sirdave wrote:My main problem is that 'Trying' equals guessing. And the guessing not only involves the choice of which number to plug in, it also applies, for the most part, to the choice of cells to start in. This isn't a judgment on those that want to do it, rather it is making clear to the uninformed what the difference is between the above method and using the methods that go beyond 'locked': fishy strategies, ALC, AICs, nice-loops etc.
I could see how _m_k picked his first cell to guess... After naked quads, swordfish etc you arrive the following state:
- Code: Select all
*--------------------------------------------------------------------*
| 1 479 45789 | 35789 569 5789 | 3489 369 2 |
| 278 3 789 | 26789 4 12789 | 189 5 689 |
| 2458 249 6 | 23589 1259 2589 | 7 139 3489 |
|----------------------+----------------------+----------------------|
| 2457 24679 4579 | 1 2589 3 | 2589 2679 5789 |
| 2356 8 1359 | 259 7 259 | 12359 4 3569 |
| 2357 1279 3579 | 4 2589 6 | 23589 12379 35789 |
|----------------------+----------------------+----------------------|
| 38 147 2 | 5789 159 45789 | 6 379 34579 |
| 467 5 147 | 2679 3 12479 | 249 8 479 |
| 9 467 38 | 2578 256 24578 | 2345 237 1 |
*--------------------------------------------------------------------*
There are only 2 cells with 2 candidates left: r7c1+r9c3={38}
So naturally, he plugged in both 3 & 8 in r7c1, and then in turn he tried out other cells with 2 candidates left to guess his way to the solution... So the choice of cells isn't the issue here... The issue is how he conveniently ignored the fact that the values he didn't pick for (such as 3 in r7c1) didn't lead to obvious contradiction, so it's not a logical approach at all to proceed to the solution...
And it's totally off the point to perform 4 levels of recursion, when this puzzle, like all others discovered so far, needs only 2 levels of recursion with singles to get to the solution (e.g. r19c2=[47])... But even this is off the point, as the special thing about this puzzle is that no 1-level recursion, even using singles, locked sets and naked pairs, can solve this puzzle...