wu :: forums
        (http://www.ocf.berkeley.edu/~wwu/cgi-bin/yabb/YaBB.cgi)
      

        riddles >> general problem-solving / chatting / whatever >> A non-Facebook version of my Numbrosia puzzle
        
(Message started by: amichail on Jan 1st, 2008, 11:21am)
Title: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 1st, 2008, 11:21am
It's here:

http://numbrosia.com

No registration required.  But do register if you don't want to lose your progress.

Any feedback would be appreciated.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 1st, 2008, 11:25am
Level one is rather trivial, and for some reason it won't let me continue to the next level..
Or do you even need to register if you want any progress?
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 1st, 2008, 11:42am

on 01/01/08 at 11:25:00, towr wrote:
Level one is rather trivial, and for some reason it won't let me continue to the next level..
Or do you even need to register if you want any progress?

Try clearing the numbrosia.com cookies and try again. It should work now.

BTW, if you register while you still have the cookie under which you played, you won't lose progress.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 1st, 2008, 1:18pm
Not bad.
Of course it helps I'm top of the board at the moment :P
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 1st, 2008, 1:32pm

on 01/01/08 at 13:18:35, towr wrote:
Not bad.
Of course it helps I'm top of the board at the moment :P

You have done quite a few levels.  If you don't want to lose that progress, register while you still have the cookie.

BTW, it's funny that I'm not particularly good at my own puzzle...
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by Grimbal on Jan 1st, 2008, 2:40pm
Thanks for listening for my complaint and implement it as a non-facebook application.

So I had a look.  I did up to level 20.

I have to say it is a quite difficult puzzle.  It is not difficult to solve, but if you want to optimize the number of moves, you really have to think ahead and find good strategies.  I think it is hard to find the optimal solution by computer.

Maybe it lacks a bit a variety.  There is nothing new to look forward to.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 1st, 2008, 3:00pm
I must say, it's a very good time waster. As these kinds of puzzles/games go, it does quite well.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 1st, 2008, 3:47pm

on 01/01/08 at 14:40:28, Grimbal wrote:
Thanks for listening for my complaint and implement it as a non-facebook application.

So I had a look.  I did up to level 20.

I have to say it is a quite difficult puzzle.  It is not difficult to solve, but if you want to optimize the number of moves, you really have to think ahead and find good strategies.  I think it is hard to find the optimal solution by computer.

Maybe it lacks a bit a variety.  There is nothing new to look forward to.

You might check back to see how your score changes over time (as the median # moves for most levels changes over time -- at least for a while).  

You can also redo levels to try to do better.

BTW, what is there new to look forward to in a popular puzzle such as sudoku?
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by Grimbal on Jan 2nd, 2008, 5:41am

on 01/01/08 at 15:47:27, amichail wrote:
BTW, what is there new to look forward to in a popular puzzle such as sudoku?

Yeah, I have the same problem with sudoku.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by SMQ on Jan 2nd, 2008, 6:16am
Registered 3 days ago and already blocked by my office firewall as a "game playing" site -- son of a :o ... I'll check it out when I get home.

--SMQ
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by mikedagr8 on Jan 2nd, 2008, 6:30am
I do not wish to lose my progress. The site is great, and much better than the Facebook version I believe.

Thanks for the fun. :D ;D


on 01/01/08 at 13:18:35, towr wrote:
Of course it helps I'm top of the board at the moment :P


I currently am ranked #1 on each level I have completed, except level 10, but it is 1:40am. It will be redeemed.  ;D
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 2nd, 2008, 10:29am
It'd be nice, though, if your best score for a puzzle counted, and not the last one..
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 3rd, 2008, 9:28am

on 01/02/08 at 10:29:39, towr wrote:
It'd be nice, though, if your best score for a puzzle counted, and not the last one..

This feature will be coming soon.

BTW, could you tell me what happened on level 91?  How did you end up doing 1000 moves (most of them being rotating column 5 up)?  Did you notice anything strange happening?
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 3rd, 2008, 9:32am

on 01/03/08 at 09:28:57, amichail wrote:
BTW, could you tell me what happened on level 91?  How did you end up doing 1000 moves (most of them being rotating column 5 up)?  Did you notice anything strange happening?
I selected rotate and kept enter pressed :P
It's not a defect in your game, but rather in the player ;) It's really boring to be the first to do a level.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 3rd, 2008, 9:34am

on 01/03/08 at 09:32:29, towr wrote:
I selected rotate and kept enter pressed :P
It's not a defect in your game, but rather in the player ;) It's really boring to be the first to do a level.

It's a good thing I use the median rather than the average for scoring!
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 3rd, 2008, 9:53am

on 01/03/08 at 09:32:29, towr wrote:
I selected rotate and kept enter pressed :P
It's not a defect in your game, but rather in the player ;) It's really boring to be the first to do a level.

So this is an easy way to cheat.  Create multiple accounts and do very badly in all but one in high levels.

I wonder how I can detect that.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 3rd, 2008, 10:34am

on 01/03/08 at 09:34:48, amichail wrote:
It's a good thing I use the median rather than the average for scoring!
Yes, indeed. I found that out when I tried an to get 1000 on the first puzzle with an anonimous; I quickly figured this would take entirely too much effort to skew the results (never mind that everyone would benefit equally much from it).


on 01/03/08 at 09:53:46, amichail wrote:
So this is an easy way to cheat.  Create multiple accounts and do very badly in all but one in high levels.

I wonder how I can detect that.
Well, I don't think you need to worry about it much; not unless someone programs a bot to play the game. It takes too much effort otherwise to get enough results to change the median. Except for puzzles where there's only a few players, but then it takes too much effort to get that high.
As for detecting it, anything over 50 moves is already a bit dodgy, and over several hundreds certainly.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 3rd, 2008, 10:49am
You might consider also giving the average score per level on the leaderboard, that way people can more easily compare their results with others even if they don't do a lot of levels.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 3rd, 2008, 11:43am

on 01/03/08 at 10:34:47, towr wrote:
Well, I don't think you need to worry about it much; not unless someone programs a bot to play the game. It takes too much effort otherwise to get enough results to change the median. Except for puzzles where there's only a few players, but then it takes too much effort to get that high.
As for detecting it, anything over 50 moves is already a bit dodgy, and over several hundreds certainly.

One possibility is to precompute the min # of moves possible for a level and use a scoring system based on how far away you are from that min.

For example, if you take k moves over the min, you might get 1000/2^k points for that level.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by Astrix on Jan 3rd, 2008, 6:15pm
So, have you figured out a way to figure out the minimum number of steps?
The next best thing is to create the puzzles manually by running the game in reverse, and counting the steps you took. There might actually be a quicker way to get there, but, unless your creation path is very inefficient, it should be close enough to establish an appropriate difficulty rating for that level.
Personally, I think the program should not only reward you for beating the median, it should penalize you for doing worse than the median. Otherwise, on any level where the median is also the optimum, a player can take a thousand steps and still score the same as the best player.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 3rd, 2008, 8:16pm

on 01/03/08 at 18:15:16, Astrix wrote:
So, have you figured out a way to figure out the minimum number of steps?
The next best thing is to create the puzzles manually by running the game in reverse, and counting the steps you took. There might actually be a quicker way to get there, but, unless your creation path is very inefficient, it should be close enough to establish an appropriate difficulty rating for that level.
Personally, I think the program should not only reward you for beating the median, it should penalize you for doing worse than the median. Otherwise, on any level where the median is also the optimum, a player can take a thousand steps and still score the same as the best player.

I don't have a good admissible heuristic yet.   This is starting to look like a major research problem.

The problem with having a penalty is that poor players will see no reason to play.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 4th, 2008, 1:40am

on 01/03/08 at 20:16:15, amichail wrote:
The problem with having a penalty is that poor players will see no reason to play.
If they just get 0 point for scoring below the median, or perhaps even in the lowest 25%, it wouldn't be too bad.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by Grimbal on Jan 4th, 2008, 2:14am
There should be a positive score for any level you have completed reasonably well, even if everybody succeeded in it.

What about a linear score of 100 if you have reached the minimum so far, down to 0 if you did the median or (min moves+5), whichever is larger.   Any higher number of moves gives 0.

If someone is the first to do a level, he would get easy 100 points, which is an incentive for other to go and beat his score.

The problem:  easy level are rewarded too much.

Or there could be a fixed 10 points for reaching the median, plus a bonus if you do better.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by mikedagr8 on Jan 4th, 2008, 2:33am
[quote author=Grimbal link=board=riddles_general;num=1199215270;start=0#22 date=01/04/08 at 02:14:48]If someone is the first to do a level, he would get easy 100 points, which is an incentive for other to go and beat his score.
quote]
If that were the case, towr would have a higher score. It wouldn't be fair to those who are starting. Possibly If you have had a best score on a level you recieve 50, and you go down from there. But I think you should be reqarded for having the minimum number of moves on a level even if someone beats it.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by Grimbal on Jan 4th, 2008, 3:06am
There are long games and short games.
And there are easy games and difficult games.

For easy games, the difference between minimum and median would be small, for hard games, it would be large.
For short games, the difference between minimum and median would also be small, partly because they are easier, partly because there is less choice.
It is easier to improve the score on long games than in short games.

I think it is reasonable to say that the median should have the same score for all games.  It should be positive.  The question is how to reward better than average scores.  Since the difference between best and median is higher with difficult games, it is reasonable to take into account the difference between moves and median.  But since it is easier to improve the score on a long game, it should be divided by the minimum number of moves.

So I'd propose something like:
 n: moves, b: best result, m: median
 points = A + B·((n-m)/b)2
   or 0 if n>m.
where A and B are constants to define.

If half of the people reach the best score, it would be A with no bonus.  Either there are not enough results to give a bonus or the level is too easy.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 4th, 2008, 3:22am
Maybe people should be able to submit their own scoring method, that way they can select whichever method that makes them look most favorable ;)
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by SMQ on Jan 4th, 2008, 5:12am
In that case I propose awarding the most points to those who don't play at all (darn corporate firewall policy). ;)

--SMQ
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by Grimbal on Jan 4th, 2008, 6:12am
That's easy!  It much more harder not to play when you don't have a firewall.  :)
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 7th, 2008, 3:02pm
One further suggestion for the game; stop using 'alert' when there is a communication failure. It is annoying. Just use a text field or something.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 7th, 2008, 4:11pm

on 01/07/08 at 15:02:45, towr wrote:
One further suggestion for the game; stop using 'alert' when there is a communication failure. It is annoying. Just use a text field or something.

How frequently does this error occur?  Does clicking OK generally work or do you often get the error again?
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 8th, 2008, 12:06am

on 01/07/08 at 16:11:50, amichail wrote:
How frequently does this error occur?  Does clicking OK generally work or do you often get the error again?
Usually it doesn't happen frequently, but in the five minutes before I posted that, I had a whole series. Within 5 seconds I clicked ok I'd get the error again, and so on a dozen times.
I suppose if it's part of the function library you used, it may be hard to change how the error is shown, but it is really bothersome when you're just working on a puzzle (which works perfectly well without communicating with the server; your progress just won't be saved).
As an added annoyance you can't do anything with the browser window, not even in other tabs, until all those errors are clicked away.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 9th, 2008, 3:45pm

on 01/08/08 at 00:06:23, towr wrote:
Usually it doesn't happen frequently, but in the five minutes before I posted that, I had a whole series. Within 5 seconds I clicked ok I'd get the error again, and so on a dozen times.
I suppose if it's part of the function library you used, it may be hard to change how the error is shown, but it is really bothersome when you're just working on a puzzle (which works perfectly well without communicating with the server; your progress just won't be saved).
As an added annoyance you can't do anything with the browser window, not even in other tabs, until all those errors are clicked away.

I took the alerts out.

BTW, one can have this puzzle in reverse, starting with zeros and aiming for a particular pattern, perhaps even a picture.  This seems more difficult to solve (if you don't cheat by going backwards).
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 10th, 2008, 12:11am

on 01/09/08 at 15:45:19, amichail wrote:
I took the alerts out.
I noticed, it's much better now, thanks.


Quote:
BTW, one can have this puzzle in reverse, starting with zeros and aiming for a particular pattern, perhaps even a picture.  This seems more difficult to solve (if you don't cheat by going backwards).
You can solve the inverse; doing those moves (in the same forward order) on an empty field will generate the original.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by Grimbal on Jan 10th, 2008, 1:16am

on 01/10/08 at 00:11:09, towr wrote:
You can solve the inverse; doing those moves (in the same forward order) on an empty field will generate the original.

Not sure about that.  You get the right numbers (provided you switch + and -) but in the wrong places.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 14th, 2008, 7:03am
The pages are becoming very cluttered with all those puzzle links on, apparently, every page. It would be bad enough if it were only on the main page.
Take a hint from google, keep things simple and uncluttered. Throwing more and more in doesn't improve it; quite the contrary.

(And I can't say I'm fond of the yellow box above the puzzle field either; did people fail to read instruction before? Not to mention it is fairly straightforward to find them out while playing.)
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 14th, 2008, 1:22pm

on 01/14/08 at 07:03:29, towr wrote:
The pages are becoming very cluttered with all those puzzle links on, apparently, every page. It would be bad enough if it were only on the main page.
Take a hint from google, keep things simple and uncluttered. Throwing more and more in doesn't improve it; quite the contrary.

(And I can't say I'm fond of the yellow box above the puzzle field either; did people fail to read instruction before? Not to mention it is fairly straightforward to find them out while playing.)

Those links were put there as a way to increase the popularity of Numbrosia (in at least 3 different ways).
If this doesn't work out, I can take them out.

As for the yellow box, it goes away when you log in.

The puzzle links are less distracting now btw.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 29th, 2008, 6:07am
OMG
How tacky it has become.. Seriously, it looks like a myspace page; and that is not a compliment.

Definitely not going to playing that anymore..
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 29th, 2008, 7:41am

on 01/29/08 at 06:07:47, towr wrote:
OMG
How tacky it has become.. Seriously, it looks like a myspace page; and that is not a compliment.

Definitely not going to playing that anymore..

NASA images are free.  I can use hundreds of them to motivate people to play more to see what's coming up.

But if there's a specific problem with readability, etc., then I could look into it.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by towr on Jan 29th, 2008, 9:10am

on 01/29/08 at 07:41:37, amichail wrote:
NASA images are free.  I can use hundreds of them to motivate people to play more to see what's coming up.
Perhaps you could try to make a poll and check whether there is actually anybody motivated to play on to see "what's coming up"; and how many people are now motivated not to play (well, ok, it'll be hard to poll them now..)


Quote:
But if there's a specific problem with readability, etc., then I could look into it.
Well, there's grey text that's not always very readable given the background; but the overall horrific ugliness is more what gets me..
It used to be fairly elegant; a google-like simplicity. Now it is smeared over with every kind of web-makeup known to man, and it looks like a grotesque clown. It doesn't need background images, it doesn't need transparent fields, it doesn't need colours fading over into each other, etc. It's a puzzle; not javascript show-and-tell. Throwing in every 'feature' you can think of does not improve it.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 29th, 2008, 12:36pm
I'll add an option to remove the background images.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Jan 30th, 2008, 8:29am
I got rid of the eye candy, except for the fade effect.
Title: Re: A non-Facebook version of my Numbrosia puzzle
Post by amichail on Feb 12th, 2008, 8:29am
What do you think of having different shapes?

For example:

X
X X
X X X
X X
X

In this way, moves may be done on rows/columns of different size.

Would this add an interesting new dimension to the game play?


Powered by YaBB 1 Gold - SP 1.4!
Forum software copyright © 2000-2004 Yet another Bulletin Board