We all save scummed in Baldur’s Gate 3—except, apparently, for my friend Jack and coworker Morgan Park. They both accidentally killed everyone at the Last Light Inn and just kept trucking. As part of Larian’s big Divinity AMA on Reddit, the game’s writing director, Adam Smith, revealed how the studio is factoring players’ save scumming tendencies into its next game’s design.
For a quick definition, I think “save scumming” is best understood as using a game’s save/load feature for anything other than quitting the game and coming back later, or continuing from a game over—no moral judgement here, mind you, just an observation of our tendencies. The incentive to save scum could be argued to be a design failure, or merely a natural function of games as wish fulfillment. We already have to live with our mistakes in real life, why do so in our fantasies?
The big Baldur’s Gate 3 example has to be quickloading every time you fail a skill check until the 20-sided dice finally lands your way. But for RPGs that are all about choice and consequence, this ever present temptation to never have to live with failure can adulterate some of a game’s chew, removing stakes from the story and preventing you from seeing interesting plot developments from failure states.
“Are there any plans to try to make failing these dialogue/non-combat checks more interesting, a la Disco Elysium,” asked user Nidies in the AMA, “to try to encourage players to accept ‘bad’ [random number generation]?”
“Our ambition is certainly to make failure more interesting,” replied Smith. “There are already a couple of situations in the game where I think the most exciting—and extensive—outcome comes from ‘failure’, but you’ll be the judge of how well we do.”
Early in 2024, I corresponded with Baldur’s Gate 3 lead systems designer Nick Pechenin about this topic for a story in PC Gamer’s print magazine—Pechinin is also working on Divinity, and answered questions elsewhere in the AMA. He was of the opinion that save scumming is a perfectly reasonable player behavior, and bristled against the tendency among some RPG fans to disparage it.
If a designer wants players to accept failure states, Pechenin argued, those failure states need to be interesting and worthwhile, which lines up with Smith’s statement in the AMA. For an example of a game nailing this, Pechenin cited a long Crusader Kings 3 run that fell apart over a small miscalculation—it was such an interesting and surprising development, he found it a fitting climax to the game.
Pechenin also pointed to quick rewind features like you see in racing games, Prince of Persia, or the more recent Shadow Gambit: The Cursed Crew as a way to better integrate the quicksave-quickload impulse into a game. That’s far from a guarantee that Pechenin or anyone else at Larian will act on this admiration in Divinity, but a one turn mulligan button strikes me as a potentially strong and versatile addition to Larian’s house style of combat.
In a similar vein, Pechenin seemed open to “static checks” as a way to avoid save scumming to redo dice rolls. Rather than augmenting a random chance with your investment in a skill, a static check simply requires a minimum investment to pass, and if you don’t have enough points, too bad. Fallout: New Vegas is a notable RPG built on these sorts of skill checks, and this is even how Divinity: Original Sin 2 handled persuasion, though its non-combat options were far more limited than Baldur’s Gate 3’s.
“We had long internal discussions and even prototypes for making some of the checks [in BG3] more static,” Pechenin revealed. “This is usually relevant to physical checks that thematically can be attempted again and again—pushing a stuck door open for example.
“In the end we did not have enough relevant cases to introduce yet another system to this chonker of a game, so we went all in on the big shiny D20 and made sure that it feels as good as it can.”
Elsewhere in the AMA, Larian further elaborated on its AI use in Divinity, committing to not using it in concept art, but retaining the option to use it elsewhere in development. The studio will also not be reprising Divinity: Original Sin 2’s armor system (but I strongly disagree with Rory about it having been “broken”—it was good!). Also? No WASD controls for you.
