Monday, January 2, 2017

CGT Facebook group

Simon Rubinstein-Salzedo started a Combinatorial Game Theory Facebook group!  Join us here:

Happy New Year!

Friday, December 30, 2016

All I want for Christmas is Tweets

In 2016, I joined the technology era of 2006 by getting myself a Wii and now a Twitter account.  Hooray!

This summer, I promised myself that I would sign up for Twitter after I submitted my tenure portfolio.  Well, I turned that all in back in October, so this is overdue.  Here's my Twitter-Sphere Presence: CGTKyle

I find myself noticing CGT things that don't quite warrant an entire blog post.  Then I forget about those things.  Now I can quickly say something about them in an attempt to garner attention.

I expect that I'll also be tweeting about general board games, teaching, voting systems, CS and math, and more.  Yikes.

My first tweet announced a little page I made for CGT events:  I realized that I always needed the links to these previous and coming events.  As is common for me, I coded the whole thing up in Javascript (feel free to look at the underlying code) so events like CGTC II is currently in the "Future" section, but should move to "Past" after it's happened.  We'll see whether I did that right.  (Also, you should all be going to CGTC2!  I'm very disappointed that I'll be missing it.)

I think I'm missing some meetings on there (there was something at Kamloops a few years back, I believe) so help me fill it out.  I hope there's something coming up this summer also (I think Games at Dal will have to wait until 2018.)

Happy New Year!

Thursday, December 1, 2016

LaTeXing Lecture Notes

I've been working hard the past 2.5 years teaching a bunch of new courses and improving old courses.  One of the things that's helped a ton is converting hand-written lecture notes into LaTeX. 

Many years ago, I started working on a style file to add a bunch of commands for lecture notes.  Most important, I created a question command to facilitate Socratic-style lectures.  I later modified this so that I could set a flag indicating that the output was for students.  Then it would hide the answers to those questions.  I started posting these versions online for students.

I've been adding stuff to this, but recently really wished I could add exercises that would put all the answers in an appendix, but hide those answers in the student version.  I recently discovered the exercise package (thanks, Stack Exchange) which can do this automatically.  Using that, I updated my style file, and recently posted the code on GitHub.  (If you're not familiar with GitHub, I made my own landing page.)

I find that it's really convenient to teach from the pdfs on a tablet, and it saves me tons of paper.  The only issue is that it's a bit more difficult to make notes on the pdfs than with a regular pen and paper.

Friday, September 30, 2016

Richard Guy is 100

Happy Birthday, Richard Guy!

Richard Guy turns 100 today.  He is one of the authors of the classic text Winning Ways for your Mathematical Plays.  At the CGT workshop at BIRS in 2011, he proposed a new subtraction game that I find myself thinking about on and off:

Given a pile of n tokens, and a subtraction set, S, each turn, you move to a new position, n (< n'), S', where either:
  • Normal thing: S = S' and n' = n - s, for some s in S or,
  • Weird thing: S' = { x } U S and n' = n - x and there are some y and z in S such that x = y + z.
...I haven't gotten anywhere with it either.

I was alerted to an excellent birthday video for Richard, highlighting his work in Geometry.

Richard is one of the nicest people I've ever met, and he continues to be active in research today!

Saturday, August 20, 2016

Integers 2016

Integers is back!

Integers is a conference in combinatorial number theory.  Traditionally it's been held every other year on the odd years, but it had to be skipped last year.  Integers is a great, low-key conference held at the University of West Georgia.  It has an associated electronic journal with no paywalls (so convenient!).  It's well run and everyone is very nice.  This year, they've moved it up to early October (6th-9th).  Here's the official Integers 2016 announcement.

Historically, there's been a contingent of gamesters attending Integers.  In 2011, I brought an undergraduate and met many of Richard Nowakowski's students.  In 2013, I met Silvia Heubach and Matthieu Dufour.

Unfortunately, I don't know of anyone attending Integers this time around.  This is partly due to the number of CGT events popping up.  Games@Dal has returned, and CGTC2 will be happening in January 2017 in Lisbon.

If you're planning to go for Combinatorial Games, please comment below (or email me).  At this point, it looks like we might not have any gamester attendees.

Wednesday, August 17, 2016

Games@Dal 2016: Tanya talks about Cookie Monster Games

Games@Dal 2016 talk: "Cookie Monster Plays Games" - Tanya Khosanova w/Leigh Marie Braswell, Eric Nie, Alok Puranik, Joshua Ziong, and Dhroova Aiylam

Tanya Khovanova shared some work with a bunch of high schoolers (whoa!) on patterns in Cookie Monster games.  Cookie Monster games are Nim games where k sticks (cookies) can be removed from multiple heaps.  Each ruleset has a different restriction on which sets of piles can be removed from.

She and her students considered rulesets where you can take from...:
  • No restriction
  • One-or-all piles
  • One-or-two piles
  • Any consecutive piles (assuming the piles are in a list)
  • One or two consecutive piles
  • Any set of piles including the first jar
  • Any odd number of piles.  (It turns out that the P positions are the same as in Nim!)
  • Any set of piles except all of them.
  • ... and more!

In one of these games, she noticed a surprising correlation with an automaton problem!  A sequence of the number of P positions is related to the number of new cells born from the Ulam-Warburton automaton.  She then looked more closely at other relationships to automaton!

Tuesday, August 16, 2016

Games@Dal 2016: Urban talks about Drawing Triangles

Games@Dal 2016 talks: "Hopeful Windows in Cellular Automata and Combinatorial Games" - Urban Larsson

Urban talked about the Hopeful Window Triangle-Placing Game, a very interesting game based on his work with cellular automata games.  In this game, there are multiple steps per turn revolving around drawing a digital 45-45-90 triangle on a grid.  The triangle always has one box at the top, and the hypotenuse descends on the left-hand side.  There is a single box lower in the grid that is the blocking box:
  • No triangle can be drawn that contains that box, and
  • The first player to draw a triangle past that box wins.
At the beginning of each turn, there are candidate boxes that the current player can start with are all in a horizontal row.  (I'll explain how those are generated as part of the turn below.)  The steps to each turn are:
  • The current player chooses one of the candidate boxes for the top of the triangle, then draws it as far down as they want.  (If they start out to the left of the blocking box, then they can automatically win by drawing a large-enough triangle.  Otherwise, they can't cover up the blocking box.)
  • Consider the boxes directly beneath the drawn triangle - the triangle's support, but increase this set by extending to the left and the right by two boxes on each side.  (This can be parameterized to be any l and r numbers.)
  • Inside of this elongated support, the next player chooses a "hopeful window" of, say, 6 adjacent boxes.  (Again, this can be parameterized.)
  • The previous player then gets to block, say, 3 of the boxes in that window.  (Parameterizeable again.)
  • The remaining 3 (in this example) boxes from that window become the candidate top boxes for the next player to choose from on their turn.
The game is certainly a bit complex, but Urban found an extremely surprising relationship between this game and automata that generate Sierpinski-Triangle figures.  By using the l, r, w (window size), and b (number of block) parameters in the formula for the triangle-creating automata, the resulting pattern of triangles tells you exactly which boxes you should choose to start a new triangle from: choose one of exactly those which remain uncolored in the diagram.  Amazing!