Othello news

Egaroucid: An Othello app review

Egaroucid was created by Takuto Yamana. Yamana-san is a 22 year old student, who enrolled at the college of engineering systems in Tsukuda, Japan, in 2020. Yamana-san began development of Egaroucid a year after enrollment and started playing the game shortly thereafter, in 2022. WOF News had a talk with him about his program Egaroucid.

"I started creating Othello AI in April 2021. I am developing it for 3 years. in November 2021, I got the 1st rank in CodinGame Othello. I started to create Egaroucid application in December, 2021, after I got the 1st rank. I spent 3 months to release the first version of Egaroucid app. After that, I update it irregularly. I started developing Egaroucid by chance. In April 2021, my friend wrote a program of Othello rules (not Othello AI). I imitated it, and then started creating AI. I created some other works related to Othello AI. "Isevot" is a set of robots that play Othello themselves with Egaroucid. "Cotechnology" is an Othello AI that runs parallel on micro controllers. "Computer Othello" is a remodeled game machine. The original machine was sold around 1980. I put a latest micro controller in it. "Retro Othello AI" is an Othello AI that runs parallel on 8 bit micro controllers (the spec is almost the same as computers 50 years ago). Before Othello AI, I used to build robots that solve rubik's cubes."

Isovot robots playing Othello using 
Egaroucid.

"Cotechnology" Othello AI running on paralell 
microcontrollers.

"Computer Othello" remodelled game engine 
from 1980.

"Retro Othello AI" running on 8-bit paralell 
microcontrollers.

Robots solving Rubik's cubes

Is there someone you look up to?

"I respect Mr.Takeshi Murakami, who is the world champion of 1996, 1998, 2000, and the representative of human in Othello AI (Logistello) vs Human battle in 1997. Also, I respect Mr.Michael Buro, the developer of Logistello. I am very interested in the match in 1997, as I am both a developer of Othello AI and an Othello player, and I would like to create an Othello AI that is very useful for human Othello players."

What are your favorite features in Egaroucid?

"My favorite features are input/output features. Egaroucid handles many formats about Othello, and I use these features a lot. Also, there is a feature "Edit and Save Board Image" function in output menu. This feature is created because my friend (6-dan Othello player) was writing a textbook of Othello. I thought this function can be useful. In addition, there is a feature that generates a board randomly. This feature is created because my friend (2-dan Othello Player) told me that she wanted to practice estimating the evaluation value of a board. So I added "Generate Random Board" function"

Were there any challenges during development?

"The biggest problem in developing Egaroucid was UI design. I am not good at designing easy-to-use UI, so I asked my friend, Kaneko Eizo to make a UI design proposal. Also, I was [and still am] struggling to make Egaroucid stronger and faster. I tried, [and still am trying] a lot of methods and found [or I hope to find] a good one."
Is there a feature you would like to include in future versions of Egaroucid? "In the future version, I am planning to add features that tell users "why this move is good/bad". This is the main issue that I have been interested for these 3 years."

Egaroucid app review

Egaroucid, just like its predecessors Wzebra and Saio, is an engine designed to play and analyze Ohello. Egaroucid was developed by Takuto Yamana. Yamanaka is an engineer enrolled at the College of Engineering Systems, University of Tsukuba.

Main features:

⦁    You can play human v human or human v the engine
⦁    You can save your game or import games for analysis
⦁    You can input the board state as a string of moves or a string of dashes, Xs and Os to denote empty squares, white discs as Os and black discs as Xs. The position needs to be thought of as starting from the top left corner of the board.
⦁    You can save and analyze branches of positions
⦁    You can take screenshots of board positions (and edit them). You can even change the pictures to monochrome and make educational material.
⦁    You can randomize the starting position
⦁    The app has good multithreadding, making it ideal to use with modern high core & thread count processors and increasing search speeds. You can set the number of threads up to 48, in case you have a Threadripper or Xeon processor.
⦁    You can set the size of the hash table to increase the search speed further (although a hash table level above 25 will double memory usage from 0.75 GB to 1.5 and 3.0 GB per level above 25).
⦁    Easy to use and intuitive UI that minimizes unnecessary clicks.
⦁    Three language options: English, Japanese & Chinese.
⦁    Online manual (in Japanese) that can be read using a translation tool

Review:
Egaroucid is nothing less than excellent! It has all the features one would want from a top tier Othello engine. Thanks to the excellent multithreadding, analysis of games is blazingly fast. Being able to compare multiple branches for analysis is nice. Standout features include being able to randomize the board starting position, which makes the engine excellent for training for XOT; and the ability to take screenshots in color or monochrome makes the engine an ideal tool for creating educational materials I suppose the only thing missing is Thor database integration.