|
< Earlier Kibitzing · PAGE 2 OF 2 ·
Later Kibitzing> |
Sep-13-10
 | | Stonehenge: Happy Birthday :) |
|
| Mar-20-12 | | Caissanist: Article in today's New York Times about a computer program that Regan has written to help detect computer-assisted cheating: http://www.nytimes.com/2012/03/20/s... . |
|
| May-28-13 | | jackpawn: Isn't this the guy that used to play Owens Defense with success? I seem to remember a win against Browne with it. |
|
May-28-13
 | | perfidious: <jackpawn>: Per the kibitz above by <Caissanist>, this was indeed the case-believe it was published in CL&R at the time and may have even made Informator or The Chess Player. |
|
| Jun-17-13 | | jackpawn: Wish more of his Owens games would appear here. As I recall his games were very interesting. |
|
| Jul-28-13 | | DoctorD: A recent and chilling article by Regan:
http://rjlipton.wordpress.com/2013/... |
|
| Jul-29-13 | | Strongest Force: Ken was a very deep thinker. We shared some interesting thoughts about where evolution was taking the human race. He was 14 at the time. |
|
| Sep-07-13 | | Karpova: His website: http://www.cse.buffalo.edu/~regan/ |
|
| Sep-13-13 | | DoctorD: In honor of Dr. Regan's work against cheating, I want to dedicate this original chess problem(*) to him, and also to test the waters in publishing chess problems in media other than traditional sources, something I am sure he would support Steven B. Dowd
Original to Chessgames.com
Dedicated to Dr. Kenneth Regan on his Birthday
White to Play and Mate in Three
click for larger viewA Rex Solus problem. As a hint, a Rex Solus (black king only) should contain interesting play and a nice variety of mates, with the black king given as much opportunity to flee as possible. I hope I succeeded. (*) I checked the Chess Problem Database Server to look for any possible predecessors and none appeared - it would certainly be embarrassing to salute someone for high moral standards with a plagiarized problem! |
|
| Sep-13-13 | | DoctorD: For those interested in composing, here is an example. After one line in the above, you essentially have a y-flight #2 that can be expressed through a number of board rotations and keys. I had to search for those too, and I certainly found examples of those, the oldest from 1915 and one example from this years. The earliest example:
Gunnar Gundersen
The Australasian 12/1915
White to play and mate in two
click for larger viewYou'll note this is a rotation of my position after one move. My decision was whether the threemover showed enough content to be considered relatively original. I think so, but also posed the question on several fora. |
|
| Sep-19-13 | | Kinghunt: I have just read Regan's paper detailing his system of intrinsic performance rating: http://www.cse.buffalo.edu/~regan/p... Overall, I was quite impressed. It seems to do a good job of matching play with playing strength. For instance, the recent Anand-Gelfand match has by far the highest level of play, which makes sense given that virtually all moves played in that match were chosen in advance by machines. Certain limitations apply in trying to make use of this system, however. The confidence intervals are very wide, and on several occasions, it picks the losing player as having played the better chess. However, that simply means over-zealous fans should not try and use this to compare players throughout history. It has more than sufficient resolution to achieve its desired purpose, namely, to identify the use of computers in chess. (And it does, in the Anand-Gelfand match.) |
|
| Feb-22-14 | | RedShield: <<In this concrete instance, however, one can make a pretty good nay-case. According to Wikipedia, the "Shannon Number" for the game-tree size of chess is estimated between 10^120 and 10^123. Right in that range is the "Bekenstein Bound" of 10^122 bits for the observable universe, which as explained by a colleague here (http://www.scottaaronson.com/democr...) is a limit on the size of any computation. Put another way, if a 32-piece Nalimov tablebase existed---even if it would be serviced by Rybka clusters in other galaxies---our entire Universe would collapse into a black hole!> http://www.chessninja.com/dailydirt... |
|
Feb-22-14
 | | alexmagnus: <RedShield> That is not true however as to compile a 32-piece tablebase one doesn't need the entire game tree - transpositions get excluded. Say, all games starting with 1.e4 e6 2.d4 and 1.d4 e6 2.e4 don't have to be listed double. The number of <positions> in chess is much smaller - about 2*10^46. |
|
| Feb-22-14 | | RedShield: Yes, but don't the transpositions have to be generated before they can be excluded? |
|
Feb-24-14
| | alexmagnus: Why should they be? Take my example. Imagine you already have an analysis of all possible 1. d4 games. And start with 1. e4. You come up with 1.e4 e6 2. d4 and see this position already occured in your 1.d4 analysis... So you don't need to calculate <anything> of this branch! In other words, you ruled out zillions of games while you generated only one position. |
|
| Feb-24-14 | | RedShield: But to see that a position has already occurred you - or the mindless computer - has to generate it to understand that there was no need to generate it in the first place. |
|
Feb-24-14
 | | HeMateMe: Yep, mainland china is putting the screws into Hong Kong, despite promises that they would have free press, after England left. <http://www.cnn.com/2014/02/24/world...> |
|
| Feb-24-14 | | RedShield: Did you mix up Ken Regan with Ken Rogoff? |
|
Feb-24-14
| | alexmagnus: <Red> But you generate only <one> position - and exclude <many> games by having generated it. In this case, all games starting with 1.e4 e6 2.d4. You don't have to generate any of the positions/games starting with this sequence anymore. While they are still there in that 10^120 calculation. |
|
| Feb-24-14 | | RedShield: < But you generate only <one> position > Yes, but you've generated it twice.
<You don't have to generate any of the positions/games starting with this sequence anymore.> But you've generated them already. |
|
Feb-24-14
| | alexmagnus: <But you've generated them already.> But I didn it once. While in the 10^120 calculation they are taken twice. I don't know how many calculations are necessary to compile a 32 piece TB, but with this reasoning it will be only slightly bigger than the number of legal positions. So it will be closer to 10^46 than 10^120. |
|
Jun-22-14
| | Stonehenge: Interview:
http://www.npr.org/2014/06/21/32422... |
|
Feb-12-23
 | | Sally Simpson: A picture here of Ken Regan (Black) playing Nicholas Ivell https://www.ecforum.org.uk/viewtopi... The spectators are Leonard Barden, Henry Mutkin, Adrian Hollis and Bob Wade. The game is not here but on Britabse: https://www.saund.co.uk/britbase/pg... Use this version. The other version in print in the BCM May 1983 had the colours wrong. |
|
Feb-12-23
 | | MissScarlett: It is now: N Ivell vs K Regan, 1983 |
|
Feb-12-23
| | Sally Simpson: That is it! |
|
 |
|
< Earlier Kibitzing · PAGE 2 OF 2 ·
Later Kibitzing> |
Kibitzing
