Sprouts2025 Wrap-up

Sprouts 2025 was a big one for us!  We broke a lot of records:

• Most talks: 13
• Most student talks: 11
• Most submitted computer players: 6  (The previous record was 2!)
• Most participants in the human tournament: 16

Other exciting things that happened:

• The computer player tournament was a real competition this year!  All six submitted players seemed better than the Gorgons players that came in last year.  You can try playing against them!  (I can say that about the Gorgons players because both people submitted better players this year.)
• Ian Grenville won the computer tournament again!  Efforts to dethrone him may be the big story next year.  
• Hiyu and Kouta participated in the entire conference, despite it running until 6am in Japan!  (I feel guilty, so I want to reiterate: it's okay for people to not attend the entire thing... especially to get sleep!)
  
• Hiyu won the human tournament and then beat HotShot to win the John Henry match.  

Things out of our control that went well:

• The student presentations were great!  I think I understood the main point of all of them.  There were a lot of really cool results!
• Everyone was very pleasant during the human tournament.  There was often a strong bias towards one player in the starting positions, but people took it in stride. 
  
• All presenters stuck to the timeline and everything ran smoothly. 
  
• The conversations during the computer tournament were great.  People were really curious and chatting about how each of the players worked.  
  

Things we had some control over:

• We had the John Henry match set up ahead of time with all the possible winners of the computer tournament so that we could run that right away.  (We didn't do it last year.)
• I used a script to organize the abstracts and talk schedule on the site which helped me keep on top of that.
• We were continuing to register people for the conference through the morning session.  (I think we might need to automate some of that in future years.)  I might consider this annoying, except that I was so excited that people kept wanting to join in.

The biggest negative unexpected thing that happened is that we went until 5pm because the human tournament ran longer than expected.  I think the human tournament ended at 4:15 instead of 3:45.  The computer tournament had four rounds, so I think that and the John Henry match ended at about 4:45.  Nevertheless, a bunch of people stayed around for the "End with Friends" until 5pm.  We closed the zoom down at 5:10.

I am so excited about how much of a success this was this year.  I hope everyone had a great time and that Sprouts continues to be this great!

Sprouts 2025 Keynote and Afternoon Session Summaries

Yesterday was Sprouts 2025.  In the afternoon session, we had our keynote by Matt Ferland of Dickinson College and six student talks.  Here are my (likely lacking) summaries of those talks.

 

Matt Ferland, Keynote: "Supernim is Much Harder than Nim"

Matt covered the Nim basics, then talked about superstars: games where all options are to nimbers.  Supernim is, then, the ruleset of any sums of superstars.  Matt talked about reductions to help relate computational problems.  (He described this using curing cancer to show the potential impossibility of solving hard problems.)  Matt covered NP and 3-SAT, then looked for a relationship from those to Supernim.  Unfortunately, an attempt at XORSAT did not yield a reduction.  Matt then described MXOR: Multi-state XOR, where variables can have one of many values instead of just boolean.  Then the literals each indicate for which of those assigned values they will evaluate to true.  Matt showed that MXOR is NP-hard, then showed the subset of Supernim positions and the reduction of MXOR to Supernim.

 

Nakul Srikanth: "GamesmanROS - A Generalized Game Playing Robotics Software Stack"

Nakul talked about his work with GamesCrafters, a long term (25+ years) project run by Dan Garcia at UC Berkeley.  Nakul's contribution has been to incorporate robotics into their automated players.  He talked about the physical constraints they deal with in general and then for specific rulesets.  He showed how the robots use string representations of positions to perform some of the logic.  Different types of games (Placement vs Capture) present different problems that they need to overcome.  He showed a great demo of it in action!

 

Jake Andersen & Cam Jefferys: "Developing Online Martian Chess"

Jake and Cam talked about creating a playable JavaScript Martian Chess program.  They described Martian Chess, a scoring game where control of the pieces changed hand depending on which side of the board they're on.  They showed the rules using screenshots, then did a quick demo.  They talked about how they trained a neural net player by reimplementing the game in python, then ported the model back to JavaScript.  They gave a high-level description of how the player model works.

 

Ethan Saunders: "Paintbucket and Paintbucket on Graphs"

Ethan talked about a partizan flood-fill game called Paintbucket.  The game started from being played among Dalhousie students using MSPaint.  Each turn you use the flood fill tool on a black-and-white image to change one contiguous color of your opponent's to your own.  He then talked about the generalization on graphs.  He was able to show that this version is PSPACE-complete.  One cool part about this presentation is that he used MSPaint to give the talk!

 

Caleb Aukeman: "Partial Jenga"

Caleb continued research from a previous study of Jenga by looking at a partizan version where the blocks are colored Blue and Red.  He talked about conditions for when layers are unplayable based on which blocks are missing.  He also categorized levels based on which blocks were red and blue.  (He had great graphics in his talk!)  He found switches with heat of 1/2 and showed families of positions that are in P.

 

Brendan Whitmire: "Gomoku in JavaScript"

Brendan talked about his JavaScript implementation of Gomoku.  He showed that he added the Pie Rule to his program, which took some tricky coding in the JS framework he is working with.  He talked about the AI "Evolved Player" he's creating.  He generated it by pitting the players against each other using an evolutionary learning algorithm using the Python NEAT library.

 

Brenden Gray: "Cops and a Destructive Robber"

Brenden talked about his work on a pursuit-evasion game on graphs.  His is a variant of Cops and Robbers where the Robber damages the vertex they start their turn on.  The cop's goal is to minimize the number of damaged vertices, with a secondary goal of capturing the robber.  (Damaged vertices do not prevent movement by either player.)  This can change the strategy from the original version because the cop may be content to not capture the robber if it will drop the amount of damage.  He showed a graph where the cop's best strategy is to stand still.

 

Sprouts 2025 Morning Talks

Sprouts 2025 was today!  In recent years, we have included early morning talks for presenters who are in so far away of a time zone that they should be asleep during our normal talks.  (At this point, this just means presenters from Asia... and one who couldn't attend later in the day.)  Even at the early hour we started, Japan is 13 hours ahead of us, so it was already quite late in the evening for them.  This year we had a record five talks during that morning session!  Here are my (spotty) summaries of those talks. 

 

Akshat Patodia & Parth Sarda: "Blippers and Flippers"

Parth and Akshat did a team talk describing two games: Blippers and Flippers.  These are inspired by generalizations of Toads and Frogs that parameterize the different allowed distances of moves and whether you can jump pieces of your own team.  Blippers and Flippers are played on 2-D grids using black and white stones instead of the usual amphibians.  Those stones are placed and removed based on rules for orthogonally-connected groups.  They were able to find some loopy positions on strips in Flippers.  

Hiyu Inoue: "On Ending Partizan Subtraction Nim"

Ending Partizan Subtraction Nim is a game where both players have the same moves (subtraction set) but the winning conditions are different.  E.g.: Left wins if there are even tokens when play reaches the terminal position, but Right wins if there are odd tokens.  This seems like a pretty evenly-biased rule, but Hiyu showed that even on subtraction set {2,3}, L shows up as an outcome class surprisingly more often then R.  This trend continues with many other subtraction sets that Hiyu studied!  He proved this is the case using a strategy-stealing argument.

Harman Agrawal: "QuadroCount: A Combinatorial Game"

Harman presented the game she's been studying: QuadroCount.  She implemented a playable version of the game online.  Each turn consists of moving one of that player's pieces in a way that lowers the "score".  The score is measured as the total size of all rectangles with each pair of opposing color stones, the "ola": OverLapping Area.  Harman has found new terminal positions.  She refers to some of them as Nash Equilibria positions because although it seems like moves should be available, all pieces are locked.  She also provided some open problems about other potential terminal positions. 

Melanie Gauthier: "Snort Strategies on Triangular Grid Graphs"

Melanie covered the basics of combinatorial games and, in particular, placement games, then described Snort.  She uses tinting to simplify diagrams during analysis (which is something Craig and I have been using in Col).  She studied winnability on initially-uncolored graphs.  She used the one-hand-tied principle to break apart triangular grid graphs after a move.  She found a ton of long graphs with two and three rows to be in N.

Kouta Kawakami: "New Rulesets and their Transfinite Grundy Values"

Kouta talked about extending nimbers by including transfinite Grundy values.  These can occur in non-loopy games that don't end in finite moves.  For example, on a Nim variant on lists where you have to remove from one pile, but then can simultaneously add to another pile with a higher index.  This yields transfinite G-values, which Kouta was able to find for instances of this game.  He also used transfinite numbers to analyze a separate single-heap non-finite game.