Go Back   Wordtwist Forums > Wordtwist.org Discussion > FAQs and Help Area

Reply
 
Thread Tools Display Modes
  #1  
Old 03-12-2018, 02:09 AM
Myxlplic Myxlplic is offline
Junior Member
 
Join Date: Dec 2017
Posts: 1
Myxlplic is an unknown quantity at this point
Default Inaccurate board stats

A couple of times I have found words longer than what was indicated on the board stats. Just now the board stats had a few six letter words and zeros all the way down for 7s and 8+s, but then I found RESPEAK which was 7 letters and it gave me the 'new record' for longest word.

Any one know why this happens? I like finding the longest word, and often I won't play a puzzle if the longest possible word has already been found (according to board stats)
Reply With Quote
  #2  
Old 03-12-2018, 08:50 AM
Spike1007 Spike1007 is offline
Senior Member
 
Join Date: Aug 2017
Posts: 389
Spike1007 has received several accolades
Default

Quote:
Originally Posted by Myxlplic View Post
A couple of times I have found words longer than what was indicated on the board stats. Just now the board stats had a few six letter words and zeros all the way down for 7s and 8+s, but then I found RESPEAK which was 7 letters and it gave me the 'new record' for longest word.

Any one know why this happens? I like finding the longest word, and often I won't play a puzzle if the longest possible word has already been found (according to board stats)
I've never seen this myself. (I play similarly to you, but generally for longer words. It's harder to tell if a 9+ letter word is longer than what's reported in the stats, especially if you avoid boards that say they're not there.) I always assumed that the stats were accurate, but I'm not about to play a board in the off-chance that they're wrong. Has anybody else ever seen this?

In another thread, we've mentioned programming things for fun. I've idly wondered what it takes to do an efficient dictionary search to find all possible words in a board. It seems like it should be expensive, but it's almost instantaneous.

Last edited by Spike1007 : 03-12-2018 at 08:54 AM.
Reply With Quote
  #3  
Old 03-12-2018, 09:17 AM
bwt1213 bwt1213 is offline
Senior Member
 
Join Date: Oct 2013
Location: Pleasant Prairie, Wisconsin
Posts: 289
bwt1213 is an unknown quantity at this point
Default

Quote:
Originally Posted by Spike1007 View Post
I've never seen this myself. (I play similarly to you, but generally for longer words. It's harder to tell if a 9+ letter word is longer than what's reported in the stats, especially if you avoid boards that say they're not there.) I always assumed that the stats were accurate, but I'm not about to play a board in the off-chance that they're wrong. Has anybody else ever seen this?

In another thread, we've mentioned programming things for fun. I've idly wondered what it takes to do an efficient dictionary search to find all possible words in a board. It seems like it should be expensive, but it's almost instantaneous.
A very long time ago, I posted the same thing -- that I had seen a board with no 9 letter words (or something like that) and then found one. No one else had seen that, and I never had anyone else say they had seen it until now. I had almost convinced myself that I had been confused by seeing a count of zero AFTER I had made my play, when in fact I had found the only word of 9 letters. As far as programming things for fun, when you do something like that sometimes you just get into a zone where it's almost impossible to make a mistake. When I wrote the library of assembly-language routines for the compiler (math, string-handling, if/then/else and all the rest of it), I wrote the whole thing in one pass over a weekend and everything worked the first time. That was back in the days when you got a macro-assembler with your operating system, so writing a compiler was easier -- but still, I was astounded that there were no bugs at all.
Reply With Quote
  #4  
Old 03-12-2018, 09:33 AM
Spike1007 Spike1007 is offline
Senior Member
 
Join Date: Aug 2017
Posts: 389
Spike1007 has received several accolades
Default

I'm always very suspicious if something compiles & runs the first time for me.

Anyway, your compiler & floating-point math stuff sounds interesting. I've basically spent a lot of time writing an interpreter (in Fortran) that would take a bunch of function definitions, partial differential equations and boundary conditions, process them to convert them into internal instruction sets (linearizing if necessary), save them as strings of instructions, then discretize & solve the resulting system on some specified 2D or 3D domain. I'd go in spurts of programming binges (followed by debugging), then waiting for the next demand on the software (like making things time-dependent). At times, it's can be hard to separate work from fun.

By the way, I know the stats beforehand should show the total available words, while stats shown during play are the remaining words for you to find, with the counts updated with every word you play. What happens if you play a word that's not supposed to be there? Does the count go negative?

Last edited by Spike1007 : 03-12-2018 at 09:38 AM.
Reply With Quote
  #5  
Old 03-12-2018, 12:31 PM
bwt1213 bwt1213 is offline
Senior Member
 
Join Date: Oct 2013
Location: Pleasant Prairie, Wisconsin
Posts: 289
bwt1213 is an unknown quantity at this point
Default

Quote:
Originally Posted by Spike1007 View Post
I'm always very suspicious if something compiles & runs the first time for me.

Anyway, your compiler & floating-point math stuff sounds interesting. I've basically spent a lot of time writing an interpreter (in Fortran) that would take a bunch of function definitions, partial differential equations and boundary conditions, process them to convert them into internal instruction sets (linearizing if necessary), save them as strings of instructions, then discretize & solve the resulting system on some specified 2D or 3D domain. I'd go in spurts of programming binges (followed by debugging), then waiting for the next demand on the software (like making things time-dependent). At times, it's can be hard to separate work from fun.

By the way, I know the stats beforehand should show the total available words, while stats shown during play are the remaining words for you to find, with the counts updated with every word you play. What happens if you play a word that's not supposed to be there? Does the count go negative?
The count never goes negative. Either it's not allowed to go negative, or the word was always there and I just thought it wasn't.

For the differential and partial differential equations, wouldn't you just use Runge-Kutta approximations? Yeah, the formulas are pretty complex, but that's not a bar if you're just writing it once.

And as far as things not running the first time, I've occasionally written things that run the first time, sometimes with a whole lot of lines of code, too. It's not exactly normal for me, but it did happen once in a while. I wrote an arbitrary precision calculator one weekend and it worked the first time. Then I tried to get cute and put in a whole bunch of new stuff and broke it repeatedly for weeks. I haven't worked on it for years and I'd have to dig it out, but if anyone needs to know roots and transcendental functions to 70 (or 2000 or more) digits of precision, I've got it. Or I had it. I'd have to see where the heck it is, since I haven't even looked at it in at least 10 years. It was a base 10 package with long integer exponents, so it would go up to decimal exponents of about two billion or so. Handheld calculators are usually good up to an exponent of about 73.
Reply With Quote
  #6  
Old 03-12-2018, 12:59 PM
Asmarandana Asmarandana is offline
Junior Member
 
Join Date: Sep 2017
Posts: 2
Asmarandana is an unknown quantity at this point
Default

Quote:
Originally Posted by Spike1007 View Post
I've idly wondered what it takes to do an efficient dictionary search to find all possible words in a board. It seems like it should be expensive, but it's almost instantaneous.
I'm probably missing something here, but if you want to find "all possible words," won't an on-line boggle-solver do the trick?

The issue, of course, is that whatever solver you choose probably has a slightly different dictionary than the one used here. But if you want to find all possible words in a board with an exact match to wordtwist, I would think the hardest part would be getting your hands on the same dictionary that wordtwist uses.

Or am I misunderstanding what you are saying?
Reply With Quote
  #7  
Old 03-12-2018, 01:15 PM
Spike1007 Spike1007 is offline
Senior Member
 
Join Date: Aug 2017
Posts: 389
Spike1007 has received several accolades
Default

If I knew I was going to find a word that didn't show up in the stats, I'd watch them carefully during play to see what happens. That seems pretty rare though. I should try to watch & see if anything strange happens.

Runge-Kutta methods are typically used for the time derivative. I started out with steady-state problems, so I was always solving a spatial system of PDEs. When I included time dependence, I generally stuck to backward Euler (of varying degrees), but one student did want an implicit 4th order Runge-Kutta, so I put that in. All the implicit methods still require setting up & solving a (spatial) PDE system at each time step, and that was the major focus of the program. That code (which kind of turned into a monster) hasn't been used much for a few years, and it's always a challenge for me to try to do something new with it. It's sometimes easier to write new code than it is to remember what the hell I was doing in old code.

Last edited by Spike1007 : 03-12-2018 at 01:50 PM.
Reply With Quote
  #8  
Old 03-12-2018, 01:30 PM
Spike1007 Spike1007 is offline
Senior Member
 
Join Date: Aug 2017
Posts: 389
Spike1007 has received several accolades
Default

Quote:
Originally Posted by Asmarandana View Post
I'm probably missing something here, but if you want to find "all possible words," won't an on-line boggle-solver do the trick?

The issue, of course, is that whatever solver you choose probably has a slightly different dictionary than the one used here. But if you want to find all possible words in a board with an exact match to wordtwist, I would think the hardest part would be getting your hands on the same dictionary that wordtwist uses.

Or am I misunderstanding what you are saying?
You're right that having a matching dictionary is part of the trick. And, of course you're right that an on-line Boggle solver would do it (given the same dictionary). My question was more about how any Boggle solver (either on-line or the one the WordTwist site uses) works quickly. It just seems like there are a whole lot of possible character strings (most of which are nonsense) in a given board to check against the dictionary, and each one has to be checked. Without shortcuts, it seems like a brute-force search would take forever. A naive upper bound for the number of just 8 letter character strings on a 5x5 board is 20 million or so, and the number generally goes up exponentially according to the number of letters in the string. I'm sure it's really quite a bit less, but that adds up to a lot of searches.

By the way, I'll try to stop myself, but I've been thinking about a program to determine just how many character strings (from 4-25 letters) would have to be checked on a 5x5 board.

Last edited by Spike1007 : 03-12-2018 at 01:48 PM.
Reply With Quote
  #9  
Old 03-12-2018, 02:49 PM
Asmarandana Asmarandana is offline
Junior Member
 
Join Date: Sep 2017
Posts: 2
Asmarandana is an unknown quantity at this point
Default

I see, thanks for the explanation!

I'm not a programmer, but I assume that one way to bring down the volume of strings that have to be checked is by having the program automatically eliminate all the strings that start with unviable letter combinations.

For example, there would be no need to check further on any strings that start with tb, sd, np, etc. (Though you'd have to make sure of that.) Then there would be 3-letter strings to eliminate (duj, joq, etc.) and so on. However, that sounds incredibly tedious to identify all the strings and program them. Now that I think about it, I'm not sure it's all that helpful.

Last edited by Asmarandana : 03-12-2018 at 03:00 PM.
Reply With Quote
  #10  
Old 03-13-2018, 04:07 AM
Spike1007 Spike1007 is offline
Senior Member
 
Join Date: Aug 2017
Posts: 389
Spike1007 has received several accolades
Default

You're right, that could cut it down a lot. If you find a letter string (whether 2, 3 4 letters or whatever) that doesn't appear as the start of any word in the dictionary, then you can avoid even considering any longer strings that start that way, so it cuts out a lot of dictionary searches. Even that way though, it seems like it could be slow. I'm sure that they have developed very efficient shortcuts somehow.

For the hell of it, I did write a program just to count the number of "legal" word paths on the 5x5 board (ignoring what letters might actually be there). I think it'll take longer to run than it did to write. (It's a little more than half way done, and it's taken about a half hour so far.) As a test, I looked at the number of "words" that just start in one corner. The number of 20-22 letter paths are each over a billion. Like you said though, if the first three letters or so weren't valid word parts, I could skip all those tests (and all those start with one of 15 different 3-letter combinations).

Anyway, I should stop. Other people are probably more interested in essay-writing help than this.
Reply With Quote
Reply


Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 04:03 PM.


Powered by vBulletin® Version 3.6.5
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.

About Puzzle Baron

The Puzzle Baron family of web sites has served millions and millions of puzzle enthusiasts since its inception in 2006. From cryptograms to acrostics, logic puzzles to drop quotes, patchwords to wordtwist and even sudoku, we run the gamut in word puzzles, printable puzzles and logic games.

Questions or Comments?

The word 'bought' has how many letters in it?