This post introduces the expansion of coloring clusters by coloring groups by grouped strong links, such as the those between value groups in ALS. This advance is especially significant as we address whether or not to exhaustively construct ALS and ALS_XZ in Sysudoku.
On the grid following a first phase AIC building without ALS, I superimposed an ALS with a value group aligned with 2r3c4. If only I could find a wink from the long AIC from 2r2c6 into the ALS value groups 4, 5, 6 or 8, the almost nice loop removing 2r3c4 would be closed
I thought I had one, but discovered three weekly posts later, I had not.
That concentrated my attention, and after adding a blue/green cluster to the flock of bv, I found the red/orange cluster connecting ALS value groups 2 and 8 with the naked pair N27. It even delivered an extra trap outside the ALS, and tied in another ALS from my later ALS building.
The red side offered the result I wanted, and more. I tried the orange side, hoping for an observable contradiction. Instead, I got an immediate collapse, with with a blue wrap and large numbers of naked pairs along the way, and a reversible rectangle near the end. It reveals a puzzle of almost two solutions, or almost three, one hard and two easy.
Next week we introduce the exhaustive generation of an ALS map, with a ©PowerPoint template to aid the process. Exhaustive ALS building takes time, but can be organized and kept in bounds. The task fits in systematic order of battle against very difficult puzzles.
Systematic Sudoku 