WorldQuant IQC: What I Learned Placing 2nd in Korea
When I was preparing for this competition, I had no seniors or experienced peers to turn to. I still remember how much I wished I could find even a small piece of useful information. I’m writing this in hopes that it reaches someone in that same position.
Table of Contents
- IQC 2026 Is Here
- What Is WorldQuant IQC?
- Competition Timeline
- Early Game
- Strategy 1: The Thinking Process
- Strategy 2: Know Your Data
- Strategy 3: Overfitting Checklist
- What This Post Doesn’t Cover
I placed 2nd in Korea at last year’s WorldQuant International Quant Championship (IQC).
IQC 2026 is about to start, and that’s why I’m finally writing this down. When I prepared, I had nobody ahead of me to ask, so I’ve tried to put together the things I would have wanted to read back then.
IQC 2026 Is Here
International Quant Championship 2026, hosted by WorldQuant Brain, is now open.
The total prize pool is $100,000 USD. Strong participants may also get the chance to become a BRAIN Research Consultant or be considered for full-time researcher roles and internships. Teams that make it through can compete at the global final in Singapore in September 2026.
For registration and full details, visit the official page: 👉 IQC 2026 Official Page
Korean-language webinars are also available, and I’d recommend registering early.
What Is WorldQuant IQC?
It’s a global quant competition hosted by WorldQuant. Participants design and submit alphas (predictive signals for stock returns) on the WorldQuant Brain platform. Coding matters less than you might think. What matters is being able to read financial data and understand market structure.
Competition Timeline (Last Year’s Schedule, Subject to Change)
The schedule below is based on last year’s competition. The 2026 timeline may differ, so always check official announcements.
- Weekly Wednesday online webinars (last year: 18:00–19:30 KST / this year I’ve heard it may shift to Tuesday or Wednesday at 20:00)
- Webinar recordings uploaded approximately 2 weeks later
- Mid-May: preliminary round closes; first OS score release
- Late June: main round ends
- Mid-July: national-stage finals (the global final comes later; in 2026 it is in Singapore in September)
The most critical moment in the timeline is the OS score release in mid-May. This is the point where your strategy needs to change entirely depending on where you stand.
Early Game: March to Explore, April to Execute
The operators and data fields provided change every year. Reusing alphas from previous competitions is nearly impossible.
March should be treated as a time for exploration and learning. Experiment with operator combinations, understand the characteristics of data fields, and get a feel for correlational structures. Build that intuition first, then shift into full alpha production from April onward.
When I first opened the Brain platform, I genuinely didn’t know where to start. Even using a single operator felt unfamiliar, and I had no sense of which data field combinations would produce meaningful signals. When my first alpha failed to meet even the basic submission criteria, I was honestly caught off guard. Rather than rushing forward, I spent time slowly working through how each operator actually functioned and how the data was structured.
More than trying to build a “good alpha” right away, understanding what each piece of data is and how it behaves turned out to be far more valuable.
Strategy 1: The Thinking Process Behind Building an Alpha
When I built alphas, I didn’t start by asking “which indicator should I use?” I started by asking “why would this data have any relationship with future returns?”
For example, rather than just plugging in a momentum indicator, I first thought about “what lookback period for momentum actually matters?” and “are there specific market environments where momentum tends to work?” That line of questioning became the idea, and the idea became the formula.
I got stuck plenty of times. When an idea I believed in didn’t work on either the train or IS window, I spent a long time not knowing what was wrong. Swapping indicators didn’t help. What helped was stepping back and rethinking the structure of the idea itself.
When you first start building alphas, there’s a strong temptation to make them complex. Stacking multiple operators, layering in conditions, and fine-tuning the formula can feel like you’re making something more precise. In practice, the alphas that actually worked tended to be built on ideas that were simple and clear. Adding complexity was often just another word for overfitting.
Strategy 2: Know Your Data Thoroughly
The Brain platform provides a wide variety of data fields, and each one has its own characteristics. An approach of “I used this field and the IS score looked good” won’t hold up for long.
Before using any data field, the first thing I always checked was coverage.
Data Coverage Coverage by data field is publicly available on the Brain platform. I prioritized fields that met a certain coverage threshold before using them. I also ran simulations and used various operators to directly analyze how much coverage was actually being captured in practice.
I think this habit, taking time to understand the data before using it, ends up deciding the quality of your alphas more than anything else.
Strategy 3: A Practical Checklist for Catching Overfitting
A good IS score doesn’t mean a good alpha. I went through the following checks every time.
Long/Short Balance Check whether the alpha is heavily skewed in one direction or overly dependent on market directionality. An alpha that is structurally betting on one side tends to be unstable over time.
Period-by-Period Performance Decomposition Don’t just look at overall returns. If performance is concentrated in a specific window, the alpha is likely environment-dependent. Separately examine periods of high volatility and periods with strong directional trends.
Read Metrics, Don’t Just Look at Them Don’t treat Sharpe, Fitness, and Turnover as numbers to maximize. Ask yourself why these figures came out the way they did, whether the performance is structurally reproducible, and whether it reflects genuine edge or just luck.
In the end, what matters most in this competition is
not the ability to build good alphas, but the ability to identify them.
What This Post Doesn’t Cover
There’s a lot more I could say. Submission strategy, the intuition for reading an alpha, and what to do toward the end of the competition are hard to put into words, and some of it you only really get by doing it yourself.
This post was written as lightly as possible, with the hope of helping first-time participants find their footing. The rest is yours to discover along the way 😄
Meeting people who are working toward the same goals is one of the more underrated parts of this competition. If there are any offline events or networking opportunities, I’d encourage you to show up. Simply talking with people who share the same questions often helps clarify your own thinking more than you’d expect.
I hope this post is useful to someone out there preparing for the competition. 🙂