Solving Techniques

Techniques

Sudoku Helper generates puzzles based on human solvable techniques. The techniques shown here are the ones used, except for guessing. Guessing is only included here in case of user error since the Hint functionality will try to guess before giving up.

For all of the examples below, when you see numbers in () like (1,9) it's referring to row 1, column 9.

Single Naked Candidate

The Single Naked Candidate technique is the simplest technique used for solving sudoku.

Definition

A possibility can be selected as the solution to a box if it is the only possibility in the cell.

Example 1

31286459.598731426746529831651973284279458163483612759865347912927186345.3429567.

This sudoku shows three ways that you can come across a single naked candidate, by row, by column and by house. A single naked candidate occurs when at least one of every other possibility has been placed in the box's row, column or house. In this case, we know that the box at (1,9) must have a 7 as the only possibility, since the other values in row 1 have been used. The same is true for the box at (9,1) where all of the values except for 1 have been used in column 1. The 8 in box (9,9) can be determined by all of the other values in house 9 being used.

Example 2

9..3..2....6985.3.......4.....2...8..1..4......3...9...5.8...7.12........3......1

Single naked candidates can arise much sooner in puzzles than just the end as shown in example 1. As you can see, the box at (2,9) shows that 7 is the only possibility. You can come to this conclusion by first looking at column 9. Since there is already a 1 in the solution, box (2,9) cannot have a 1. Looking at house 3, where box (2,9) resides already contains a 2, 3 and 4, so box (2,9) cannot be 2, 3 or 4. Finally, looking at row 2, the numbers 5, 6, 8 and 9 are already part of the solution and in the same row as box (2,9). Therefore, we can safely conclude that 7 is the solution for box (2,9).

Single Candidate

Definition

A possibility can be selected as the solution to a box if it is only found in one box in a given row, column or house.

Example 1

.6.51......9.....4..4..8.7...........536..1....7..1.45....6...7...18.3....5..981.

As possibilities are removed, it often happens that only one box in a given row, column or house can be the acceptable solution for that box. In this example, we can see two such cases highlighted by the possibility highlighting feature. If we focus on the the box at (1,6) we can see that 4 has to be the solution for this box since it it the only 4 left in this row.

But how did we get to 4 being the only possibility? As you can see, the puzzle started out with 4 as a solution to box (2,9) and box (3,3). This means that the 4 slot has been occupied for rows 2 and 3. This leaves us looking at house 2. Since house 1 and house 3 took up the other 2 rows, we know that the 4 in house 2 must be in row 1. Boxes (1,4) and (1,5) already have solutions to them. Therefore, the only place left for a 4 to be in house 2 is at box (1,6).

It is not always the case that a solution number will be the cause of a single candidate. Utilizing other techniques that eliminate possibilies may also leave a single candidate in a row, column or house.

Single Sector Candidate

The Single Sector Candidate technique is the first technique used to remove possibilities instead of selecting a solution.

Definition

  1. If there is only one house in a row or column that contains a given possibility, the remainder of the house may be cleared.
  2. If a possiblily in a house is only in one row or column, then the possibilites can be removed from other houses sharing that row or column.

The easiest way to catch onto this pattern is to look at some diagrams (these are some of the possibilities, not all):

     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
     
The four diagrams above all show both potential cases. Assume the pink boxes have at least 2 of them with a possibility. If the blue boxes do not contain the possibility in question, then the possibity in the boxes that are colored green can be safely be removed. The inverse is true as well. If the green boxes do not contain the possibility, then the possibility can be removed from the blue boxes.

As you can see from the two diagrams above, the order of the rows, columns and houses is irrelevant.

Example 1

..257.1.....19.2781.72..9...7.9615325613278..239845617.2......1..671.32..1..52.8.

This example shows how rule #1 of the single sector candidate technique can be used to make eliminations. Looking at 4 as a possibility, we can see that by looking at column 7, only house 9 contains spots where a 4 can be, (7,7) and (9,7). Since we know that a 4 needs to be placed in one of those two boxes, that the 4s at (7,8), (8,9) and (9,9) can be eliminated. This is an example of eliminating possibilities out of green boxes because no possibility lies in the blue boxes.

Example 2

....7..3.5.8...7...7..5...6.6274..5...3....17...9.....2..468.7..8....2.....237..4

In this example, we can see that house 8 does not have a 1 set yet. However, the only possible places for the 1 to fall are in row 8 at (8,4), (8,5) or (8.6). By using the second rule of the single sector candidate technique we can remove the 1s at (8,1), (8,3) and (8,9) since a row can only have a number once, and we know it must be somewhere in the row in house 8. This is an example of eliminating possibilities out of blue boxes because no possibility lies in the green boxes.

X-Wing

Definition

If two rows have only two boxes with a possibility in the same columns, then all other boxes in those columns that contain that possiblity can have that possibility removed from them. The same is true if you exchange rows and columns.

Example 1

8...1.......5......1.34..29481...9...7.49..31953...2.416827439539.8...42........8

This example of the X-Wing technique is unique since there are actually 2 distinct X-Wings that will remove the same possibilities, one by row and one by column.

Starting with the row example, you can see that rows 3 and 5 only contain the possibility 8 in columns 6 and 7. Since there are no other places in these rows that an 8 can appear, we know that each one must contain one or the other. Therefore we can remove all other possibilities from columns 6 and 7. In this case boxes (2,6), (2,7) and (6,6) all contain an 8 as a possibility in columns 6 or 7 and can be removed.

The column example works equally well to remove 8 as a possiblity in boxes (2,6), (2,7) and (6,6). Columns 5 and 8 contain the possiblity 8 only in rows 2 and 6. Therefore we can remove 8 as a possibility from any boxes in rows 2 or 6.

Swordfish

Definition

Swordfish is an extention of the X-Wing technique to use three rows or colums instead of two. So, if three rows have a possibility in only three distinct columns between them, then all other boxes in these columns that are not part of the three rows can have the possibility removed. The inverse of rows and columns works as well. The rows (or columns) in question do not need to have all three columns have the possibility, they can have only two.

Example 1

..4357...752896143368124975..5.13..44...8..3..3..4.7...239714...41238..7...4653..

This swordfish example uses rows 1, 4 and 8 to make it's deductions. Row 1 has 9 as a possibility in only columns 1 and 2. Row 4 has 9 as a possibility in only columns 1, 2 and 8. Row 8 only has 9 as a possibility in columns 1 and 8. As you can see, this means that 3 rows have their only possibilities in 3 columns (1, 2 and 8). Therefore, we know that eventually these three rows will have 9 as a possibility in all three of these columns. This means that we can remove 9 from column 1 at (6,1) and (9,1), from column 2 at (5,2) and (9,2), and from column 8 at (6,8) and (9,8).

Naked Pair

Definition

If two boxes in a row, column or house only contain the same two possibilities, then these two possibilities may be removed from the other boxes in that row, column or house.

Example 1

732169548159482673.4.57329137.2..819.21.38754.847..3262.3....8....3..46.46.8...3.

If we look at row 7 in the above example, we can see that boxes (7,2) and (7,7) both contain only the possibilities 1 and 9. Since putting a 1 in box (7,2) will force a 9 in box (7,7) and putting a 9 in box (7,2) will force a 1 at box (7,7) every other instance of 1 or 9 can be removed from row 7. A 9 is removed from (7,4), 1 and 9 are both removed from (7,5) and a 9 is removed from (7,6).

Hidden Pair

Definition

If only two boxes in a row, column or house contain the same two possibilities, then these two possibilities are the only possibilities for the box and any other possibilities may be removed from the box.

Example 1

52.6.7..8.6...2...4......3..3.....2...1......28..6.7.9..6...4758574...1..42.7....

If we look at row 3 in the above example, we can see that the possibilities 2 and 6 each exist in only 2 boxes in the row, and they only exist in the same boxes. Since we know that row 3 still needs both of these possibilities and there are only 2 boxes which they can go into, all of the other possibilities in these boxes can be removed. In this case, 1, 5 and 9 can be removed vrom box (3,7) and 1 and 7 can be removed from (3,9).

The same logic could have been used for house 3 instead of row 3 as well to obtain the same results.

Naked Subset

Definition

If any row, column or house contains X possibilities in X boxes without any other possibilities in those boxes, then these possibilities can be removed from the rest of the boxes in the row, column or house. X will always be either 3 or 4. It is important to note that not all the boxes need to contain all of the possibilities for this technique to work.

Example 1

....843..43...796818.9.3745....3.8..3.97.8.5..68.4.1....1....8..74...2....3....9.

Row 4 contains our first example of a naked subset.If we look at boxes (4,1), (4,3) and (4,8) we can see that these three boxes contain only the possibilites 2, 5 and 7. Box (4,1) contains all three. Box (4,3) only has 5 and 7. Box (4,8) only has 7 and 2. You can see that choosing any of the possibilities in Box (4,1) will allow you to use simple techniques to solve boxes (4,3) and (4,8). Since we know that these three boxes are the only possibility for 2,5 and 7 they can be eliminated from boxes (4,4), (4,6) and (4,9)

Example 2

....843..43...796818.9.3745....3.8..3.97.8.5..68.4.1....1....8..74...2....3....9.

52.6.7..8.6...2...4......3..3.....2...1......28..6.7.9..6...4758574...1..42.7....

Expanding on Example 1 a few steps later, another naked subset appears. Unlike our first example that contained 3 boxes, this one needs four. Looking at column 4, boxes (1,4), (2,4), (4,4) and (6,4) all only contain a subset of the possibilities 1, 2, 5 and 6. We can therefore eliminate these as possibilites from (7,4), (8,4) and (9,4).

Hidden Subset

Definition

If X number of the same possibilities are only found in X number of boxes in a row, column or house, the all other possibilities can be removed from these boxes. X will always be 3 or 4. All X possibilities do not need to be in all of the boxes. Sets like 1,2 - 1,3 - 2,3 will all still count for X = 3. The important thing is that the set of possibilites only exist in the same number of boxes as there are possibilities in the set.

Example 1

8..6.7.......8..6..6.31........43..17.1..5...58.1.69..3.5..84......32.7.27...13..

Looking at house 3, we can see that only boxes (1,8), (1,9) and (2,9) contain the possibilities 3, 4 and 9. Since there are only three possible boxes for these three possibilities we can eliminate all possibilities that are not 3, 4 and 9. In this example we can remove 1 and 2 from (1,8), 2 from (1,9) and 2 and 7 from (2,9).

Coloring

Before getting to the definition of coloring, it is important to go over how to make coloring work. The basic idea is to look at one possibility at a time and find all of the conjugate pairs (only 2 of a possibility in a row, column or house). Congugate pairs are important because if one of them is the correct solution, we know for certain that the other one is not the solution and vice versa. Coloring is a way to visualize these pairs by assigning one color to one of them, and another color to the other one. Usually this is done with a dark and light version of the same color.

As you can see in the above example of basic coloring we have found 4 opportunities to apply coloring to the possibility 3. Looking at column 4, you can see that there are only two possibilties for that column. Therefore, we color one of them dark red, and the other one light red. Why do we have multiple colors on the form? Each conjugate pair get's its own color to start with. First we used red for column 4, then green for column 9. But what about the blue 3s?

The power of coloring comes when we start chaining conjugate pairs together. We can start to see this by looking at row 8. It has a conjugate pair that we have colored blue, and light blue. Our next step would be to color the conjugate pair in house 9. However, one of the 3's is already colored light blue. When one of the pairs is already colored, simply reuse the color and take the opposite. In this case, since the 3 in box (8,7) is light blue, we make the 3 in (7,7) dark blue.

While we were able to apply coloring to the set above, we were not able to make any conclusions just yet. The chains were too short.

Definition

Coloring is a technique that makes deductions off of conjugate pair chains. There are 2 different deductions that can be made using the coloring technique.

  1. If a non-colored possibility shared a row, column or house with both a light version, and a dark version of a color, it can be removed as a possibility.
  2. If the same version of a color (light or dark) is in the same row, column or house more then once, all of that version may be removed as possibilities. The opposite version can be used as a solution in the puzzle.

Example 1

..35.6.78...38....18...7..5...73928..3.2.89.48921.4..7.5.8.37.9..8.7...33..6.58.2

Looking at the red chain in this example will show us the benefit of the first rule of coloring. Box (8,7) is not part of any conjugate pair chains. However, it shares a row with box (8,6) and a column with box (1,7). Both of these boxes have a color, one dark and one light (in the same color chain). Since there is no contradition in the red chain, it is safe to say that either box (8,6) or box (1,7) will have a 1 in it eventually. This means that the 1 at (8,7) can be eliminated as a possibility.

It is very important to note that you can not make any inferences between the two different major colors (red and green in this case). For example you can not remove the ones in boxes (2,7) and (2,8) because the box at (2,6) is light and the box at (2,9) is dark. If the dark red and light green were eventually determined to be correct, then at least one of the boxes at (2,7) and (2,8) would have to be correct as well.

Example 2

591..7..2.6..591...........48.79.2159.51.4..817.5..9346..37...13..9.2...8..4.....

This example of coloring shows the contraditions in the conjugate pairs of the possiblity 3. Row 3, column 2 and Row 3, column 3 both have two dark blue boxes in them. Since we have a contradition, all of the dark blue boxes can be removed as possibilities and the light blue boxes can be set to solutions.

Multicoloring

Definition

Multicoloring is a way to reduce multiple coloring chains down into one larger chain that the coloring technique can then be used on. Chains are combinded when one variation of the first color shares a row, column or house with one variation of the second color and then the opposite variations of each of the colors also share the same row, column or house.

Example 1

Before
After

.8...319.......2.8...182.6.5.32687.9.984.75.6.6.9.58..8513.6.72.2.871.5.37.52.681

By looking at the before and after pictures above, you can see that the red and green colors have been combined. How was this done? In row 2, box (2,6) is colored dark green and box (2,8) is colored dark red. This satisfied the first condition where a version of each color is in the same row, column or house. Now, looking at column 7, we can see that there is both a light green and a light red sharing a column. This satisfies the second condition needed for joining chains since the opposite colors of each chain are in a common row, column or house.

Now that we know they can be joined, how do we join them? You can now treat all of the instances where the the two colors meet as being conjugate pairs. Since the red chain was shorter, the red color is converted to green. The colors are easy to switch now that we know were the two color chains intersect. Since the row 2 green variation was dark, we know that the dark red variation will need to be changed to light green. This can be done for all of the dark red variations. The light red variation is then switched to the dark green variation. We end up with a nice long coloring chain shown in green in the after picture.

The coloring technique can now be used to make eliminations on the non-colored possibilities. In this case possibilities in boxed (1,3), (2,1), (2,2), (2,3), (2,5) and (8,7) can all be removed.

Remember that the actual light/dark value associated with a color is just an indicator. The example above would have worked just as well if the light and dark values in the red chain had been switched (2 light and 1 dark).

You might be tempted to merge the blue chain into the green chain as well, but remember, you need opposite of both to be merged. Dark blue / light green and light blue / light green is not a valid way to combine strings.

XY-Wing

Definition

XY-Wing is a pivot based technique involving 3 pairs of possibilities. The pivot of the three pairs will have possibilities (A,B) the two wing pairs will have possibilities (A,C) and (B,C). The wing pairs must share a row, column or house with the piviot pair. Possibility C can then be removed from the intersection of the rows, columns or houses of the two wing pair boxes. The boxes that you are looking at must have only 2 possibilities in them for this technique to work.

Example 1

1.....2..4..2....8..7.9.....5.3.1.4..31.6.7.5......381.....6.5.5.2..7..39.6......

The piviot pair in this example is found at box (5,1). This box contains the possibilities 2 and 8. We are now looking for two other pairs that share a row, column or house with box (5,1), each of which have one of the possibilities (and not the same possibility). The box at (4,3) fills the need for the first wing since it has an 8 as one of it's possibilites. The box at (5,8) fits for our other pair since it has a 2 as one of it's possibilites. Now, we can use this as an XY-Wing pattern since they both contain 9 as their other possibility. This is really just a tricky way to do a naked subset where the constraint of being in a row, column or house has been removed.

The removals are now where the rows, columns and houses of the wings intersects. In both cases, the rows of the wing boxes intersect the houses of the other wing boxes. This intersection would include the boxes (4,7), (4,8), (4,9), (5,1), (5,2) and (5,3). Now you are never going to be able to remove possibilities from all of these boxes, however it is important to note that there are usually a bunch of places you can look. In the example above, boxes (4,7) and (4,9) both contain possibility 9 which can be removed since we know that either row 4 or house 6 will already have a 9 in it.

Guessing

If you generated your sudoku using Sudoku Helper then you will only get to guessing if you did something wrong. However, if you entered your sudoku manually it is entierly possible that the only way to solve it is to guess. Sudoku Helper has psychic powers and will always guess correctly with a Hint. However, you can always take a guess and if it doesn't work out use the Undo button and try again.