Playstack discovered Balatro because one person—head of discovery Patrick Johnson—manually scrolled through every new Steam listing daily. He spotted the game the day its page went live, around May or June 2023. LocalThunk, the solo developer, had roughly two or three followers on Twitter at the time. No algorithm flagged it. No viral moment triggered it. A human being doing repetitive grunt work found what became PC Gamer's 2024 Game of the Year.
What Actually Happened
At a talk highlighted in the 2026 GDC Trends Report, Johnson laid out the process plainly. It's boring on purpose. "One thing that we do at Playstack is we scout," he said. "We look at games across all sorts of platforms. Part of my job is looking at every game that goes up on Steam every day, if that's possible."
He starts in the morning. Scrolls through every newly listed game. Does this daily.
That's how Balatro surfaced. "I saw the game the day it went up on Steam," Johnson explained. He then went to Twitter and contacted LocalThunk directly.
Let's put a fine point on what made this unusual: Playstack was still actively using Steam wishlists as a primary indicator of commercial potential. Balatro had none of that signal. It was a complete unknown from a developer with essentially zero audience. The publisher's own heuristic would have told them to pass. They didn't.
The entity → mechanism → outcome chain here is clear: Playstack's discovery infrastructure (entity) relied on manual daily Steam scouting by a named individual (mechanism), which produced early contact with a solo dev before any market signal existed (outcome). This is the part other publishers struggle to replicate, because it requires paying someone to do unglamorous work with no guaranteed hits.
Why This Matters Beyond Balatro
Why didn't algorithmic discovery catch Balatro early?
Algorithms optimize for existing momentum. They surface games that already have wishlists, followers, click-through rates, or social proof. Balatro had none of those when its Steam page went live. The game's core hook—a roguelike deckbuilder built around poker hands and scoring multipliers—doesn't immediately signal "genre-defining hit" on paper. It signals "niche card game from an unknown."
Steam's own recommendation systems, which drive significant discovery for established titles, are essentially blind at the moment of first listing. There's no behavioral data yet. No engagement history. The algorithm sees a new page with zero wishlists from a developer with no track record.
This is why Johnson's manual process matters. Patrick Johnson (entity) bypassed the algorithmic cold-start problem (mechanism) through systematic human evaluation of every new listing (outcome). The mechanism isn't scalable in the way tech platforms prefer, which is exactly why it works for catching what algorithms miss.
What does this say about indie game discovery right now?
The uncomfortable truth: the most reliable early discovery system for indie games in 2023–2024 was one guy scrolling through Steam pages every morning. Not a recommendation engine. Not a content creator amplification pipeline. Not a press outlet. Manual scouting.
This isn't unique to Playstack. Smaller publishers and indie-focused scouts have operated this way for years. What's different here is that Playstack's approach actually connected to a publishing deal, and the game in question went on to win major awards including recognition from BAFTA and multiple outlet Game of the Year picks.
For developers, the implication is straightforward but cold comfort: your Steam page needs to communicate its hook immediately, because the person deciding whether to reach out might be scanning hundreds of listings that day. Balatro's page did that effectively enough for Johnson to stop scrolling and make contact.
The Detail Nobody's Talking About: LocalThunk's Position
Johnson noted that when he contacted LocalThunk, the developer "had maybe two or three followers" on Twitter. This isn't a rounding error. It means there was effectively no community, no pre-existing audience, no influencer network ready to amplify the game.
LocalThunk (entity) had no platform or audience base at the time of contact (mechanism), which means Playstack's decision to pursue the game was based entirely on the game itself—not the developer's reach, track record, or social proof.
This runs counter to how most publishing decisions get made. Even indie-focused publishers lean on signals: has the developer shipped before, do they have a following, are they plugged into any communities that could generate early wishlists? All of those signals were absent. Playstack looked at the game, not the metrics around the game.
(Whether this is scalable as a publishing strategy is a different question. It worked here. That's one data point, not a blueprint.)
What's Still Unknown
Several pieces of this story remain unaddressed in public reporting:
- What exactly did Balatro's Steam page look like in May/June 2023? The page that caught Johnson's eye—the specific trailer, screenshots, description—isn't documented in the available sources. Knowing what "worked" on that day would be genuinely useful for developers, but that artifact isn't publicly accessible.
- What happened between first contact and the signing of the deal? We know Johnson reached out. We know Playstack ended up publishing the game. The negotiation timeline, deal terms, and what other publishers (if any) were also in contact remain undisclosed.
- How many other games has this scouting process produced hits for? Johnson mentioned "games like Balatro," plural. The other success cases from this specific daily scouting method haven't been detailed. One flagship success is notable. A pattern would be more instructive.
- Does Playstack still maintain this scouting method at the same scale? If the process is dependent on one person's daily routine, what happens when that person is unavailable, or when the volume of new Steam listings exceeds what one person can reasonably evaluate?
What to Watch Next
If this scouting model interests you—whether you're a developer hoping to get found or a publisher evaluating your own discovery process—here's what's worth tracking:
Playstack's next moves under new ownership. According to related reporting, Playstack could soon belong to the same private equity firm that owns Fandom and GameSpot. Whether a PE-owned Playstack maintains labor-intensive manual scouting, or shifts toward more "efficient" (algorithm-reliant) discovery, will tell you whether this approach has institutional staying power or was a product of specific people at a specific moment.
Whether other publishers adopt formalized scouting roles. Right now, "head of discovery who manually scans Steam daily" is not a standard job title. If that changes—if more publishers create dedicated scouting positions after seeing the Balatro outcome—the competitive dynamics of early indie discovery shift significantly.
Steam's own discovery improvements. Valve has iterated on recommendation systems, curator tools, and discovery queues. If Steam improves its ability to surface zero-wishlist games that have strong page-level hooks, the value of manual scouting diminishes. If it doesn't—or if volume continues to outpace algorithmic curation—human scouts remain the primary discovery path for unknown developers.
LocalThunk's next project. The next game from the Balatro developer will be a test of whether the initial publishing deal produces an ongoing relationship with Playstack, and whether the developer's now-established reputation changes how discovery works for them. A known quantity doesn't need to be scouted the same way.
The Hard-Stop Verdict
The Balatro discovery story isn't a parable about believing in indie games or trusting your instincts. It's a story about operational grunt work that happened to connect to an exceptional game. The lesson isn't "publishers should take more risks." It's "someone has to actually look at every new game, and most publishers don't want to pay for that."
Playstack did. It paid off once, spectacularly. Whether that's reproducible—by Playstack or anyone else—is genuinely uncertain.
