I agree about the recent genre of questions Marilyn has been answering. Most of them are pretty lame.

I do like the Numbrix game as well. But can it be more challenging? I love the concept, but so far it’s just too easy to be enjoyable. Maybe she will step up the “challenge” factor after it’s been introduced for a while? I think a bigger grid would be more challenging, as would allowing diagonal paths.

Who ever wrote the lead in for this page should use spell check or go back to school.

Just this page? Hey, I’m doing pretty good. I thought, you know being a computer guy and all, that every page was likely to be suffering from frequent gross errors.

My wife surprised me a little bit ago by actually reading the site and pointed out a few errors on the post at the time. I showed her how to edit them for me. I guess what I’m saying is I do have a part time editor.

Hi Joe,
I found your website by googling
“Numbrix”. I was wondering from the name of (your?) website if you or your subscribers have thought of making your own expanded version of Marilyn’s new game? You see, I tried making my own, but it kinda defeats the purpose if you already know the answer! What kind of program can you use to create a grid on the computer?

Ah ha! What a great idea. Write our own numbrix. Or even better, could a program be written that output numbrix? The task is to write a program that builds an uninterrupted path that completely fills a grid (for bonus points, have that grid be variable size) and then only output the numbers around the edge. Naturally I can see situations where this will generate lame numbrix’s but at the same time the pseudo-random flailing of the computer could surprise. Of course an excellent generator could check to make sure that the numbrix forced a ‘unique’ solution as a clue as to whether it generated a lame puzzle, tho I’d want to see the ones that had more than one solution; of which we’ve already seen if you’ve been following it daily.

Now, if you’re doing it by hand you can naturally use something as simple as an spreadsheet for making your grid, if that’s what you were asking.

I’ve set myself to the task of writing a Numbrix generator and wanted to give an update of the process. Making a Numbrix starts out by making a path that winds through a grid and occupies the whole grid. It turns out this is called a “unicursal” maze. I know this because two forums (pretty much at the same time) pointed me to the excellent Think Labyrinth website.

The algorithm described on Think Labyrinth for generating Unicursal mazes I found entirely too restrictive, so I put myself in contact with Walter Pullen, the man behind Think Labyrinth. In an e-mail to me he had this to say:

There’s another technique of producing Unicursal Mazes in “Daedalus”, although it takes a few steps. Starting with a single line across the Maze area (easiest way is to make a standard perfect Maze and then solve it, leaving just the solution path). Then make that line buckle in the middle, randomly spreading throughout the area like yarn being stuffed into a box (this is accomplished with the “Tweak Passages” command, when configured appropriately). This is slow, tends to produce a Maze with a texture similar to intestines, and leaves a few unfilled cells at the end, however it is a completely alternative approach.

Now that’s more like it, only I have to figure out how to avoid the few unfilled cells. I have a theory that involves insuring that there are no contiguous blocks of an odd number of unfilled cells. There are nuances that have to be worked out still, but on paper this approach looks pretty good.

The Parade site requires installation of extra software. Does anyone know of an online site where Numbrix puzzles can be played directly? Ed

The Parade site simply uses Java. As far as playing them directly, it’s in print in the parade magizine.

Today’s Numbrix fiddles with the formula. Two Numbrixs, the first is 9×9 and meander’s somewhat more than usual, allowed by the greater size making it slightly more challenging.

The second puzzle is also 9×9, but instead of telling you just the outside edge it goes for a checkerboard pattern around the edge. Watch out Marilyn, afew more rules changes and you’ll you’ll be making Hidato’s before too long. Actually, I had thought to make a few format changes as options to my numbrix generator, which I will get to eventually.

I also enjoy the Numbrix, but find them too easy. I enjoy Nonograms much more.. on that note, and I’ll apologize if there’s a better place to ask this, but what is a good book of nonograms to purchase? I have “More O’ekaki” and I’m almost through it, and was wondering where to turn now.

Thanks!

Nonograms (aka paint by numbers aka Picross aka griddlers aka figurepic aka…) rock. I’ve seen a few books of them, but I’m not so into them that I can recommend any.

I found a site while searching for Nonograms (http://www.puzzle-nonograms.com/). Their Nonograms are random which kinda defeats the point, but they have other logic puzzles that are pretty cool that I’ve never seen before. I like this one.

A few comments above I mentioned format changes. Today’s numbrix was EXACTLY what I was thinking of, but I didn’t spell it out so I don’t get to say I predicted it.

So to insure that won’t happen again, here are a few more changes I predict:
o Move the ring one more step inside.
o Leave the ring where it is but only do every other hint.
o Do the corners and 4 blocks in the middle (for even edge sizes they’d be a square, for odd sized they’d be the 4 corners surrounding the middle, leaving a cross shaped gap.)
o Just to random squares (tho that’s getting a little too Hidato.)
o Two lines, one that goes 1, 2, 3, and another that goes a, b, c. No guarantee on how long they are. The nature of the puzzle is that they will be partially running in tandem, but there will be “bunching” on one or another. Plus if you make one enumerating up and one enumerating down it can obfuscate matters somewhat.

This looks to be a rip-off of Hidato:

http://www.hidato.com

please let me know where i can purchase the Numbrix puzzle. It would be a gift for my mom.

Thanks

Well, glad to see that traffic is being driven to my site from this keyword at least.

Numbrix are puzzles created by Maryln Vos Savant and distributed through Parade Magizine. I don’t produce Numbrix puzzles. They’ve only been running a short time now but I would not be surprised to learn that they will be producing a book soon. It’s just a matter of patience.

Let’s set the record straight. She didn’t create this. The original is called HIDATO invented by an Israeli– check out http://www.hidato.com for proof. I hate when some take credit for others inspirations. But if you want an addicting math puzzle challenge, KENKEN is the answer. http://www.kenken.com

I don’t think MVS ripped off anything intentionally. More likely Numbrix was developed independently of Hidato (which has been linked to before on this page). Kenken is not a bad puzzle, except unlike Numbrixs Kenken requires math, the exact sort of thing Numbrix’s were designed to steer away from to appeal to beginners.
Personally I prefer Shikaku or Bridges.

(Another reader who got here from a google search :^)

I tried to do the java-based Numbrix game at the parade web site, and kept seeing … in boxes. I finally realized those are 2 digit numbers, and the font they are using is too big for the boxes they are using. Grr!

Your post and ensuing comments was useful, because it directed me to challenging and fun games. Thanks!

As I asked a month ago (above), why can not the Parade site be set up so that I do not need to install anything more (with security risks, etc.) in order to play the current (or other) puzzles online?? Almost all other sites do this, and surely Parade must have the funds and/or access to technical help to set it up.

Also, I do not see a need to change the format of the Numbrix puzzles (location or number of initial givens). (I have looked at some of the “Japanese games” sites mentioned, and those games do not seem identical, even if they may have inspired the Numbrix.) A puzzle should be designed so that there is known (provably) to be only a single solution to each. Marilyn’s puzzles have increased from 7×7 to 8×8 and now 9×9: larger ones would take a bit longer to solve, but I do not think they would bring new thinking into play (I am not convinced that the huge sudoku-formats added much to that puzzle.)

I surmise that Numbrix may run their course in a few more months, since they are highly repetitious: about half of the open squares may be filled by forced moves looking only at neighboring cells; then one just needs to see what chains are not placed, and experiment with possible connections to decided cells.

I will say the more I try KenKen, the more addicting it is, and definitely has the sustained play value that Numbrix lacks. At first i thought the math component would be a turn-off, but actually it’s surprisingly and amazingly fun.

Numbrix is a great game.. but the one you have shown is not sovable - the 60 is trapped and no way to get 59, 58, etc to it..
Look again. This is one of the early ones. It’s 8×8, so it ends on 64. Give it another try.