ronk wrote:The term "2-digit unavoidable" is used frequently in this thread. But what exactly does it mean?
It means a subset of the unavoidable sets, the unavoidables that consist only of two digits. In this thread we have only examined these smaller sets. The maximum amount found in any grid is so far 78 in the Pt-grid. As you noticed in the Megaclue-thread, there is actually a huge amount more unavoidables in any grid, most however use more than two digits.
ronk wrote:Since some of the 2-digit unavoidable sets are quite large, it can't mean only two digits appear in the set. So does "2-digit" mean each cell in the set can only be one of two values?
The maximum size of a two digit unavoidable set is 18 (2x9). The 127 FE-grids are special as all two digit unavoidables are of size 18. Essentially it means that if you pick any two digits in any of the grids, all 18 instances will only form one minimal unavoidable set.
ronk wrote:For example, consider this unavoidable... Is it a "2-digit unavoidable?"
No, it seems to be a 6-digit unavoidable.
ronk wrote:How does one know which of the unavoidable sets output by checker are 2-digit unavoidables ... or are they all 2-digiters?
I suppose the only way of knowing is to compare them to the grid. They are not all 2-digiters, and there is 2-digiters that don't get listed by checker (the largest ones). I don't know of any publically available program that counts the 2-d unavoidables, but it's not hard to program one your self. Mine works on the simple principle:
- Code: Select all
for i=1 to 8 {
for j=i+1 to 9 {
load the grid
for all cells in the grid {
if cell A = i {
set A=0
find all cells with value (j) seen by A and set their value to 0
find all cells with value (i) seen by those cells, set value to 0
find all cells with value (j) seen by the new cells, set value to 0
...until no more cells (i,j) can be seen by any of the cells
increase unavoidable counter
}
}
}
}
I figured that would be the easiest way to do it without having to code any solver or pattern recognizer. I've been trying to come up with a way to find 3-digit and larger unavoidables in a similar way, but it's just not that simple...
RW