Corey Davis, design director at Psyonix, has spent close to a decade making the same game, twice. In 2008 his studio released Supersonic Acrobatic Rocket-Powered Battle Cars. Then they spent eight years polishing the game. At the Game Developer’s Conference in San Francisco today, he said that the end result, Rocket League, is “the game we wanted the first game to be.” In the middle of all that, though?
In the middle, Rocket League was a lot of really weird stuff.
Most people, he says, only hear about SARPBC in articles– like this one!– about Rocket League. It got a 67 on Metacritic. “We weren’t sure what went wrong… we weren’t really sure what we were doing with our first independent game,” he says. But he game was really “sticky” for the people who got it. It was one of the first PSN games with native Youtube export, and its fans made so many video montages of it and sent in so much positive fan mail that Psyonix knew that had something there– some kind of really special game hidden beneath the surface.
Their first attempt to remake SARPBC was kind of a trainwreck. It was called World Battle-Car League, and they put a huge emphasis on trying to differentiate it visually from the original game. They knew they wanted to make an esport, and “we thought to be this serious sport, it needed to be clear that these were awesome human-sized vehicles.” They spent ages experimenting with the scale of the cars, trying to make them look and feel like giant monster-trucks, and were pretty scared that people would mistake them– as they’d done in SARPBC, apparently– for toy RC cars.
But this bad focus led to a lot of directionless experimentation. “The fact that we were changing everything from scale, to physics models, to how cars worked all the time” meant that they were never able to make a good testable build of the game, or polish the way the cars moved and how the physics worked.
“So we did what every developer did earlier in this century and started making it open world,” he laughed. He showed a video of a Rocket League-style car flipping and bouncing down a snowy mountain slope, dodging large, cartoonish trees and rocks. “So if you want to drive to a stadium and play soccer, that’s one mode available to you,” he said.
At the time, it was “insanely insanely out of scope for our studio. At the time we were only 15 guys.” And the most important part of the game– the car physics– was still out of whack. “The split focus made it impossible to make any part of the game good. We were serving too many masters at once.”
So they scrapped open-world Rocket League and re-assessed their goals, which were
- Dedicated servers,
- Getting the game to a PC alpha build ASAP, so they could polish the physics, and
- Having it be free-to-play.
Rocket League was apparently F2P for quite a long time– the PC alpha had an elaborate crafting system– but they eventually dropped that plan, too. Spending so much time trying to figure out F2P systems made them feel less like game designers. They also did the math and discovered that to break even, they’d have to make two million F2P item sales. With a $20 game, however, they’d only have to sell 50,000 complete copies. Selling the game for straight-up moneydollars let them break even with a much smaller fanbase.
Little did they know it, but they’d have a gigantic fanbase, and that wouldn’t even be a problem. But Davis pointed out a lot small decisions and lucky surprises they’d had which made it easier for Rocket League to become the breakout success it did. One was their decision to keep the game very pure and simple. He said that a lot of players would ask, “Why don’t the cars have stats? I want a car that’s better at playing goalie or a car that’s better at striking.” But because roles can change in Rocket League so quickly, they decided to leave those mechanics out. But the real decision-maker? The fact that they didn’t have enough time to polish a stats system.
This same rush for time led them to cut a lot of other features which, in retrospect, it was good for them to leave out. They were worried that people wouldn’t want to pay $20 for one game mode on one map, but they didn’t have time to polish any of the other sports modes. But “at the point of release, we’d been working on the soccer mode for seven years!”
Keeping the release version to basically one kind of map also had an unexpected benefit. That very pure map design had been polished so thoroughly for so long, “every angle and corner and surfaces encourages fun.” The lack of other options frustrated their SARPBC veterans, but that was “a downside we had to accept.”
Davis then took some time to talk about how physics in Rocket League isn’t “real physics,” but a highly-tuned “fake physics” also designed to maximize fun. “Our car and ball interaction are actually completely fake,” he said. Real physics are too random and unpredictable to be fun. “Predictability wins off here,” he said. “It’s important that when you touch the ball from a current angle it’s going to bounce fairly consistently.” They had old versions of the game where cars had extremely accurate hitboxes and real ball physics, but they discovered that it was actually a lot less fun to wrangle real physics than to simply use very ordinary box collision and fine-tune it for the purposes of game balance.
As the game evolved into what Rocket League is today, some of the old SARPBC fans were not pleased with the extremely simple direction the game was taking, or with the changes to the core mechanics. One angry fan had this to say about the game:
Ironically, this player actually went on to play Rocket League anyway, and has actually won tournaments. In general, these older players were a huge benefit to Psyonix, since they were already comfortable with the (admittedly very wierd) idea of “car soccer,” and were good enough to show off their skill on Twitch during the beta period and its free PSN launch. “Veteran players gave us a built-in pro scene,” Davis said. They made it obvious to newbies that it was possible to actually get good at the game. “Aspirationally, that’s amazing. ‘I want to get that good,’ it set a really high bar for what as possible in the game,” he said.
Davis listed a lot of other reasons why he thinks the final version of Rocket League was successful. It made really good Twitch material! It is easy to learn but hard to master! It was lucky to get press attention, and dominated games news coverage for days! But Davis warned that “random luck can look a lot like genius in retrospect.” He reminded the audience– mostly developers themselves– that the advice to “just make a really good game, and you’ll succeed– that’s not true! We were really lucky!”
And we’re lucky, too, that we’re not playing downkill-skiing open-world Rocket League, or monster-truck Rocket League, or Extremely Realistic Physics Free To Play Rocket League. I’m not gonna say we’re living in the best of all possible worlds, but we may be playing the best of all possible Rocket Leagues.