After the lcls steps, I found manually a simple two-step solution without forcing chains.
1. W-wing
(7=4)r5c5-r6c4=r6c8-(4=7)r4c8 => -7 r4c4, r5c7
- Code: Select all
.----------------.--------------.---------------.
| 689 69 3 | 78 27 1 | 4 5 267 |
| 4 2 7 | 3 5 6 | 9 1 8 |
| 168 16 5 | 478 247 9 | 67 3 267 |
:----------------+--------------+---------------:
| 2 149 189 | 14-7 6 3 | 178 d47 5 |
| 3 146 16 | 5 a47 8 | 16-7 2 9 |
| 7 5 168 | b14 9 2 | 168 c46 3 |
:----------------+--------------+---------------:
| 1569 7 169 | 2 18 45 | 3 89 146 |
| 16 8 2 | 9 3 47 | 5 67 1467 |
| 159 3 4 | 6 18 57 | 2 89 17 |
'----------------'--------------'---------------'
2. After 5 singles and some other lcls steps, there is an XY-Wing:
(4=6)r7c9-(6=7)r8c8-(7=4)r8c6 => -4 r7c6, r8c9; sste
- Code: Select all
.--------------.-------------.----------------.
| 89 69 3 | 78 2 1 | 4 5 67 |
| 4 2 7 | 3 5 6 | 9 1 8 |
| 18 16 5 | 78 4 9 | 67 3 2 |
:--------------+-------------+----------------:
| 2 19 189 | 14 6 3 | 178 47 5 |
| 3 4 16 | 5 7 8 | 16 2 9 |
| 7 5 168 | 14 9 2 | 168 46 3 |
:--------------+-------------+----------------:
| 56 7 19 | 2 18 5-4 | 3 89 a46 |
| 16 8 2 | 9 3 c47 | 5 b67 167-4 |
| 159 3 4 | 6 18 57 | 2 89 17 |
'--------------'-------------'----------------'
Actually, if we eliminate a digit or force a digit into a cell with a forcing chain (or a forcing net), we might need three of more chains, so is it better than a two-step solution with two chains described by standard patterns?