| 
 
	
		| View previous topic :: View next topic |  
		| Author | Message |  
		| coolio Guest
 
 
 
 
 
 
 | 
			
				|  Posted: Sun Jan 15, 2006 10:01 pm    Post subject: pencial marks |   |  
				| 
 |  
				| [i]whenever i do this i always have pencial mark from erasing and ersaing is there anyway to help me figure out how to play so that i am able to have not as many pencial marks this way it will be easier to see the # that i have put in[/b] |  |  
		| Back to top |  |  
		|  |  
		| alanr555 
 
 
 Joined: 01 Aug 2005
 Posts: 198
 Location: Bideford Devon EX39
 
 | 
			
				|  Posted: Mon Jan 16, 2006 2:51 am    Post subject: Re: pencil marks |   |  
				| 
 |  
				|  	  | Code: |  	  | > I always have pencil marks.
 > Is there anyway to help me figure out how to play so that I
 > am able to have not as many pencil marks.
 
 For Easy and Medium puzzles, it should be possible in most cases
 to solve the puzzle without pencil marks - although some Medium
 puzzles have thwarted me without marks!
 
 A lot of the Hard and Very Hard puzzles can be progressed a long
 way using the Mandatory Pairs marks. By their nature, these are
 added and removed during the solution process and there will
 never be as many of them as the marks for candidate profiles
 if the latter are fully completed. However, there are quite a few
 puzzles where resort to candidate profiles reaches the solution
 much quicker than relying upon just Mandatory Pairs.
 
 It mus be remembered that Mandatory Pairs is just a reasonably
 efficient method of recording "information" gleaned from the
 puzzle in a way that enables it to be used later - whereas using
 Candidate Profiles is akin to simulating the work of a computer
 and recognising patterns that will reveal the solution.
 
 The Mandatory Pairs method (M/P) falls short when the puzzle
 is based on pairs/triples etc which span more than one 3x3
 region. Recognising this I find that more time needs to be given
 to "Sole Position for Digit" searches if the simpler M/P techniques
 fail to uncover the solution.
 
 The standard technique "Counting" will reveal what numbers are
 missing from any line (row/col) or region. There are then two
 useful techniques.
 
 a) Sole Candidate: This applies when there is only one digit that
 will fit into an empty cell. The method is to consider each empty
 cell and use "counting" to determine which digits are eliminated
 by being in rows or columns intersecting with that cell or in the
 region that contains the cell. If it is possible to find EIGHT such
 digits then the ninth MUST be the solution for that cell.
 
 (Of course, there is also a useful variant whereby the technique
 is applied to two cells in a line and the same two digits are found
 to be the only candidates but this is "Hidden Pairs" - demonstrating
 that Sole Candidate is but a simplified form of the Pair/Triple etc)
 
 b) Sole Position: This technique considers each digit in turn and
 looks at each empty cell in a line (of course one ignores digits
 that have already been resolved within the line!) to determine
 if it is possible to place the digit in each empty cell. If it is shewn
 as possible in more than one cell one moves on to the next digit
 to be considered but if only ONE cell in the line can contain the
 digit under consideration then that cell is resolved. This can be
 applied also a region in a similar way to a line.
 
 Both these techniques are contained implicitly in the candidate
 profile approach. The first occurs when there is only a single
 candidate for a cell and the second when the profiles for a line
 (or region!) contain a digit which appears only once. Of course
 finding that involves a closer inspection than sole candidate!
 
 With all of these techniques the REAL skill is based on
 
 a) knowing which technique to use next
 b) Selecting which line, region or digit to be the focus for the
 next application of a technique.
 
 One of the advantages of Mandatory Pairs is that it can 'root out'
 a lot of the easier stuff and many of the "hidden pairs" in a far
 more transparent manner than working through the candidate
 profiles. This, in turn, reduces the number of occurrences with
 which one is dealing in preparing the candidate profiles if it
 becomes necessary to use the latter. Also some of the Candidate
 techniques become redundant. For example the rule that "if the
 only occurrences of a digit in a line are all in one region that digit
 may be eliminated from the other two lines in the region" can be
 difficult to spot. With M/P it becomes almost self-evident - certainly
 obviating the need to scan profiles in the hope of spotting a pattern!
 A similar situation arises with the rule about all occurrences of a
 digit within a region being in a single row or column.
 
 +++
 To return to the main question - pencil marks.
 
 I never make "pencil" marks - as I always use a pen! However, if
 I were to use a pen and eraser there would be definite advantages
 in using Mandatory Pairs - the marks would be erased.
 
 Two other tips:
 
 1) When I have completed the puzzle as far as I can with M/P and
 decide to progress to Candidate profiles, I change the colour of
 my pen. This means that I have a simple "break-point" should I
 make a mistake with the profiles. My experience is that mistakes
 are MUCH more common with candidate profiles than with the
 M/P part of the exercise. Applying a break point was a lesson
 learned the hard way - although in most cases it is not necessary
 to implement it!
 
 2) With so many digits littering the page (and not using an eraser!)
 I use the convention of circling the resolved digit (also written large
 near to the centre of the cell) to distinguish it form the workings.
 
 This may not be necessary if one is working in pencil and erasing
 the digits and the alternative of using different colours for final
 and working digits would involve frequent mistakes in which pen
 to use.
 
 I am aware that some commentators use the circling of numbers
 to denote particular types of working - but I do not recall having
 seen reference to this on THIS site. Those other people refer to
 BIG numbers to distinguish the solution from the "tiny writing" as
 one columnist referred to it in Spring 2005. My own view is that
 "bigness" is not a adequately clear visual distinction - unless one
 is a pedant for writing size, which I am not. However each to her
 or his own when developing such conventions.
 
 My personal convention is to put Candidate Profiles at the top left
 of a cell and Mandatory Pair digits at bottom left - leaving the
 bottom right for the "resolution sequence" number (this is just an
 ascending sequential number incremented for each cell resolved
 and which allows for tracking back). However, I am very prone to
 errors with the these tracking numbers (more duplications than
 omissions) as I pay heed to the resolutions rather than the next
 number in the series and so the technique does not always prove
 useful and does demand printing the LARGE version of the puzzles
 as there is insufficient room for sequence numbers on the so-called
 "medium" grids.
 
 (PS: Why is "medium" the default for printing the grids? Only one
 grid appears at a time and so have a smaller print does not save
 paper. The default just means having to change it each time!!)
 
 +++
 
 So, yes - there are ways to "manage" pencil marks.
 
 One way is to get Sudoku San to print them with the grid - as
 SamGJ has not yet acceded to the request to have a "sweep"
 option included with the initial print facility in "Draw". As the
 facility develops a solution in order to determine a grade for
 each grid entered it MUST derive the initial profiles. All that
 was requested was an option for them to be printed as
 superscripts in a manner similar to the one used by Sudoku San.
 
 Another way is to use Mandatory Pairs as the first resort if the
 puzzle is clearly one not to be tacked in a "mark-less" way.
 
 The "BIG" numbers could be distinguished by circles and key
 points in the solution process could be marked by a change of
 colour of the writing instrument used.
 
 There are probably several more.
 The challenge of Sudoku is to be creative!!!
 
 Alan Rayner  BS23 2QT
 
 | 
 |  |  
		| Back to top |  |  
		|  |  
		| dotdot 
 
 
 Joined: 07 Dec 2005
 Posts: 29
 Location: oberseen
 
 | 
			
				|  Posted: Mon Jan 16, 2006 8:57 am    Post subject: pencil marks |   |  
				| 
 |  
				| See also the discussion here. 
 I reserve erasure for correcting (immediately recognised) mistakes.
 |  |  
		| Back to top |  |  
		|  |  
		| PseudoCool Guest
 
 
 
 
 
 
 | 
			
				|  Posted: Fri Feb 17, 2006 5:38 pm    Post subject: Re: pencial marks |   |  
				| 
 |  
				|  	  | coolio wrote: |  	  | [i]whenever i do this i always have pencial mark from erasing and ersaing is there anyway to help me figure out how to play so that i am able to have not as many pencial marks this way it will be easier to see the # that i have put in[/b] | 
 
 When I find a need for pencil marks, I use tiny dots to indicate a cell's candidates, according to this pattern:
 
 1 2 3
 4 5 6
 7 8 9
 
 E.g., candidates 1, 2, 6, 9 would be marked like this in the cell:
 
 
* *
 *
 *
 
 When I've identified that a given number is constrained to one of two cells, I'll indicate that with a slightly heavier mark in those cells.
 
 Then I need only erase tiny dots, and can enter a cell's number with a pen.
 
 HTH!
 |  |  
		| Back to top |  |  
		|  |  
		|  |  
  
	| 
 
 | You cannot post new topics in this forum You cannot reply to topics in this forum
 You cannot edit your posts in this forum
 You cannot delete your posts in this forum
 You cannot vote in polls in this forum
 
 |  
 Powered by phpBB © 2001, 2005 phpBB Group
 
 |