03 November 2017
|
Board Game Finder uses machine learning to suggest recommendations
With so many games out there, it can often be a bit overwhelming knowing what game you should try next.
One team trying to solve this problem has come up with Board Game Finder, a search engine that allows players to put in the games and gameplay mechanics they already enjoy and get a list of recommendations based on their preferences.
Put in ‘worker placement’, for example, and you’ll be pointed towards Caverna, Agricola and Lords of Waterdeep. But if you specify that you’re after a card game, it’s Targi and 51st State at the top of the list.
Each game has a Netflix-style match rating based on your interests, as well as featuring its BoardGameGeek rating and general information – number of players, time to play and so on.
It’s an interesting concept, and the tech behind it is surprisingly complex. The recommendations are produced by a combination of looking at which games certain BoardGameGeek users have liked – so if you like Catan, you might like Pandemic and Dixit, for instance – with a form of artificial intelligence called machine learning, which aims to fill in gaps in players’ rating history by analysing data to predict what they might think of particular games based on their feelings about games they have rated.
It’s not always perfect – one of the highest-ranked recommendations for fans of Codenames is Terraforming Mars, which might be a great heavy strategy game but couldn’t be much more different from the light word game – but it’s still very interesting putting in different combinations and seeing what gets spit back out. And, hey, any way for more people to discover more great games is a-okay in our book.
Comments
Login or register to add a comment
No comments