Tuesday, 20 July 2021

Locked out from Twitter

 I'm unable to log in to Twitter, since I don't have a mobile phone and am unwilling to let them know my landline telephone number, which already gets too many unwanted calls. The same would apply to any other online site. 

A notice on my twitter page, @mayhematics, it says there has been some unusual activity on my account. But the only difference from usual is that I used my old computer to access the site last week, since it was cooler in the back room where that computer is.

They would be better devoting their resources to stopping people who send nasty messages or fake news. I just try to entertain with my knight tour discoveries and occasional humour and logic.


Thursday, 15 July 2021

Emperor Solutions to the Onitiu Problem

Despite considerable efforts I have not made much further progress on the Quaternary Onitiu problem, that is of constructing knight tours with 90 degree rotational symmetry and with the square numbers in a knight path. These diagrams show the best I have found on the 10x10 and 14x14 boards in the form of Emperor tours which use wazir moves as well as knight moves.

10x10 Emperor tour
with 28 wazir moves

14x14 Emperor tour
with 40 wazir moves

I am now convinced that a knight tour solution on the 10x10 is impossible, and probably also on the 14x14 board, though I don't have a simple conclusive proof. In the above diagrams the square numbers are shown on the red line. The other coloured lines are rotations of the red line.

ADDENDUM (20 July 2021)

I have now solved the same problem on the 18x18 board, using only two wazir moves in each quarter. Can this be modified to make a knight tour solution? 

18x18 Emperor tour
with 8 wazir moves

This case is particularly interesting in having eight intersections instead of four. I hope to improve the diagram.




  

Friday, 18 June 2021

Onitiu Problem on 22x22 Boards

These diagrams are the results of a month's struggle to find a knight's tour with 90 degree rotational symmetry with the square numbers in a knight path (the red line). The tour on the circular board was completed first. Both tours use the same central pattern of knight moves. Both boards are of 22x22 = 484 cells. The yellow, blue and green paths are rotations of the red path. The circled cells mark where the coloured paths meet. These are at the points numbered 121, 242, 363, 484.

Circular Board (radius 13)
Dots mark {0,13} and {5,12} moves from centre.

Square Board 22 by 22
With square numbers and their rotations noted.

Addendum: 
Here is a black and white version that shows the overall and its symmetry pattern more clearly.

Black and White version.


These are the only solutions to the quaternary Onitiu problem I have found so far. Whether it is possible on smaller boards (of sides 10, 14, 18) is unknown. 


Friday, 28 May 2021

Circular Boards Again

 These new versions of the previous four tours keep strictly within the circle marked by the black dots. There are no cells whose corners go outside the circle. 

Radius 5

Radius root-50

Radius root-65

Radius root-85

As before the fourth shows 180 degree rotation, the others 90 degree.


Thursday, 27 May 2021

Smaller Circular Boards

Here are four tours on the smallest circular boards, with at least 12 fixed corners on the circle. These show boards with radii 5, root-50, root-65, root-85 corresponding to the smallest double-pattern fixed-distance leapers. The first three tours show 90 degree rotational symmetry but the fourth only has 180 degree rotation, due to the quarter-board containing an even number of cells (64). 

Radius 5

Radius root-50

Radius root-65

Radius root-85

There are 17 other two-pattern fixed-distance leapers to consider before the first three-pattern leaper of this type (root 325) is encountered, as used in the large tour previously shown. I doubt if I will get round to constructing examples of all of these! 

There is a certain arbitrariness concerning which cells, whose corners go beyond the circle, should be included or not.






Monday, 24 May 2021

Circular Board

 This is a knight's tour with 90 degree rotational symmetry on a circular board of 964 cells. The 24 black dots on the circle are at exact distance root-325 (approx 18.03) from the centre point. The four corner cells that make possible the birotary symmetry by ensuring the quarter-board has an odd number of cells (241) extend slightly beyond this circle to root-338 (approx 18.38).




The black dots are at points with coordinates {1, 18}, {6, 17} and {10,15} relative to the centre, while the corner points are at {13, 13}. 


Sunday, 23 May 2021

Antelope Pettern

Here is another to add to the list. I hope I've got it right. It was difficult to be sure a move wasn't missed.  



I suppose this is a sort of "painting by numbers".


Friday, 7 May 2021

Leaper Move Patterns

 I was inspired to produce these patterns as a contribution to the #Maydala season on Twitter for mathematics and art. They use rainbow-sequence colouring in place of the numbers of moves from the centre to other cells. 

Zebra {2,3}


Giraffe {1,4}


Knight {1,2}


Wazir {0,1}


 I suppose the Giraffe pattern could be cut off at the 9x9 size, if we stop when the first 8-move cells are reached.


Tuesday, 4 May 2021

Dawsonian Tours on Odd Boards

 It occurred to me that I haven't seen any examples of Dawsonian tours on odd square boards, so I composed a set. Of course closed circuits using all the square numbers are not possible. Other formations have to be considered. Here simple up and down steps.




A 5x5 knight solution is not possible but there is a simple wazir tour with the squares in this formation with a uniquely determined path, and also emperor solutions using knight and wazir moves.

 Diagrams:




Tuesday, 27 April 2021

The Onitiu Problem with Birotary Symmetry

 I've spent some time considering whether it might be possible to solve the Onitiu problem of constructing a knight tour with the square numbers in a knight chain but showing birotary symmetry (i.e. unchanged by 90 degree rotation) which is possible on boards of oddly even side (10, 14,18, 22, 26, etc). So far without success, though I can see no simple argument to prove such a tour impossible. 

Here is an Emperor (Knight + Wazir) tour that shows the general idea.

The red line indicates the square numbers. The other colours show its rotations. The black dots in the middle are the positions of 25, 50, 75 and 100 in this tour. It starts with 1 at d5. 

***

Earlier today I had my second vaccination dose against the coronavirus.

  


Wednesday, 14 April 2021

The Onitiu Problem 24x24

 I have now solved the Onitiu Problem on the 24 by 24 board. This is the first case where there are six intersections of the circuit of squares (shown in red) with the circuit of complements (shown in blue). The six square numbers where this occurs are 1, 36, 196, 289, 324, 484, the common difference being 288 which is half of 24x24 = 576.


   Once again the squares and diamonds feature strongly but even so I found this a difficult task.



Tuesday, 6 April 2021

The Onitiu Problem on Larger Boards

 I have decided to publish my results for the 12x12, 14x14 and 16x16 boards here rather than go to all the trouble of writing the subject up for publication in a journal. I no longer have the patience for that frankly. These were solved in reverse order, the largest 16x16 first because I thought its division into 16 areas 4x4 would make the solution easier by use of the squares and diamonds method in each quarter, and this proved to be the case. 

12x12

16x16

14x14

I'm not sure why Blogger has put these images in the wrong order! 

In the 14x14 case the sequences of squares (red) and the sequence of complements (green) have no points of intersection, but in the 16x16 case there are two (16 and 144 differing by 128), while in the 12x12 case there are four (two pairs, 9 and 81, 49, and 121, with difference of 72). 


Wednesday, 31 March 2021

The Onitiu Problem

 In 1939 in Fairy Chess Review the Romanian problemist Valeriu Onitiu solved the difficult problem of constructing a symmetric chessboard knight tour with the squares in a knight circuit. It turned out that there is only one possible solution. 


I have been looking at this problem to see if it is solvable on larger boards. Initially I found it very difficult. However eventually on 27 March I found a solution on the 16x16 board (which can be split up into 4x4 sections which makes the solution by 'squares and diamonds method' feasible). This quickly led to solution on the 14x14 board, then on the 12x12 board, and finally tonight I solved the 10x10 case. I posted a copy on Twitter (misdated 2921). Here is a better presentation using coloured lines to mark the sequence of squares and the sequence of their complements.



I think I will reserve the other solutions for publication elsewhere in some chess or mathematics journal, where I can write it up in more detail. 


Demolition in Progress

 Back in 12 May 2019 and 4 May 2020 I showed photos of the Crewe town centre area where preparations were being made for demolition of many of the old shops including the British Home Stores (BHS) building on the corner of Queen Street and Victoria Street, and also the Clock tower which I had hoped could be saved. Here are some recent photos of the actual demolition in progress. 







Photos taken in January (2), February (1), and March (2) of  2021.

Most of the buildings have now been reduced to piles of rubble.


 

Monday, 15 March 2021

Trees

 Some snaps of trees from recent walks round Crewe. 





The above four are from the Brooklands area between Ford Lane and Broad Street.


This, taken a few days before, is in front of the police station, library, and swimming baths, and on the corner of the grounds of the old bombed-out church.


Saturday, 27 February 2021

Dawsonian 12x12 Tour

 Following on from the 10x10 tours with square numbers in knight chains it suddenly occurred to me that the 12 squares on the 12x12 board could be arranged to show a 3-4-5 triangle.

3-4-5 triangle

It will be seen that my solution makes much use of squares and diamonds in the nine 4x4 areas that the board divides into. It may be improvable in this respect.


Thursday, 25 February 2021

Dawsonian 10 by 10 Tours

 These tours that I have been posting initially on Twitter are part of some work I am doing to put an updated page on Figured Tours on my Knight's Tour Notes pages. I found I had no examples of Dawsonian tours on larger boards. That is tours showing the square numbers in knight paths, preferably symmetric circuits. T. R. Dawson constructed a complete set of such tours on the 8x8 board.

Rectangle 1x4

Hexagon

Octagon

Constructing these tours makes a good puzzle for solving. Probably the larger board makes it a little easier to complete the paths than on the 8x8 board. My procedure is to start with the shorter sections 1-4, 4-9, 9-16 and so on, trying not to leave inaccessible unused cells. Fitting the last two segments 64-81, 81-100 are of course the tricky part of the problem. Sometimes one ends up with two loose ends that do not connect by a knight move.

Addendum: A fourth example to complete the set of symmetric convex polygons.

Rectangle 2x3


Tuesday, 9 February 2021

Tiles and Key Patterns

 In my Knight's Tour Notes on Wazir tours and in various other places I included the following diagram outlining the various ways of forming a frieze with vertical and horizontal moves. 

Wazir Paths

In Twitter I have been following a series of items by Tom Ruen on tiling with a wide range of various shapes. His most recent articles inspired me to put together the following two Greek Key Pattern borders. They are constructed in each case of a single rectangular (metasquare) tile pattern, in two colours and various different orientations.

Key Pattern 1

Key Pattern 2

This would make a good frame for a shaving mirror!