Product1 distinct publisher3 min readPublished
Gambit borrows Chess.com's bots, review mode and Elo-style rating while ruling out real money entirely, which leaves the rating ladder carrying the whole job of motivating players and converting them to a paid plan.
The Product Desk · Product desk
Compiled by The Product DeskSomething wrong?How this is made
Martin turned out to be the tell. The Verge's writer logged into Gambit and found Chess.com's lowest-skill chess bot sitting at the bottom of the poker site's bot ladder too, doing the same job under the same name [6]. Around it are bot matches, games with friends, tournaments, a review mode for studying past hands, a learning mode, and the parent site's design language [5]. What crossed over was the furniture a Chess.com user already knows how to operate without reading anything, not poker expertise.
The money works the same way as the furniture. Chess.com's business is a subscription that most players never buy, because most people play free and pay only for unlimited game review, puzzles and lessons [9]. Gambit copies that model, with the rating system as its backbone [8]. Chief growth officer Albert Cheng told The Verge the rating is the main motivator to improve, since players are not risking their own money and something has to counterbalance the lower stakes [7]. He also ruled out real-money gambling outright, saying the whole industry makes money when beginners lose [2]. That is a coherent position, and it also means the ladder is load-bearing in a way it is not on a real-money site. Hands dealt would not settle it; what would is whether players who start reviewing their hands are still reviewing them in week six, and nothing public says.
Other operators should look twice at the build cost. Cheng and one other employee took Gambit from prototype to launch inside a year, using AI mainly for coding [11]. Roughly two person-years, then, for a platform [1]. The specific help he describes is unglamorous: AI read the existing codebase and design system and applied consistent patterns to a different platform [12]. It was no use for brand and feature design [13], and the live site has "zero or close to zero" AI-made assets, which were confined to early prototypes [14]. Without it, Cheng says, the company would have needed a bigger team or accepted more disruption to the work of building and growing Chess.com itself [15]. Go, backgammon and mahjong are next, each with its own Elo-style rating, on no announced date beyond development "moving quite rapidly" [10]. At Gambit's rate that is about six more person-years [2], which for a company that spent nearly 20 years on one game [16] is a rounding error rather than a strategy meeting.
So the forcing function for anyone tempted to franchise their own playbook sideways has two axes. First, does the adjacent activity have a skill ladder your users will accept as legitimate. Second, do your existing users recognize your furniture well enough to skip the learning curve. Both yes is poker for chess players, and it is the only quadrant where two people is a plausible headcount. Ladder yes, recognition no, and you are launching a cold brand and buying users. Recognition yes, ladder no, and you get visitors with no reason to ever see a paywall. Neither, and what you have built is a marketing page with a login.
The honest state of the evidence: one beta launch from May now getting its first livestreamed event and creator collaborations [1][17], three unnamed dates, and no retention or conversion number for the site that is supposed to prove the pattern.
Ranked by verification strength, evidence, and original report placement.
Chess.com quietly launched Gambit, a free online poker website where players learn without risking real money, in beta in May, as its first companion site.
Chess.com chief growth officer Albert Cheng told The Verge there will never be real-money gambling on Gambit: "Real money is the problem we built Gambit to get away from, the whole industry makes money when beginners lose. We'd rather build a place where people can actually learn."
Cheng said Chess.com chose poker for its first companion site because, like chess, it has learning and social elements on top of playing.
Gambit borrows Chess.com's design style and includes similar play modes: bot matches, games with friends, tournaments, a review mode for studying past games, and a learning mode.
Gambit borrows Chess.com's design style and includes similar play modes, including bot matches, games with friends, tournaments, a review mode and a learning mode.
Gambit includes Chess.com's bot characters with matching skill levels; The Verge's writer immediately recognised "Martin", the lowest skill level bot on Chess.com, as also the lowest skill level bot on Gambit.
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 30, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
build
The FCC read HOVERAir's Versa by what it was built to do, not how it was boxed2 distinct publishers
build
Chess's one-game-a-day cap is a two-core bill, and WebAssembly moves it onto your laptop1 distinct publisher
product
Google's new Preferred Sources button hands publishers the recruiting job2 distinct publishers
product
Google's Pet Memory names your pets, then calls all three cats Smokey1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
One interview, one hands-on check
The Verge did verify the part it could touch — the shared design language, the modes, Martin sitting at the bottom of Gambit's bot ladder just as he does on Chess.com. Everything else traces to a single conversation with Chess.com's chief growth officer: the subscription mechanics, the two-person headcount, the AI contribution, the promise about real money. No documents, no second voice, no numbers.
Live and datable, but uncounted
Two things are real and checkable: the site has been open in beta since May, and the first Gambit Cup is on the calendar for August 31 with Rensch and Botez fronting it. What no one supplies is anyone playing — not registered users, not paying subscribers, not hands dealt. Go, backgammon and mahjong exist only as "moving quite rapidly", with no dates.
Promises running ahead of proof
The reporting is calmer than the pitch it carries, but two claims outrun what anyone can check. "There will never be real-money gambling" is a pledge with no mechanism behind it, and the AI-made-this-possible framing depends entirely on Cheng's guess about the team he would otherwise have needed. To his credit he shrinks his own story where it matters — brand and feature work resisted the tooling, and the shipped site has "zero or close to zero" generated assets.
The only voice is the seller
Timing tells you something: the interview runs the day before Gambit's first livestream, and the sole source is the person whose title is chief growth officer. Two of his messages do commercial work at once — free poker positioned against an industry that "makes money when beginners lose", and a two-person build that makes three more games sound almost free to attempt.
Single account, verifiable surface
One publisher, one interviewee — but the claims are specific, dated where they can be, and much of the product side is checkable by anyone who opens the site. The business half stays soft until Chess.com or someone else puts a number to it.