Poker has no single inventor. It evolved from European card-playing traditions and is now played at casino tables worldwide and on digital platforms like PokerStars and GGPoker, primarily targeting adults 18 and older. The session I played was Texas Hold’em with a group of friends using poker chips. In this Critical Play, I argue that Poker puts players at risk for addiction not because of pure randomness but because it convinces you that your losses are a skill problem. Unlike roulette, where a loss is just a loss, every poker hand gives you a reason to stay: you read the opponent wrong, you called when you should have folded, you bet too small. The randomness provides the losses; the skill element provides the rationalization to keep going. That loop is what makes poker one of the most persistently compelling gambling games ever designed.
The hand that made this clearest came near the end of the session. The community cards were 5-7-4-3-8, and I held A-Q off suit (Figure 1). A-Q is a premium starting hand, in the top five percent of hole card combinations, so I had played it confidently from the beginning. But the board never connected: no clubs fell for a flush draw, no straight was possible, and neither an ace nor a queen arrived to give me a pair. Two other players held 5-3 for two pair and 7-6 for a made straight. I shouldn’t have been in that hand by the end. But I had stayed, recalculating whether my overcards were live and whether a well-timed bet could win. When the cards flipped, my first thought was not “I was unlucky.” It was: “I should have pushed harder on the turn.” That instinct, to find the decision that explains the loss, is exactly the mechanism Natasha Schull points to when she said that “almost hitting the jackpot increases the probability that the individual will play the machine, although this reinforcer costs the owner of the device nothing.” In poker, the near miss is not a visual trick programmed into a chip. It is built into every hand. A bad loss is always reconstructible as a near win, and that reconstruction compels another hand.
Figure 1. The losing hand: board of 5-7-4-3-8 against my A-Q. Two opponents held 5-3 (two pair) and 7-6 (straight). The hand looked winnable until it wasn’t.
This is what separates poker from other games at similar positions on the chance-skill spectrum. In class, we placed games along an axis from pure skill (chess) to pure chance (Candy Land), with poker sitting closer to the chance end but still carrying meaningful skill. Slay the Spire sits nearby on that spectrum, where deckbuilding matters but the draws do not. The critical design difference is containment: a Slay the Spire run ends, you die, you restart, you learn. Poker never ends on its own terms. One player at the table, having lost most of his chips, said flatly, “I’m not leaving until I at least break even.” That statement would make no sense in Candy Land, where there is nothing to improve on, and is structurally impossible in chess. It only makes sense in a game where the skill layer is real enough that improvement feels possible, but the variance is high enough to keep that improvement perpetually out of reach. The game gives you just enough control to make quitting feel premature.
Compared to slot machines, poker looks transparent. In poker you can calculate pot odds, study hand ranges, assign probabilities to opponent holdings. The deck is not hiding anything from you. But this transparency does not protect players; it amplifies their exposure. When the odds are visible and you still lose, the natural conclusion is that you made an error, not that the game beat you. Roulette does not offer this. Candy Land does not offer this. The more skill a game contains, the more a loss can be attributed to something correctable, and the harder it becomes to stop. Teamfight Tactics, which also sits in the chance-skill middle range, sidesteps some of this by placing losses inside a matchmaking queue: the match ends, you return to a lobby, you make an active choice to queue again. Poker has no lobby. There is no designed moment where the game signals that the session is over.
Chance in game design is morally permissible when it generates meaningful tension within a bounded experience where financial harm is not a real outcome. Poker fails on both counts. The randomness is real, but paired with financial stakes, the skill element actively gives players reasons to stay past rational play. The problem is not randomness itself: Slay the Spire uses it responsibly, because each run is finite and nothing real is on the line. The moral line is whether chance creates meaningful decisions or sustains a loop that exploits the player’s belief in their own improvement. A designer who adds chance to a financially consequential game with no session limits, no loss caps, and no designed exit has made an ethical choice, even if it’s a passive one.
