I solved the maze! Code=210 Judy Riddle, Vince Goodman, Alasdair Trotter, Ben Krage, Jan Chuvi, Nick Gardner, Eva Li, Tom Winchester, Andrea & Brian Boltwood, Tom Winchester, Kwx, Mark J. Tilford, Matthew Urick Carl Hoff: Just wanted to let you two know I figured it out. The code for maze 10 is 276. I just needed a nights rest to allow me to take a fresh look at the puzzle. By the way there are actually two solutions as I can swap the order I hit two of the goals. Was this intentional? Jayasankar G Nair: I solved the maze. It's easy. :) Cyndy: Wow that was a really interesting maze. J It was fun. On to level 10!!! Adrian Tuck: My name is Adrian and it took me about an hour to solve the thing. Joseph DeVincentis: Quite a challenge. Hi!^_^ My name in Y.F.T. I live in PinTung in Taiwan. David Pollitt: Woooo Hoooo!!!!! Allen Corcorran: Fan of Robert Abbott. Took me a bit, but I got number 9. Woohoo! Alan Kwong: Whoo-hoo! Oriel Maxime: Very nifty. Darwin D. Smith: Thanks Ed, Robert A., Andrea G., et al.; I love these mazes! Scott Purdy: These are very nice. I'm not convinced they're as hard as Robert suggests, though 10 looks truly frightening. Andrea Gilbert: Yes... that was just as hard the second time! Very sneaky of Bob to save it long enough that I had to solve it all over again :) Chris Dougal: I have done it! Thank you. Paul Gould: These things are addicting; I couldn't stop working on one until I solved it. Clint Weaver: That one was pretty darn challenging! But certainly lots of fun! ---------------------------------------------- Dear Ed, I hope you read all of this message. I combined two topics into one e-mail. 1. Thanks to Robert Abbott and you: the Eyeball Mazes were lots of fun to solve*, by hand and on the computer. Since I am not terribly good at puzzling, I ended up giving up and writing a program to solve the mazes; among all the uninteresting programs I write, this was refreshing and very fun. (Solving all ten mazes took a combined total of under .2 seconds.) 2. I was solving Dyson's Telescope Mazes, because for me they were even more fun than the Eyeball Mazes and because I thought writing a program to solve them was entirely nontrivial. So here's my question: has anyone reported (or have you gotten) below-par results on any of the levels? I got a 17 on the par-18 level 24, and the scorecard feature of the applet seemed to confirm it. /Unfortunately/, I have tried to repeat the feat and failed to get below 18. Therefore, I now think my "accomplishment" was a fluke in the program, because I was clicking pretty fast when I got 17. (And I'm willing to admit that maybe I didn't really get a 17 "legally".) Echoing my questions: Have you gotten below par on this or any other level? (I didn't go higher than 25 - maybe par is not always optimal.) Has anyone reported that they had done so? I really want to write a program that will either find a 17-move solution or "prove" that none below 18 moves if possible, but I am burdened by other work /and/ I think it's harder than it looks. I might try over the next few days, late at night or something. If you'd like, I can let you know what happens. Sincerely, (and thanks for listening) - Boris Alexeev. -------------------------------------------------- Woot. :-) Nice little puzzle, I have to admit -- I saw a few of the features right away (the 'red loop' in the upper half and the need to go through it counter-clockwise to solve the upper portion of the puzzle, meaning that the lower half needs to be cleared first) but it took me a little while to hit on reversing direction through the upper half to get back to the lower half properly oriented, and then a few minutes more to find the specific mechanism for clearing the lower half while keeping access to the squares that needed to be reached. Very cool. Incidentally, on a side note: I'm not sure I agree with you that telescope mazes are efficiently solvable. Given that scopes can modify the positions of other scopes, it suggests that the configuration space is exponentially large, and while I don't have enough experience to be able to put together a proof immediately I wouldn't be surprised to learn that it's NP-hard at least, and quite possible PSPACE-hard. Steven Stadnicki --------------------------------------------------- Eric Solomon http://www.ericsolomon.co.uk/ Having failed to solve eyeball maze 9 by 'eyeball' I convinced myself that there was an error in its presentation. So I wrote a program to solve this type of maze. It solved 9 in 50 seconds. So I am impressed by Boris Alexeev whose program solved all 10 in 0.2 seconds. Admittedly my program was in Visual Basic 3 which is not all that fast, and it spends time 'economising' the solution. Anyway, I have the feeling that Bob Abbott is not all that keen on programs that solve his mazes. But being an honest sort of fellow I didn't submit the solution to your site ;-). I agree with Steven Stadnicki that writing a program to solve telescope mazes is not as easy as I first thought. But even if they are NP-hard the small scale (of those I have seen) should mean that it is feasible. By the way it seems unlikely that anyone would beat par if the mazes are designed backwards from a solution, and the number of steps quoted as par.