One Game to Rule Them All
An exploratory analysis of board games listed and rated in BoardGameGeek.
“What makes a great board game?”
Casual and experienced board game players worldwide have varying opinions when it comes to this question, even me.
I’ve always had a passion for board games, but it grew exponentially during the first lockdown. I was lucky enough to live with 3 other people, so we tackled our stress and boredom with plenty board games. We expanded our collection through the months and played together almost every night; today we have now amassed more than 10 board games that we consistently play together!
As we collected more games, we asked each other the same question: While some of us considered effective use of strategic play to make a great board game, others considered high inter-player activity, multiple game mechanics or reasonable game time to be the most important features of the game.
This discussion became a great opportunity for me to develop my data science skills, particularly in exploring the features of highly rated board games with data. Luckily, I came across a dataset published by Gabriele Baldassarre, available on his Kaggle.
For this exploratory analysis, I am following the CRISP-DM process and using the pandas libraries in python. You can access the Jupyter notebook I created in this Git repository to follow my analysis in more detail.
Introduction to the Data: The Reign of Board Games
Before I explored specific questions, I wanted to understand the dataset and what it has to offer. This dataset contains the details, attributes, and measures for around 94,000 board games and expansions from BoardGameGeek, some dating as early as the 1960s!
Throughout the set, we see an exponential growth in both the yearly number of published board games as well as a corresponding increase in the average rating of those games. This can be linked to a movement known as the Eurogame Revolution, where designer games, family games and strategy games took form in the early 1980s in Germany, quickly followed by Europe and the rest of the world.
Eurogames focused their marketing and design on creative game mechanics that fit the theme of the particular game. The variety of gameplay styles along with popular themes such as medieval fantasy, World War 2, and Lovecraftian horror were marketed to the masses, hence attracting a dedicated fanbase.
It looks like most board game designers are well informed on their consumers, and generally receive at least a 6 out of 10 in their reviews. This brings me to my first question:
What Correlates to a Board Game’s Rating?
After cleaning the data through imputing missing values and filtering biased values, I decided to investigate a correlation heatmap for this dataset. At first glance we immediately see a relatively high correlation of 0.56 between game complexity and the average rating of a game.
Complexity usually means that the game is rich in mechanics or lengthy in play-time, but that doesn’t always mean the game is considered difficult. On the other hand, game complexity almost always guarantees replayability; this is prominent in board games like 7 Wonders, Twilight Struggle, and Articulate.
We also see that there is a correlation of 0.36 between game complexity and minimum age to play. This comes as no surprise, as a more complex game incurs a higher barrier of entry. This also supports the idea that the minimum age to play positively affects the average rating of a board game, as seen on the correlation heatmap. Users on BoardGameGeek tend to rate complex games very well.
How Does a Game’s Complexity Affect its Rating?
Coming upon this interesting correlation, I wanted to examine the distribution of game complexity in the dataset, and consecutively, how the average game rating changes.
The following shows a scatter plot that answers both of those questions:
There are two clear observations here. First, most of the board games listed in BoardGameGeek are considered simple, with most of the dots plotted between a range of 1 and 3 in game complexity. These include games meant for large groups/parties like Cards Against Humanity, family games like Game of Life, or short duration games like Exploding Kittens.
However, the range of ratings for these games is very wide. We see that a board game with a lower barrier of entry to play or one that is simple in gameplay does not necessarily make it great; the average rating for these groups lies at approximately 5 out of 10.
The second observation is that while the total number of board games listed decreases as their respective game complexity increases, the quality of these games converges to a higher average. Complex board games such as Scythe and Villainous are designed with attention to detail; ensuring gameplay is flexible, replayable, and creative for board game enthusiasts to thoroughly enjoy. This takes up a lot of effort on the publisher’s part, but it clearly pays off with their board games being highly rated on BoardGameGeek.
Is a Complex Game Considered Popular?
From the results above, I wondered if there was a distinct trend on complexity and popularity. I created a new calculation based on the average rating of a board game and the total ratings that game received. Once again, a scatter plot is the best fit in showcasing the relationship between game complexity and game popularity:
The verdict, is vague, as the outliers with massive popularity scores provide little information to the trend. However, we can see a very slow descent in popularity as the complexity of the game increases. This could be supported by the fact that most complex games are highly rated in niche markets, particularly consisting of dedicated board game collectors and players.
Ultimately, there seems to be no distinct correlation between the two, and these two factors should be considered independent towards the making of a great board game.
Conclusion
We have seen how game complexity, game popularity, and other details like minimum age affect the quality of a board game. However, there are many more factors that should be taken into consideration, including the types of game mechanics and the theme of the game.
The analysis above is observational, not the result of a formal study. So, the big question remains: