E-Fanorona is running for the igam4er.org 2014 competition.
Why do e-fanorona fit into this competition?
Via Nelli-Studio, the Madagascar-based game studio behind the e-fanorona mobile app, we are competing for igam4er.org as professionals.
Apart from the cultural aspects of e-Fanorona, the Fanorona concept in itself involves a lot of mathematical concepts – especially decision tree algorithm.
The challenge for Nelli Studio while implementing this board game resides into the implementation of artificial intelligence so that the player can have three levels of difficulties when he wants to measure himself against his smartphone (or tablet).
On the other hand, a special thought have been integrated so that visuals are appealing while confirming Nelli Studio’s aim to promote malagasy culture.
Fanorona: the game
Fanorona is a malagasy (from Madagascar, the island, not the movie) traditional board game similar to chess.
There are different flavors of Fanorona -requiring different algorithms. As an example, the usual « Fanoron-tsivy » that is implemented into the e-fanorona app requires both players to start with 22 pawns. However, with the « Vela » rule (also implemented into the e-fanorona app), the second player can only start to « eat » his opponent’s pawn once the setting goes back to 22 vs 5 (22 pawns for the first player and 5 pawns for the second one)- this makes the game a lot more interesting.
Choice of platform for efanorona: mobile devices
Mobile devices have become part of most people’s daily lives. Nelli Studio is making the challenge of bringing traditional malagasy board game to a brand new set of users.
Nelli Studio value the cultural dimension of the games it is developing. The user will notice the reference to malagasy culture all along the game. Moreover, the efanorona team is mainly composed of malagasy technicians and creatives.
Being a mobile game development studio, Nelli Studio have developed the e-fanorona project on mobile devices (smartphones as well as tablets). We specifically have developed a version for e-Fanorona for two major mobile platforms: Google PlayStore and AppStore. Feel free to download the version that fits to your mobile device.
The scientific approach
Fanorona is a a Malagasy ancestral game that is still practiced by Malagasy people nowadays. It requires memory aptitudes as well as strategy thinking that we usually find in games like Chess.
Through this project, we want to answer the following questions: can artificial-intelligence-based efanorona application beat human champions of Fanorona? Can the e-fanorona project find new winning-strategies that practitioners of traditional Fanorona haven’t experimented yet?
Through the e-fanorona project, we want to create a similar experience to the famous human VS Computer Chess challenge opposing Garry Kasparov VS Deep Blue (http://en.wikipedia.org/wiki/Deep_Blue_versus_Garry_Kasparov)
There are few differences worth mentioning though:
– Fanorona has a whole different set of rules that make it a very challenging concept to implement when it comes to Artificial Intelligence.
– This time, we won’t have Deep Blue computer, we want to put Artificial Intelligence inside an ordinary smartphone running Android or IOS so that the lambda user can measure itself to world-class computer-based Artificial Intelligence.
We assume that champions of traditional Fanorona have a lot of experiences as well as winning-strategies they trust on.
We also assume that the smartphone or tablet where the e-fanorona project is installed have enough computing power so as to fully implement the artificial intelligence part of the project.
- A smartphone (or tablet) running Android or Apple IOS
- The e-fanorona app that can be freely downloaded from Google PlayStore or AppStore
Procedure and rules of the Fanorona game:
The “Fanoron-tsivy” starts with 22 pawns for each opponent. An player wins the game when his opponent has no pawn left.
One can only move following the lines, however, one cannot move to the same direction twice, and one cannot go back to a position where one have already been for the same/continuous move/combo
Eating opponent’s pawn
There are two ways for “eating”: either by approaching, or by “leaving”. Opponent’s pawn can only be eaten altogether if they are contiguous.
- Implement the rules of Fanorona into the e-fanorona app: define authorized moves of the pawns. We use 8-bit binary masks so as to define the authorized moves based on each position on the Fanorona Board. We will have different set of authorized move matrix depending on the “fanorona” rules we want to implement. For this project, we only consider the following rules:
– Fanoron-tsivy: the board is composed of 9 columns. Each component start with 22 pawns and both can start “eating” his component’s pawn from the beginning. This is the game setting where we introduce Artificial Intelligence
– » Vela »: this is the same setting as with “Fanoron-tsivy” with the major difference that the second player who plays the Vela can only start “eating” his opponent’s pawns when he’s only left with 5 pawns while his opponent still has 22. Although an extension of the Artificial Intelligence part of e-fanorona, we haven’t yet implemented it.
- Implement and optimize the Artificial Intelligence: we mainly use decision-tree algorithm so as decide on the next move. Weight of each section of the decision tree are defined according to variables ad formulas that we set based on two elements:
Territory : is the number of positions that can been controlled by a pawn.
Liberty : is the number of positions where a pawn can moove safely