The first version of OP.GG lived in Sangrak Choi’s house, on a homemade server worth roughly a thousand dollars. It was the sort of machine a talented young developer might regard with pride: bought in parts, assembled by hand, adequate for a side project. Then the players came. The box could not carry them. Traffic overwhelmed it almost as soon as the site launched, converting a pleasant hobby into an urgent lesson about demand.
The usual startup retelling would treat the overloaded server as destiny calling. Choi’s account is more useful because it stays close to the work. Building a replacement saved about a hundred dollars. Assembly took a day. When the new machine produced neither video nor a network connection, troubleshooting took three more. In another telling, a friend sold his car quickly to help pay for stronger hardware. The romance lasted about as long as the uptime.
What mattered came next. For roughly a year, Choi and his collaborator absorbed feedback and improved the service around the clock. The site answered a question that League of Legends players already carried with them: what just happened, and what should I do differently next time? OP.GG made the answer fast, legible, and shareable. A match ended. Curiosity began. The search box sat between them.
A childhood spent shipping
OP.GG did not arrive as Choi’s first encounter with strangers on the internet. On his personal site, he writes that he started making Harry Potter and StarCraft fan sites at 11. Before university, he built game services for Nexus TK, TalesWeaver, and MapleStory. One community gave visitors custom avatars assembled from game assets and mini-maps extracted from the client. A later Minecraft server reached 20,000 registered members and used SMS verification to limit trolling.
These projects form a peculiar kind of apprenticeship. They teach code, but they also teach the sociology of buttons. People ignore the feature you polished. They find the edge case you missed. They invent uses, conventions, and complaints. A young builder learns that software is never just the system imagined at a desk. It is the system plus every person who touches it.
Choi carried that apprenticeship into paid web work and a social app company, then back into games. He started OP.GG as a hobby in 2012. The beta arrived in early 2013. By his account, he still had a day job, so the second shift began after dinner. In one interview he remembered sleeping about three hours and staying attached to the messenger window. The company was incorporated in 2015. In 2016, during his third year at Keimyung University, he left school to concentrate on it.
“Coding is practically my entire life.”Sangrak Choi, on his personal website
That sentence sounds sweeping until you inspect the evidence. Choi lists more than 20 years of web programming and experience in PHP, Java, Python, C++, C#, Objective-C, Swift, and Go. His professional profile says he built and designed the original OP.GG site, engineered databases handling thousands of inserts per second, and managed the visual work through 2015. He has the classic engineer-founder tension: he can think in tables with billions of rows and still notice a pixel that sits one place off.
The product was the next decision
Statistics products are easily mistaken for warehouses. Gather enough numbers, arrange the shelves, and wait for a user to admire the inventory. OP.GG’s more durable insight was to make data point toward an action. A player could inspect an opponent’s preferred champions, win rates, runes, rank, and recent matches. The information mattered because another game was coming.
Choi’s memory of an early MMR feature captures this bias toward usefulness. He did not know all the mathematics he needed, so he wrestled with formulas and spreadsheets through the night. He watched the community respond and changed the calculation hundreds of times. The feature has since disappeared, but the method remains recognizable: publish an answer, observe where it fails, make the answer less wrong.
The loop was also OP.GG’s marketing. Early on, the team worked through game communities, applied feedback quickly, and told users when changes went live. Choi has said the company grew without much conventional advertising or public relations. The brand traveled because the tool sat inside a recurring ritual. Players checked themselves, checked teammates, checked opponents, and returned after the match to check again.
A good utility can look inevitable after it becomes habit. It was not inevitable when Choi chose to compete with existing communities and statistics services. His thesis was almost embarrassingly specific: a better interface could matter. It did. Design was not decoration applied to a database. Design determined whether a pile of records became an immediate judgment.
When the side project becomes an institution
The product widened. OP.GG added games, global languages, mobile apps, a desktop client, communities, esports projects, and services for players beyond the original search page. The company launched a professional gaming team, worked with publishers, sponsored League of Legends Champions Korea, and in 2022 acquired OGN, the broadcaster that had helped turn Korean competitive gaming into a spectator culture.
The acquisition carried a certain symmetry. As a child, Choi watched StarCraft and built communities around games. OP.GG had grown by helping players interpret competitive play. Buying OGN brought a historic storyteller of esports into the same orbit as the data platform. Search, discussion, competition, and media now looked less like adjacent experiments and more like pieces of one gaming environment.
- OP.GG beta launches for players
- The hobby becomes an incorporated company
- Desktop launches and OP.GG gains K-Unicorn status
- OP.GG acquires OGN and earns major design awards
- Choi speaks about community in Morocco and adds US gaming leadership
Scale also changed Choi’s job. He says he leads more than 100 people, with 12 managers reporting to him, and has conducted thousands of interviews. He has hired hundreds and trained more than 20 leaders. Yet his stated preference is for teams of three or four. It is an intriguing management confession: the founder of a large platform still thinks fondly in the dimensions of a small room.
Small teams compress the distance between observation and action. Large companies produce distance by default. Managers, plans, and dependencies gather between the user and the person making the change. Choi’s challenge is no longer proving that he can build a feature quickly. It is designing an organization that preserves the attentiveness of the original two-person project without preserving its three-hour nights.
“We always make decisions with the gamer at the center.”Sangrak Choi, describing OP.GG’s operating philosophy
Still near the workbench
Choi’s public GitHub suggests he has not retreated entirely into executive abstraction. The repositories move across AI translation, Laravel, macOS utilities, VS Code extensions, flight search, and tools for AI coding agents. The subjects vary, but the gesture is consistent: notice friction, package a solution, ship it where another person can try it.
His personal site adds photography, design, business, behavior, and psychology to the picture. Travel images from Jeju, Hong Kong, Los Angeles, New York, Macau, Okinawa, and elsewhere sit beside descriptions of programming and management. The combination makes sense. Product builders spend their careers deciding what deserves attention. A camera makes the same demand in a different form.
In 2024, Choi took the community argument to Morocco Gaming Expo. He described feedback as a source of unfiltered insight into what users actually want and spoke about building esports ecosystems. In early 2025, he announced former Blitz founder Adil Virani as OP.GG’s chief strategy officer, connecting the hire to a deeper relationship with American players. The geographic ambition has grown, but Choi’s language still returns to communities and the signals they send.
His broader aspiration has been to make OP.GG a brand players recall regardless of which game they open. That is a harder project than building one excellent statistics page. Games have distinct cultures. Communities resist being flattened into a generic audience. Expansion therefore asks OP.GG to repeat its first trick in new contexts: learn the local question before presenting the global answer.
What builders can borrow
The transferable lesson is not to start a game-statistics company. It is to treat usage as correspondence. A crash says the capacity plan was wrong and the demand estimate was wrong in a happier direction. A complaint says the user expected a promise the product has not kept. A repeated search says a question has become a habit. Each signal is a note from the other side of the interface.
Choi’s story also argues for the value of a long prehistory. The fan sites, freelance jobs, game communities, servers, interfaces, and failed formulas did not look like a tidy founder curriculum while they were happening. They accumulated into judgment. By the time OP.GG found its audience, he had already rehearsed the central act many times: build something for people who care, then remain present when they tell you what it got wrong.
The homemade server is now a charming artifact from a company origin. The more consequential machine was the loop around it. Player asks. Builder answers. Product breaks. Builder learns. The loop turns again, first in a house, then across a company, then across countries and games. OP.GG grew because the loop kept running.