Olevicor

Quantitative research & tradingChicagoEst. 2023

Systematic trading in U.S. index options.

Olevicor is Corey Olson’s independent research and trading operation. Its systems trade multiple strategies autonomously, with his own capital, on infrastructure he built.

The approach Read the research
Founded2023
MarketsU.S. index options
HorizonIntraday, same-day expiry
CapitalProprietary only
01How it works

How an idea becomes a trade

01.1

Research

Each idea starts as a written, testable claim. Most are discarded before they trade.

The lifecycle
01.2

Execution

Orders go through an in-house engine, and fills are reconciled against the broker. The code that trades is the code that was backtested.

The replay ladder
01.3

Risk

Hard limits on loss and exposure, checked before every order.

Risk controls
02Research

Latest note

All research
Note Oct 1, 2026 6 min read

The at-the-money call has no price history

A rolling at-the-money selector splices bars from different contracts, so any moving average or rolling statistic over it measures the splice. Compute per contract, and use the selector only to choose which value to read.

Read the note
Fig. 1 · Stitched EMA vs each contract's ownIllustrative
Both contracts' own EMAs rise steadily, but the alias EMA drops at the bar-7 re-point and by bar 10 matches neither contract.
03AI

Where the AI sits

Agents run as supervised, long-lived services that claim work from a queue. A model can propose a change; it can’t approve one.

  • Eight statistical gates accept or reject each proposed change.
  • The deploy call stays with a person.
How the pieces fit
Who decides whatAuthority
  1. Human Sets the problem and makes the deploy call
  2. Engine Decides what is admissible; eight statistical gates accept or reject each change
  3. LLM Proposes changes, searching only inside what the engine allows
04Platform

Written and run in-house

For any trade, the system can show what it saw, when it saw it, and why it acted.

  • Live is the reference: each faster replay mode earns trust by matching the one above it.
  • Replays are reconciled against live trading every night.
Explore the platform
Each mode adds one production layerReplay ladder
  1. Mode 1 Discovery
  2. Mode 2 Screening
  3. Mode 3 Integration
  4. Mode 4 Replay
  5. Mode 5 System
  6. Mode 6 Simulation
  7. Live Production
05Who builds it

Corey Olson

Twenty-one years building software. Before Olevicor, he founded and ran Vanilla Video for fifteen years, serving more than 4,000 clients on systems he built, and wound it down by choice. He also built Pullboard, a work board for AI agents.