I previously mentioned my struggles with looking for a new mobile game, and it appears the team at Second Dinner heard my prayers. Marvel Snap, which launched this week on iOS, Android, and Steam, is a fun and "snappy" collectible card game that's scratched the "play on the bus" itch I was looking for.

Developers who've played Blizzard Entertainment's Hearthstone will recognize many of that game's design principles in this Marvel-themed free-to-play game. Cards smash into the board with a satisfying "crunch," there's a tiered energy system that raises the tension as the game progresses, and there's a healthy amount of CCG synergy that encourages players to experiment with niche heroes and villains from Marvel history.

After assembling a deck, players go head-to-head in match-ups where they aren't trying to destroy each other's heroes, but rather controlling at least two out of three lanes (called "Locations" in Marvel Snap). Locations also contain game-twisting bonus powers that can either benefit both players or reward the player controlling the Location.

That design gives the game a nice bit of narrative framework (the player isn't a superhero themselves, but rather a superhero team leader sending their favorite characters into battle), but there's a more subtle mechanic at play that brings the whole experience together: players take their turns simultaneously, rather than one at a time.

Why is that so thrilling? It's time for some game theory.

Marvel Snap's simultaneous turns make for short and sweet games

Okay, actually we need to put a button in the game theory. First, let's do some math.

Second Dinner was founded by ex-Hearthstone developers like Ben Brode, and in their previous game, matches could last a lengthy amount of time.

Turns in Hearthstone max out in 75 seconds, and player-gathered data shows us that as of February 2022, the average game lasts about nine rounds.

If Hearthstone players make quick decisions, (let's say, they take 15 seconds per turn), you might have rounds that last 30 seconds, making for an average game of 270 seconds. Games like that will only run about four minutes and 30 seconds.

But what if players use their full 75 seconds each turn? That happens sometimes for organic reasons (players are in touch matchups and want to think carefully) and inorganic ones (players are trolls). A 9-round Hearthstone game could then run as long as 22 minutes and 30 seconds.

The purple button at the bottom right doubles as a timer.

Those don't happen often, but I'm very sure some Hearthstone games I played back around 2014-2015 were at least 15-20 minutes. Sometimes they felt like great, hard-drawn duels; other times I was being ground down by trolls.

There's an obvious multiplication problem here that causes the potential match length to run so long: each round's length is detiermined by the combined amount of time each player spends on their hand. If one player rushes and the other acts slowly, one of them might feel irritated at the other's pace.

In Marvel Snap, players select their cards in the same round which maxes out at 35-40 seconds (I was trying to juggle a stopwatch and playing a card, it wasn't precise).

With games locked at six rounds (short of Hearthstone's open-ended average of nine), every game has a fixed maximum length: about three to four minutes.

Right away, you can see the difference: Marvel Snap rounds are short and fit very easily into someone's lunch window, bus ride, or commercial break. But round length isn't enough to guarantee fun or player interest. What do the simultaneous turns do to make the game so compelling?

Now it's time for some game theory

All turn-based multiplayer games rely on some level of projection or patience for what your opponent is going to do. What Marvel Snap illustrates is that there's a huge difference in how you try to make those guesses when you and your opponent play your cards at the same time.

In Hearthstone (or Yu-Gi-Oh, Magic the Gathering, the Pokémon Collectible Card Game), players take individual turns and can make assessments based on the choices their opponent makes in an isolated environment.

There's still strength in this kind of design! When games are structured with individual turns, CCG players begin to pick up the same skills as poker players. They look for tells, watch how their opponent reacts to their own decisions, and observe the minute choices-within-choices that can define a player's turn.

But when players play their cards at the same time, the only information each of them has is the cards that were played the previous round (particularly true in a digital environment. Second Dinner also doesn't rely on the cursor arrows that appeared in Hearthstone).

There's plenty of information in those cards. And with Marvel Snap's three-Location structure, players are also being asked to guess not just what kinds of cards their opponent will play, but where they'll play them. Early cards in the games' starter deck actively encourage new players to start making that judgment call, and it's a call that's easier or harder to make based on what's been played.

For instance, the starting Hawkeye card gains more power if its owner plays another card in the same Location on the following turn. If I see my opponent play Hawkeye, I know I can safely play a card like Star Lord, which gains power if my opponent plays a card in the same Location in the same turn.

But what if my opponent is willing to give up Hawkeye's bonus power in a digital head fake? What if their strategy was built on another card that could act alone in another Location? What if the Location we were placing our cards in is revealed to have a secret power that nullifies someone's strategy? These choices are all amplified by Marvel Snap's simultaneous play.

You can find examples of this kind of design all over the board game world, and it's already influenced video games through games like Teamfight Tactics or Auto Chess. But Second Dinner's introduction of the mechanic to the digital CCG market brings freshness to the genre.

This mechanic also feels like it does more for user retention than other gimmicksbeing trotted out at the moment. Do players really want to truly own digital cards? Or are they more interested in quick rounds that tickle their brain and give them a dopamine hit?*

It might be both! Maybe there's a peanut butter and jelly sandwich of a game out there waiting to happen. But for now, I want to applaud the team at Second Dinner for making something so fresh in a crowded market. It's not just the Marvel brand that makes this game so compelling—it's a good set of design and programming principles that should inspire other developers.

*Yeah yeah, we can't get out of here without acknowledging the occasionally shady ethics of dopamine hits in video games. The always-excellent Celia Hodent has spoken at length about this subject.

Update: This story previously misspelled the name of Second Dinner co-founder Ben Brode. It has been updated to correct the error.