OpenEuphoria

Re: contest #2

• Posted by petelomax at blueyonder.co.uk Mar 17, 2002
• 339 views

I'm a bit disappointed with my results (see below), but I'll make two
observations just to cheer myself up:

1) The worst cases (*'d) are where I knew they would be. Depending on
how Rob generates the random test data, they will either be quite rare
or extremely common. If he picks random words from the dictionary, and
makes a patterns from them, I might be OK. If he chooses to do:
	pattern=repeat(0,rand(10))
	for i = 1 to length(pattern)
		pattern[i]=rand(26)+(rand(2)-1)*64
	end for
I'm spannered, (& I wouldn't consider that a fair test).

2) I doubt Rob's test set will be < 50% with fixed letters and > 50%
without, but if it is, maybe I'll not do quite so bad.
(for a seven letter word, the above will generate 127 words with fixed
letters for every one with none, I think)

Maybe a bit late to realise this, but I guess I subconsciously assumed
there would be a bias towards patterns with no fixed letters, given I
realised when using such a routine for problem #3, no fixed letters
would ever be passed, unless you use the routine again rather than
filter/test the list it already gave you.

Any comments Rob? Does the 90% rule for problem 3 apply to the
patterns for problem 2 as well, or not?

Oh well, we'll see. Down to luck again I see, C.K.

Pete

Here are my (disappointing) results:
constant testset= {
--# pattern			matches	tone		chris	pete		
{1,2,3,6,4,7,3,10,4,8,7,'S'},	1,	1.58,		0.28,		0.57,		
{'A','P',1,2,3},				6,	3.22,		1.43,		51.45,	*
{'A', 'P', 'P', 1, 2},			2,	0.35,		0.39,		2.3,
{'E',7,8,9,7,9},				0,	0.96,		0.16,		0.54,
{1,2,'C','K',1},				3,	1.8,		3.9,		3.12,
{4,'I','E'},					7,	3.64,		0.66,		10.95,	*
"RABBIT",				1,	2.48,		0.11,		8.52,
{'T',1,2},					40, 	1.12,		0.44,		10.49,	*
{1,2,3,'S'},				283,	10.75,	2.63,		35.76,	*
{1,2,3,4,'E','D'},			331,	55.39,	5.2,		66.66,	*
}
--					totals  	 81.29,  	15.2,		190.36!!!