dobrichev wrote:Thinking about earlier dead-end branch identification, the UA of size 4 and 6 are hit with the first few clues and are out of interest.(Is this true for an average grid or the number of exceptions is ~100% ?)

Hi mladen,

For sure, early detection of a dead branch is a key point (the key point ??)

Mc Garry made some negative comments on another process using late building of higher degree UAs. I am not quite sure that this is definitely true.

Regarding your point, let me answer in the following way.

The table of blue shows an average 27 UAs of size <= 6.

Even if the maximum number of disjoint uas in that lot is weak (says 4 to 5 for example) each branch hit only a small number of them at the start.

Usually after ten clues taken in the smallest uas, you still have available 4_6

here is an example out of the "test solution grid above".

We are on the path leading to the 17 seen, so this is a king of worst case for your "conjecture".

With 10 clues, your conjecture is nearly verified, so I take the situation with 8 clues of the expected 17.

the last item shows the next clues in the path leading to the 17

we start in bands 1+2 with 36 uas of size <=6.

Here is the status

- Code: Select all
`......7.........2....1.....2..........79......3......4`

..1.....1.....1..1..1..1.............................. i=11

....1..1..1..1.....1.....1............................ i=13

.....1..1..1..1.....1.....1........................... i=14

.1..1..1...........1..1..1............................ i=23

.1..1.....1..1.....1..1............................... i=26

...........1..1..1..1..1..1........................... i=30

.....1..1.....1..1.....1..1........................... i=31

..1..1.....1..1.....1..1.............................. i=32

..1..1..1...........1..1..1........................... i=33

..1..1..1..1..1..1.................................... i=34

..1.....1..1.....1..1.....1........................... i=35

......7.........23.8.1.....2..........79......3......4 next clues