Friday, June 7, 2013

Collecting the data

Here's a chart which we might use as we collect our data.


A
B+
B
C+
C
D+
D
E+
E
I
100%
100%
100%
100%
100%
100%
100%
100%
100%
II
100%








III
100%








IV
100%








V
100%








VI
100%








VII
100%








VIII
100%








IX
100%







0%




While the presence of blockades is dependent on how crowded a square is, the actual usefulness of blockades is dependent on which array we're working with, so we'll put the (+) symbol to indicate a blockade with the rank for data points rather than with the rank for squares.



It's pretty clear that a space with only one possibility left will be solved every time, and that an intersection which has four data points influencing it will give us a hit every time. It's also clear that an empty square with no data influencing it won't give us anything to work with at all.

What we need to find out is where the transition is -- at what point are we looking at diminishing returns. And we need to think about whether we'll count finding pairs equally with finding deductions, or if we want to split that information out. 



A
B+
B
C+
C
D+
D
E+
E
I
100%
100%
100%
100%
100%
100%
100%
100%
100%
II
100%
100%
100%
100%
100%
100%
100%
100%
100%
III
100%
100%
100%






IV
100%
100%
100%






V
100%
100%
100%






VI
100%
100%
100%






VII
100%
100%
100%






VIII
100%
100%
100%






IX
100%
100%
100%





0%


For example, a square with 2 spaces left will always have at least two pairs, and a square influenced by three data points (B) will always find at least one pair, whether or not there's a blockade involved.  If we count pairs equal to hits our chart would look like this.  Splitting the information would require two charts, or extra lines.


No comments:

Post a Comment