... by: K9Dr

I am amazed by and in awe of the people who can solve these puzzles in their "heads".
My solver first eliminates all matches in rows, columns and boxes. It then solves any single occurrence of a number in a row, column or box and then repeats match elimination. If the puzzle is easy, this is the only routine necessary for solution. If the puzzle is hard i.e. unsolvable, the program will randomly select (Trial) one of the 2 numbers of a naked pair, if available. Rarely, there are no naked pairs available, so the program will use a naked triple. It prefers matching naked pairs in a row, column or box, if available. It then goes forward to eliminate all matches in a row, column or box and solves any single occurrence, if available. If no more eliminations or square solutions are available, it will randomly select another naked pair and randomly select one of the numbers of this naked pair. If during elimination, there are duplications in 2 squares of a row, column or box (Error) or all of the numbers in a square get eliminated (Error), the program goes back to the last naked pair Trial and uses the other number of the naked pair and then goes forward as above. If this Trial has already been reversed, the program goes back to the preceding Trial. This program could be improved by using the statistical likelihood of a Trial, but I am not smart enough to code that ... yet. Even when using statistical likelihood, the solution is still Trial and Error. My program will also graphically display and explain (with words, not talking yet) each step of its solution process. Program written in VisualBasic6. I am old school

... by: Dieter

@K9Dr:
We are interested in strategies.
It's not so important to know only the time spent on solving the puzzle.

... by: K9Dr

My VB6 solver (self-coded) took 1.467 seconds, 414 steps, and 60 Trials.

... by: Frans Goosens

#718

With trial and error

Combination C2=268 and H9=468

************************************************************
C2=2 H9=4 No solution, Fixed

Combination 2-digits cells

C9=6 Wrong, Undo calculation
C9=9 Wrong, Undo calculation

All reset to initial position

************************************************************
C2=2 H9=6 Wrong, Undo calculation

All reset to initial position

************************************************************
C2=2 H9=8 Wrong, Undo calculation

All reset to initial position

************************************************************
C2=6 H9=4 Wrong, Undo calculation

All reset to initial position

************************************************************
C2=6 H9=6 Wrong, Undo calculation

All reset to initial position

************************************************************
C2=6 H9=8 Wrong, Undo calculation

All reset to initial position

************************************************************
C2=8 H9=4 Wrong, Undo calculation

All reset to initial position

************************************************************
C2=8 H9=6 No solution, Fixed

Combination 2-digits cells

A1=3 Wrong, Undo calculation
A1=7 Wrong, Undo calculation

All reset to initial position

************************************************************
C2=8 H9=8 ( I9= ) Solved,

Total solving time is : 304 sec.
Number of logical steps is : 38559

... by: Yves

There are 879 two cells trivial solutions in this #718

one is J8=9 & F9=6

... by: Dieter

#718 - initially populated cells: 21 = 25,9 %
No solver used, solver = brain.
basics: %

Strategy: numbers 1,3,5,7 occur most frequently, combine two cells with almost same candidates.
combinations of D8 = (1,3,5,6) and D9 = (3,5,6,7,8);
after basic checks 4 matrices left: 3/5, 3/7, 5/7 and 6/5.
D8=5 & D9=7 solve with J1=5.

Optimization:
D8=5 & J1=5 or D9=7 & J1=5 solve.

... by: Serban

# 718 Basics to 21

1. G4 = A6 = 5 solve

2. E7=9 & F7=8 solve

3. F6=9 & F7=8 solve

... by: Norman Powell

This puzzle is solved on the fourth pass of Forcing Networks, which finds the solution when checking to a possibility..
AS's Solver Opens in the same way and finds 6 eliminations through Forcing Networks before running out of strategies, indicated [AS# order of discovery below]. My implementation reports eliminations as they are found. AS reports in order of increasing length of post-hoc inference chain that would reveal elimination.
Here are the details:
Hidden Pair: Column 9 for Values 5, 7 in Cells B9 & D9, removing other possibilities from those cells. [As AS]
Confined to Box: Value 3 is confined to Column 8 and Box 6, removing other 6s from Column 8.
Forcing Network First Pass: Eliminate the follow since they produce contradictions when ON:
2D1[AS#3], 1D8, 8E5[AS#5], 1F1, 6F6, 8F6 [AS#1], 9F7[AS#6], 4G2[AS#4], 8J4[AS#2].
Forcing Networks Second Pass: Eliminate the follow since they produce contradictions when ON:
2C1, 6D1, 6D3, 6D8, 2J1, 4J9.
Confined to Box: Value 6 in Row 4 and Box 5.
Forcing Networks Third Pass: Eliminate the follow since they produce contradictions when ON:
8A6, 6C1, 4C4, 2D3, 7D4, 8D4, 8D6, 9E8, 2F2, 2F6, 2G2
Hidden Single: in Row F, F1=2.
Confined to Box: Value 8 is confined to Column 6 and Box 8.
3D Medusa: Rule 5 1C5 can see both colours, eliminate.
Forcing Networks Fourth Pass: Eliminate the follow since they produce contradictions when ON:
7A1, 9A3
However the solution is found when testing 88A4 for being ON.
Solver VBA in Excel [inspired by AS's Solver and Sudoku Wiki], with lots of graphical cues, so very slow:
2:57 [min:ss]
Forcing Networks gradually moving up the methods for efficiency, eclipsing X-Cycles, AIC, Grouped X-Cycles and Grouped AIC. Almost Locked Sets after Forcing Networks and is still triggered by some examples that get past Forcing Networks.