top of page


   Liubo is an ancient Chinese board game originally played from about 500 BCE to 500 CE. The rules of the game have been lost to time, but our team is attempting to reconstruct the game based on what we do know. Today you will be helping us test our digital version of the game.

Goals for the project

1. Reconstruct and playtest this “lost” ancient board game, based on available artifacts, texts, records, English/Chinese-language scholarship, consulting with experts, etc. – plus our team member’s own ideas about player experience and which version of the game seems the most fun.

2. Create a digital method (a game, an app, etc.) to teach contemporary audiences how to play our proposed reconstruction of Liubo. Also convey some of the history of the game and information about how it’s been reconstructed.

Summary of the game

Liubo is a game about birds catching fish. Each player controls 6 pieces representing birds. These birds fly around the board in order to reach the center square, which represents a pond full of fish. Birds catch fish by crossing over or landing on this pond, and then score points by dropping the fish in the circle “Nest” spaces on the opponent’s side of the board. However, you can also score points by capturing your opponent’s birds before they reach the nests! Liubo takes between 10 to 45 minutes to play; the first player to 6 points wins!

Design & Iterations

Path Design:

According to the ancient books, there are nine types of positions related to the board, but their meanings on the board are unclear. For convenience, We use "A" to "I" to refer to them.

Now, here's another quote from an ancient book that provided a formula for playing Liubo, which I have replaced with A to I for ease of reference.


According to this formula, we believe that positions A to I connect with each other to form a path. Connecting all four sets of A to I, we got a diagram of a path with four branches connecting the corners to the center.

Path Iteration:

We playtested this pattern of paths in the game, where the pieces move from the end of the branches to the center, and we found the problem was that sometimes it was difficult for the two sides to encounter each other, and when they do encounter, they often block each other, making it too easy to either defend, attack or even cause a deadlock.

To solve this issue, we drew inspiration from the contemporary game "Pachisi" and connected the branches to form loops, so when the two sides confront, it's easier to escape, reinforce, or attack with a siege. This greatly increased our diversity of strategy.

截屏2023-07-06 下午2.36.45.png
截屏2023-07-06 下午2.44.15.png
截屏2023-07-06 下午2.48.12.png
截屏2023-07-06 下午2.48.23.png

Helpers & Labels Design:

During playtests, we found the positions and paths were too hard to remember, as it's unlike any game people are familiar with.

So, we translated the nine positions into English and marked them on the board, which helped players understand the functions of specific positions. We also added a layer showing the connections between the spaces to help players plan for their target ahead of time.

截屏2023-07-06 下午2.58.17.png

Hard to remember sptial relationships

截屏2023-07-06 下午2.56.36.png
截屏2023-07-06 下午2.56.44.png

Final Result:

Use of Components Iteration:

Our remaining mechanics concern the physical components of the game and how they relate to one another. This includes rules for pieces capturing and blocking each other, as well as the use of the game’s throwing sticks.

截屏2023-07-06 下午3.06.11.png

We combined our observations from these two sources into paper playtests, through which we iterated and reached our final version.

截屏2023-07-06 下午2.33.48.png

Capturing Design & Iteration:

For example, when it came to capturing we started with quotes like this:

"My King… why don’t you use wisdom like using an Owl: Eat if it’s safe, stop if it’s not safe."

Quotes like this became the basis of a capturing system, in which pieces “ate” their opponents by landing on them. However, in many of our early attempts to design a capturing system, capturing was too easy, leading to games that were too quick.

截屏2023-07-06 下午3.17.03.png

Capture relationship draft that I drafted

It was around this time that we developed a more robust paper prototype. Instead of using slips of paper to represent the pieces, we now built them in 3D out of cardboard. This allowed more of their unique properties to stand out. We already knew that they were able to stand up on end, but now we discovered that they stacked nicely on top of each other as well. However, our current rules provided no opportunity for stacking.

截屏2023-07-06 下午3.20.21.png
截屏2023-07-06 下午3.20.32.png

Robust paper prototype

To solve our capturing problem and allow for stacking, we tried out a rule requiring two of your pieces to capture one opponent’s piece.

Another Capture relationship draft that I drafted

This rule slowed down the pace of capturing to a satisfying level, and it also created a new “contested” state in which opposing pieces stacked until a third piece enters to complete the capture. This rule change tested well.

截屏2023-07-06 下午3.25.43.png

Capturing Example

However, it also led to a new question–if opposing pieces are allowed to stack, what does it mean when two friendly pieces stack? Which seemed like a natural extension of this rule.

截屏2023-07-06 下午3.26.48.png

Blocking Design & Iteration:

This brought us back to an outstanding quote from our research that we hadn’t been able to incorporate yet:

“The owl is capable because one uses the normal piece to support it.”

We didn’t yet have a direct way for pieces to support one another. To solve this, we looked at one of Liubo’s related games, Pachisi.

截屏2023-07-06 下午3.31.56.png

In traditional pachisi, the shell or cone shape pieces stack by nesting on top of each other. When they do so, they create a blockade that prevents all other pieces from moving through them.

These rules were a natural fit for our game. Blockades provided a direct means for pieces to support one another, satisfying our research, and testers reported that they added a layer of defensive strategy to the game.



Dice/Sticks Design:

Finally, we had to make a decision on how to roll random numbers for our game. Our research pointed to a few alternative methods, including 4-sided throwing sticks, 2-sided throwing sticks, and an 18 sided die.

截屏2023-07-06 下午4.31.32.png

We tested each of these versions first using stand-ins such as dice and coins, and later using our own reproductions. We were looking for the method that created the best balance of strategy and luck.

The two-sided sticks were the stand out, because of the way they are counted.

The sticks are counted in two groups of three, differentiated by color. Each stick has two sides: an ornate rounded side that you can think of like “heads” on a coin, and a flat, plain side that you can think of as “tails.” Heads is worth 1, and tails is worth 0.

For example:

截屏2023-07-06 下午4.39.37.png

This is 2

截屏2023-07-06 下午4.39.46.png

a roll of 2 on the red sticks and 1 on the grey sticks

截屏2023-07-06 下午4.43.38.png

And the one special rule is that rolling all “tails” is actually worth 4 instead of 0.

Dice/Sticks Iterations:

截屏2023-07-06 下午4.46.23.png

The important part is that this method produces an uneven probability curve, while the other rolling methods produced a flat curve. This curve allows for much more interesting gameplay decisions, as players weigh the likelihood of events in the game, such as being able to evade capture or reach a nest. Playtesters of this version reported enjoying the balance of strategy and luck, whereas testers of the other versions were often frustrated by the level of randomness.

As a bonus, playtesters told us that rolling sticks gave our game an enjoyable and very unique physicality that we then worked hard to preserve in our digital game.





Playtests & Tournament
截屏2023-07-06 下午4.59.01.png

From our physical prototypes to our digital version, we have received valuable feedback both from ETC faculty and on playtest days. In addition, we have hosted tournaments and playtested with board game communities to observe how players can enhance their skills through repeated play. From our tournament champion’s feedback, he said the owls should be more powerful.

截屏2023-07-06 下午5.01.50.png

We found some strategies that players were using during playtests.

First one is set the trap, our tournament champion use this strategy to capture opponent’s owl instead of scoring in the nest. He used his normal pieces to set around of the pond and wait for opponent’s owls get out of the pond.

The second strategy is build a bridge before you move your owl out of the pond. As you can see, players use normal pieces to set a path for the owls can move on the top of the normal pieces step by step and each step will become a blockade which is safety enough for the owl to get to the nest. So use your normal pieces to protect your owls.


截屏2023-07-06 下午5.06.09.png

Here are some majority feedbacks that we received in this semester. Art is gorgeous, animations are good especially rolling sticks and the game process is relaxing and it feels like a chess win which means slow and tactical. However, some people said they unclear that order doesn’t matter when using 2 pieces to capture and worried about intricacy of the board.

​Also, I provided some playtest note here just for example:

Tutorial Design:

For the Tutorial part, we split our rules into 12 sections. Each section has multiple pages information and an animation for visualize the rules.

Originally, we plan to do an interactable tutorial into our digital game. Unfortunately, we are running out of time. Thus, we decided to change the tutorial style into Rule books which inspired from comic books. Even the rule book is not our original plan, it still successfully teach players how to play our game perfectly based on playtests.


We Also split our game features into few sections in order to design the tutorial easier and more intuitive. Here are the example of our tutorial work flow:

截屏2023-07-06 下午5.20.16.png

More details, please click the link

Week 1

We are working with our instructor to research in the history and mechanism of Liubo.

research brainstorm.jpg

Composition Box

Current Schedule:

  • Keep research about Liubo.

  • Set up weekly meeting times with team members.

What Went Well:​

  • After a brief meeting, other designers and I decided to implement the tutorials and player vs player first.

  • The next stage of the project is to use AI and machine learning to implement player against computer.

Issues For Support:

There are many versions/interpretations of the game available from different people and their own interpretations. Need to pick some and deisgn with my own ideas.

Week 2

Yicheng’s Prototype Time takes to play the game: 30 minutes Problem: rules too complicated.

Lochlan’s Prototype Time takes around 30 mintues Problem: once a player have an owl, the owl is too powerful and the player can easily win. (other designer on team)

Jeffery’s Prototype Time takes to play the game: 5 minutes Problem: all player avoid turning to owls first. (other designer on team)

Current Schedule:

  • Consider if we want to collect data from players.

  • Pieces' color: white and dark.

  • Sound of jade.

  • Playtest different versions of Liubo:

    • Lochlan's version Jan 23​

    • Yicheng's version Jan 25

    • Jeffery's version Jan 27

Project Summary:

  • Three of out team member each create their own version of Liubo and will playtest next week.

  • We are aiming to build a prototype of physical version of Liubo and write the rules by ¼ (Friday Feb 3).

What Went Well:​

  • First logo design was completed.

  • Lochlan, Yicheng and Jeffery’s first design completed.

some prototypes.jpg

Some physical prototype of the game

Issues For Support:

  • Need to order physical materials and some literature from taobao and amazon.

  • Need to improve the rules to make the game more fair and fun to play

My Prototype Update:

  1. Time takes to play the game: 30 minutes.

  2. Problem: rules too complicated.

my proto v1.png
Week 3

Current Schedule:

  • Release our game at the end of the Semester, but not sure what platform ( need to apply for license).

  • Start to combine different versions.

  • Iteration the prototype.

Project Summary:

Each of the physical prototype have a iteration and the goal for this week and next week is to combine the prototypes.

What Went Well:

  • Improvements on first version of prototypes are made.

  • First prototype of the combined version is made.

Issues For Support:

  • Need to play test.

  • Problems need to solve for the game

  1. Whether normal piece can go in Circle.

  2. Whether owl can turn around in the middle square.

  3. Whether normal piece can pass by middle square to get fish or only land.

  4. Whether combination of sticks make any difference.

Prototype Update:

Basic elements of the game

  • 2 player, each 6 pieces.

  • Catch fish in the middle.

  • Sticks

    • 6 sticks for rolling to get number.

    • Other sticks to count score.

  • Board with circle in the middle.

  • Owl go to circle, drop off fish at the corner.


Colored lines indicate how the player should move, it is not part of the board


  • Normal go to the middle 方, catch fish, become owl

    • Bring back to nest get 2 points.

    • Owl try go back to the original side.

  • Owl caught normal: Normal ones are gone forever.​

  • Owl caught owl: does not do anything.

  • Normal caught owl: owl being downgrade

    • Get 1 point for catching owl, become owl.​

    • Get 2 points when bring it back.

  • Normal caught normal: share space.​

Tech Update:

A stimulation has been made for Liubo. Right now the functions allow dropping a piece on any part of the board and it will automatically go to the closest red spot because that’s the correct location for the piece to be on.

prototype update 6bo.png

Art Update:

This is our model of Liubo pieces Maya’s Substances Painter and models of Liubo Pieces.

Below is the references of the pieces.

Rest of the blog details please go to the blog page.

bottom of page