Growing up, Kevin Binkley had always gravitated toward the same idea, and that was… could a computer ever play card games well, as in, you’d feel like you were sitting right across from a worthy human opponent?
“In high school and college I played a lot of bridge,” he likes to say. “I’ve always liked card and board games.”
And that interest turned into a career for him, rather early, in fact.
Flash back to the 1990s, and you’d have found Kevin and his friend group starting a company, called Electric Gravity. It focused on making online versions of classic games, including bridge, chess, hearts, spades, etc.
The timing was spot on too. This thing called the internet was catching on, and naturally, multiplayer online games (also a new contraption at the time) lent themselves very well to this newfangled technology.
Eventually, all of their work caught the eye of Microsoft, which ended up acquiring Electric Gravity. Even so, that same question “how do you make an AI play smarter?” still nagged at Kevin.
So he did something about it… he went back to school.
Kevin would go on to conduct research for a PhD in computer science, spending years studying artificial intelligence, including topics like neural networks and evolutionary computing, just to name a couple.
He had built a neural network for the board game Reversi (also known as Othello, a classic strategy board game) and then realized it could be a great fit for mobile phones. That’s because it didn’t require too much computation; just enough for an early Android phone to handle.
“A neural network would not require extensive deep search, thus saving computing power,” Kevin explains.
So he did what a lot of serial entrepreneurs do: he shipped it.
First, there was the app on Android under the name “NeuralPlay,” a project that ended up also giving him his business’ name.
The company, NeuralPlay LLC, was established in 2017, though some of its apps had been on Google Play a lot earlier. In fact, one of them was called NeuralPlay Hearts, first made available in December 2010.
Then, as time went on, the mission began to crystallize for Kevin. He was building practice partners for people who love trick taking games.
“NeuralPlay’s mission is to create engaging and intelligent computer AIs for four-player, trick taking card games,” he says, referring to card games where players play cards in rounds (or tricks), “offering flexible rule options so players can enjoy the game in the kind of way they’re accustomed to playing it.”
Flexible-rule-options would turn into an important signature, a very intentional design choice for Kevin, showing up everywhere across his apps.
Just look at NeuralPlay Hearts. It lets players assign point values to specific cards. That way, they can simulate the same rules they had followed growing up. He even calls out the Hooligan Hearts variant as an example, where the 7 of Clubs is worth seven points.
And that’s the kind of detail his audience appreciates, the very same people emailing him all the time. “I receive emails from users who play the classic card games in person at community centers or other clubs,” he writes.
And they’re not shy when it comes to giving feedback either. There are some who’ve been emailing him for more than five years now, submitting a bug report and making suggestions here and there. And Kevin is happy to oblige; he’s implemented many of their ideas.
By his estimates, NeuralPlay’s most downloaded apps include Bridge, Spades, and Twenty-Nine, each boasting over a million downloads.
