How Provably Fair Casino Games Changed Online Gambling Forever

In April 2012, Bitcoin was trading at around $5. The whole thing still felt like a forum hobby. Mt. Gox — the world’s largest exchange at the time — had started life as a website for trading Magic: The Gathering cards. Mining was still possible on a gaming PC if you didn’t mind the noise; nobody had heard of an ASIC yet. The community lived on a single forum, Bitcointalk, where regulars tipped each other in satoshis for high-quality meme posts, sold each other VPNs and web hosting for BTC, and argued about whether this stuff would ever be worth $100.

That was the world into which a 27-year-old named Erik Voorhees announced, on the 24th of that month, a website called SatoshiDice.

There was no signup. No login. No deposit screen, no balance, no account. You picked your bet by sending Bitcoin to one of a few dozen public addresses, each one tagged with its odds and payout multiplier. Within seconds — often before the block even confirmed — the answer came back as another Bitcoin transaction: a single satoshi if you lost, your stake times the multiplier if you won. The address was the bet. The transaction was the wager. The blockchain was the receipt.

Within months, SatoshiDice was responsible for up to half of all Bitcoin transactions on the network — much to the irritation of every full-node operator who suddenly had to store millions of one-satoshi “you lost” notifications in the chain forever. The community had a lot of feelings about that. By July 2013, Voorhees had sold his ownership interest for 126,315 BTC — about $11.5 million at the time, and roughly $9.8 billion at today’s BTC price. It was the first major acquisition in Bitcoin’s history. The buyer never put their name on the deal, and to this day nobody knows who they were.

That single site started a category. Everything modern crypto casinos do — Dice, Crash, Limbo, Mines, Plinko, the whole genre we now call Originals — descends from a handful of products built by a handful of people between 2012 and 2015, for an audience of crypto nerds, libertarians, miners with too much idle BTC, and the sort of person who would write their own auto-play bot for a dice game before going to bed. They didn’t trust anyone – and that was the whole point.

This is how that happened, and why it still shapes what we build today.

SatoshiDice wasn’t really “provably fair” by modern standards. There was no commit-reveal, no client seed, no per-bet nonce. What it had instead was public verifiability of the output: every bet was a Bitcoin transaction, every payout was a Bitcoin transaction, and the entire ledger of wins and losses was sitting on the blockchain for anyone to audit.

The mechanism was elegant in its bluntness. Voorhees and his team kept a secret number. When your transaction confirmed, the site combined that secret with your transaction ID, hashed it, and produced a number between 0 and 65,535. If it fell in the winning range of the address you sent to, you won. Periodically, SatoshiDice would publish the secrets that had been used, so anyone could retroactively verify every outcome.

It was clever, transparent enough to win trust, and very profitable for such a niche site. In August 2012, Voorhees did something stranger still: he listed SatoshiDice as a dividend-paying stock on MPEx, a fully unregulated, Bitcoin-denominated securities exchange run by a Romanian writer named Mircea Popescu. Public shareholders owned 10% of the company and were paid monthly dividends in BTC. No SEC filing, no regulator, no jurisdiction — just a stock that paid in Bitcoin. The SEC eventually came asking questions. The exchange simply shrugged. Roger Ver — the Bitcoin evangelist already known as “Bitcoin Jesus” — was an early backer.

The house edge was around 1.9%. That is not a small number by modern Originals standards, but compared to the slot floor of any conventional casino, where 4–15% is normal, it was already a revolution.

SatoshiDice’s on-chain model had a ceiling. Every bet was a Bitcoin transaction, which meant every bet paid Bitcoin fees and waited for blocks. By 2013, fees were rising and the model started to feel slow.

The solution came from a site called BitZino, which is generally credited with the first commit-reveal “seeded” provably fair system. The pattern looked like this:

This is the trick. The casino committed to the random source before they knew what you were going to bet, and they can’t change it after the fact without breaking the hash. The player participates in the randomness, so the casino can’t pre-mine a seed that’s bad for them either.

You don’t have to trust the operator. You verify.

This pattern — server seed, client seed, nonce, HMAC, public commitment — became the template for almost every provably fair game that followed. And once it existed, the games didn’t need to be on-chain anymore. They could be instant.

In June 2013 — a month before Voorhees would sell SatoshiDice — a Canadian developer named Chris Moore, known on Bitcointalk as dooglus, launched Just-Dice.com.

The game itself was simple to the point of austerity: pick a number, pick whether to roll over or under, pick your bet size, click. The house edge was 1% — already the modern standard, set right there at the start.

In its first month, Just-Dice processed 429,600 BTC in wagers — around $38 million at the time, and tens of billions of dollars at today’s BTC price. The interface was a single page. There was no marketing, no bonus, no welcome offer. Just a dice slider and a balance.

Just-Dice’s real innovation wasn’t the game. It was the investor bankroll. Anyone could deposit Bitcoin into Just-Dice not to bet, but to act as the house. You earned a share of all profits — and absorbed a share of all losses — proportional to your stake. It turned gambling into a two-sided market: gamblers on one side, capital on the other. Modern casinos still use variations of this structure. So do prediction markets.

Dooglus was also famous for an incident that defined the culture of early crypto gambling. In July 2013, a player won 1,300 BTC and asked to withdraw it. dooglus paid out the 1,300 BTC to the player’s wallet — but forgot to deduct it from the player’s in-site balance. The site still showed 1,300 BTC sitting in their account, and the player promptly gambled it back and lost the lot. On paper the casino had won 1,300 BTC; in reality that BTC had already left the bankroll in the original withdrawal, and the player had been gambling a phantom balance. Because Just-Dice’s bankroll was investor-funded, those investors — third-party users who had put their own BTC up to act as the house — were now down 1,300 real coins through no fault of their own. dooglus covered the shortfall out of his own pocket, and the Bitcointalk community responded by pledging to donate future winnings back to him until he was made whole. That was the audience.

The other big genre to come out of this period wasn’t dice. It was crash — a game with no real-world casino analogue, invented entirely by and for crypto players.

The original was called MoneyPot, introduced on Bitcointalk in July 2014 by a Canadian developer named Eric Springer (handle: espringer). A multiplier started at 1.00x and climbed in real time. Every player who had placed a bet could cash out at any moment, locking in bet × current multiplier. At some random point, the multiplier “crashed” and anyone still in lost their bet.

It was beautiful. It was social — everyone watched the same number climb at the same time. It was tactical — your decision was when, not what. And the math was clean enough to make provably fair trivial.

In 2015, Springer sold MoneyPot to another Bitcointalk regular, Ryan Havar, who rebranded it to Bustabit and added the things that define the modern crash format: an investor bankroll (borrowed from Just-Dice), a custom scripting API so players could write their own auto-play bots, and the famous “last-longer bonus” that funded itself from a 1% cut of every wager.

Every crash game on every crypto casino today — Stake’s Crash, BC’s Crash, Aviator (which brought the format into the fiat world), all of them — is a descendant of those two products. Same core idea. Same math. Same culture of players writing scripts to play their game for them. (Our own variation, Turbocrash, takes the format further with faster rounds for players who want the loop tightened up.)

In 2013, two Australians who had been banned from RuneScape for running an in-game gambling scheme started a dice site of their own. Their names were Ed Craven and Bijan Tehrani, and the gambling scheme they had run in RuneScape was called “staking.”

PrimeDice was, on the surface, just another 1% house edge dice site in a year that had a lot of them. What made it different was the operator quality. The site loaded fast. The chat worked. The withdrawals were instant. Affiliates got paid. By 2016, Craven and Tehrani were ready to do something bigger, and in 2017 they launched the brand they had clearly been building toward all along: Stake.

The line from PrimeDice to Stake to the entire modern Originals genre is direct. Dice was the proof of concept. The product taste they developed making one game work well at scale is what defines every Originals lobby today.

Modern casino marketing imagines a customer who wants free spins, lobby art, a welcome bonus, and a celebrity face on the homepage. The customer who built the Originals genre wanted none of that.

The early provably fair player was, roughly:

That customer profile is what makes the architecture of these sites so different from a traditional casino — and it still shapes the products today, even now that crypto casinos serve a much broader audience.

A conventional online casino in 2013 had a wallet that looked like a bank account. You deposited fiat by card or wire, you saw a balance, and you withdrew back to the same source after a multi-day review.

A SatoshiDice “wallet” wasn’t a wallet at all. There was no balance, no login, no withdrawal. Every bet was its own deposit, every payout was its own withdrawal, and you never had funds sitting at the casino at all. The model was radical: don’t custody player funds.

Just-Dice, PrimeDice and Bustabit moved to a more conventional balance model — you deposit, you play, you withdraw — but kept the rest of the radicalism. Deposits and withdrawals were on-chain only. Withdrawals were instant or near-instant — there was no human review, no “pending” hold, no withdrawal limit calibrated to slow players down. You could deposit one BTC, play for an hour, and have everything back in your own wallet inside a minute of pressing “withdraw.”

This was a different relationship with the customer. The site held your money as briefly as possible, and the cryptography made the act of holding it irrelevant — the seed commitment proved the game was fair before you ever placed a bet, and the blockchain proved the casino had paid what it owed.

You can draw a straight line from those mechanics to what modern crypto casinos still do better than fiat ones: faster withdrawals, lower edges, public APIs, scripting tolerance, real-time stats and proofs. The Originals genre carried the cultural DNA of provably fair forward even as it grew into a mass-market product.

The audience has broadened — Stake and its peers now serve millions of players, most of whom never read a hash in their lives. But the product DNA hasn’t changed.

A modern Dice game still uses HMAC-SHA256, server seed + client seed + nonce, public seed commitment, and a 1% house edge — exactly the design dooglus shipped in June 2013. A modern Crash game still uses the same multiplier-climbing mechanic, the same provably fair seed commitment, and a structured auto-bet UI that descends directly from Bustabit’s scripting subculture (even if today’s version is a form, not a JavaScript console) — all of it Bustabit’s contribution. The whole “Originals” category is a label for house-built games that descend from this lineage, distinct from licensed slot content the operator is just reselling.

Cubeia Originals is our take on the genre this post is about — a full suite of provably fair games built in the lineage of SatoshiDice, Just-Dice and Bustabit, and brought up to the standard a modern operator and a modern player both expect.

The library covers the core Originals canon — Dice, Crash, Limbo, Mines, Plinko, HiLo, Wheel, Keno, Coinflip, Roulette, Blackjack — plus a few of our own additions like Turbocrash, Tower, Loop and Chicken Cross. Every game ships with verifiable HMAC-SHA256 seed commitments, in-depth auto-play, per-bet odds and payout transparency, and the granular volatility controls that make a single Original cover the same ground as an entire slot category.

What makes Cubeia different from the rest of the field is how we ship them: as deeply brandable web components rather than locked-down iframes, with a careful styling and theming process so each game looks and feels native to the operator running it.

If you want to see the games in action, try them at originals.cubeia.com — no signup, just the games themselves. If you’re an operator and you want to talk about putting them on your platform, that’s what we’re here for.

Further reading: