an empty magic cell.
\n
\nsorry if my english is not so good.
\nbest regards.","dateCreated":"1289506309","smartDate":"Nov 11, 2010","userCreated":{"username":"dustinluis","url":"https:\/\/www.wikispaces.com\/user\/view\/dustinluis","imageUrl":"https:\/\/www.wikispaces.com\/i\/user_none_lg.jpg"}},{"id":"30047035","body":"You will always find magic factors, to map N relevant occupied bits to an N-bit index range for that particular square.
\n
\nThe challenge is to find factors, with enough constructive collisions yielding in a N-1 range to halve the table size for that square.
\n
\nThat seems not possible for all squares. Empty cells indicate that there is no magic factor so far or not published with property.
\n
\nThis page is also a bit outdated due to Volker's Fixed shift Fancy.
\n
\nRegards,
\nGerd","dateCreated":"1289507679","smartDate":"Nov 11, 2010","userCreated":{"username":"GerdIsenberg","url":"https:\/\/www.wikispaces.com\/user\/view\/GerdIsenberg","imageUrl":"https:\/\/www.wikispaces.com\/user\/pic\/1202793136\/GerdIsenberg-lg.jpg"}},{"id":"30053915","body":"thanks, I understood now.","dateCreated":"1289514260","smartDate":"Nov 11, 2010","userCreated":{"username":"dustinluis","url":"https:\/\/www.wikispaces.com\/user\/view\/dustinluis","imageUrl":"https:\/\/www.wikispaces.com\/i\/user_none_lg.jpg"}}],"more":2}]}],"more":false},"comments":[]},"http":{"code":200,"status":"OK"},"redirectUrl":null,"javascript":null,"notices":{"warning":["You do not have permission to edit this page."],"error":[],"info":[],"success":[]}}