'No matter what we changed, somebody was going to get mad:' After hearing what an Obsidian dev went through to get RPG difficulty right, I get why nobody's ever gotten RPG difficulty right

In a talk at the Game Developers Conference, Obsidian senior systems designer Robert Donovan spoke about dialing in the difficulty of The Outer Worlds 2⁠—a great game you ought to play⁠⁠—but one I found to have a particularly noticable reverse difficulty curve.

Instead of the challenges ramping up as you get more familiar with the game, RPGs tend to start out hard and get easier. They're fantasies of growth and accumulation⁠—number go up⁠—which runs counter to a game getting harder as it goes on, and I don't think I've ever played one where the challenges keep up with how strong you get.

The Dreadlord of the Doomkeep is just never as tough to beat as a level one goblin with an iron dagger that can kill you in one hit. It's not just a TOW2 problem⁠—you can see it in Fallout, Baldur's Gate 3, even talkies like Disco Elysium.

It shouldn't be surprising that people who make RPGs for a living are thinking about this even harder than we are as players. Donovan's talk focused on how he approached hit points and damage while making The Outer Worlds 2, and how that translated into difficulty.

"My game director, Brandon Adler, said that we wanted to maintain the feeling of Emerald Veil throughout the entire Outer Worlds 2 game," said Donovan. Emerald Veil was the opening area of the first Outer Worlds, and the team thought it stuck the landing in terms of difficulty and balance. "Weak things are weak, and strong things are strong," as Donovan put it.

When first designing The Outer Worlds 2, Obsidian mapped tiers of enemy health and damage directly onto a curve pulled from Emerald Veil, with an initial target of a more "flat" character progression: Instead of having your guy be exponentially more powerful at level 30, like they might be in D&D, you'd be twice as strong numerically.

But that "crushed" progression just didn't feel good. "When you play the game and you pick your upgrades, getting a 1.5% upgrade didn't feel meaningful to the player," said Donovan. "We wanted the choices to hit harder than some other competitors in the genre, and so what we wanted was something like a 10% boost so as soon as you hit that level, as soon as you pick that skill, you felt it immediately." And this really did translate to the final game: One of my favorite parts of TOW2 is its allowance of varied, weird, and overpowered builds.

Testers consistently found TOW2 too easy, eventually leading Donovan to a "hits to kill" system for determining enemy strength: Hit points can be so abstracted as to be meaningless, so Obsidian instead focused on making sure the number of shots it took to kill an enemy with a set weapon on Normal difficulty made sense and felt fair.

"Virtually every player is going to play the game on Normal," said Donovan. "There are some real sickos out there that want to start on Very Hard, but unfortunately, there are not very many. So we knew that getting normal right was our priority one.

"Each of our difficulties, from Story to Hard and Very Hard has some straight multipliers on top of the balance that we had already done. It would be nice to have a more robust difficulty set of options, a suite of cool stuff: More enemies, fewer enemies. But ultimately, we felt like if we got the balance right, we could just multiply it up and down and succeed at our goals."

Not an unreasonable strategy by any means: Doom designer John Romero has said that he personally tested every level he made on Ultra Violence, the second-highest difficulty of five. One thing that struck me was the diversity of players Obsidian had to account for: There are going to be story-focused or less hardcore players who could bounce off TOW2 if it demanded too much optimization from them. On the flip side, what about the buildcrafting freaks?

"We thought if we got it right for the intended player, then it would be close with players who really didn't want to engage with the systems, or for players who were super engaged with the systems," said Donovan. "It turns out, players who are super engaged with the systems are absolutely going to blow your balance out of the water as soon as possible.

"And again, there's not much you can do about that without hurting the game for everybody else. You've got to let the super optimizers run wild and hope that they crank up the difficulty on their next playthrough."

In one playtest, equal percentages of respondents found the game too hard, too easy, and just right. "No matter what we changed, somebody was going to get mad," said Donovan. "But at least we were close. It felt like we were close."

That division seemed to hold for post-launch reviewers: Donovan quoted reviews that found TOW2 tediously difficult, way too easy, and even just right. Personally, I came down somewhere between the last two: It started out just right on Normal, but by the end of the game I was cranking things up and still finding it a bit too smooth.

"We found that the final boss and that sort of capstone to the whole game, the big climactic finale, was just a little bit too easy, so we put our thumb on the scale and we made it harder," recalled Donovan. "Arguably, we didn't make it hard enough, but that's probably a talk for another time."

I wasn't able to find out either way in my review playthrough: In classic Obsidian fashion, I was able to talk the final boss into killing himself, one of the hallmarks of a Good RPG™. So maybe the real difficulty curve was the friends we made along the way⁠—TOW2 still turned out a great game, with my main mechanical complaint being something it shares with the best of the best.

"Perhaps we did not quite nail that fun in the end, but someone on Reddit said they were having a great time," Donovan quipped. "I guess we're gonna call it a W."