UFC Fight Night recap: Allen vs. Costa
Reading our own card. The model had Andre Petroski at 61.6%; Cody Brundage won. Per-pick grade live.
Insights
Thinking
out loud.
Hawk-Eye, Second Spectrum, and the 29-Point Question
The NBA's shift from 2D dot tracking to 29-point 3D pose tracking unlocks a qualitatively different analytics layer. Most franchises are still buying derived metrics from yesterday's system.
UFC 328 recap: Chimaev vs. Strickland
Reading our own card. The model had Khamzat Chimaev at 75.1%; Sean Strickland won. Per-pick grade live.
The Maycee Barber Miss: A UFC Model Postmortem
Our UFC model gave Maycee Barber a HIGH-confidence edge over Alexa Grasso. She got knocked out at 2:42 of round one. The postmortem, the four code changes that shipped twenty-four hours later, and what happened when the headline backtest result later got superseded by a stricter validation audit.
What the 2026 NFL Draft Looks Like When You Run It Through a Predictive Stack
Draft valuation is a prediction problem with sparse features and a long feedback loop. Franchises that treat it that way extract more surplus value from picks 30 through 150.
Lineup-Adjusted Efficiency, Explained Without the PhD
Raw offensive and defensive ratings lie to you. Lineup-adjusted efficiency is the correction that turns possession-level data into roster-construction signal.
The Third Circuit Just Settled Prediction Market Jurisdiction. Here's What It Means for Infrastructure Buyers.
The April 7 Third Circuit ruling is the first durable federal signal that prediction market infrastructure is a national product. Here's what changes for vendor procurement.
Why Calibration Matters More Than Accuracy in Franchise Analytics
A miscalibrated analytics model doesn't just give a franchise the wrong answer. It gives the coaching staff and front office the wrong amount of confidence in the right answer, which is worse. The difference between a model that says "72% chance this prospect becomes a starter" when the true rate is 72% versus when the true rate is 55% is the difference between a sound draft pick and a wasted first-round selection. Most franchise analytics departments optimize their models for accuracy. They sh
Look Again. It Is A Wide Open Market
Two weeks ago, Teamworks acquired PFF's enterprise business for $100M+. As someone building an independent sports prediction company, I've been thinking about what this means. Teamworks now owns Zelus Analytics, Telemetry Sports, Sportlogiq, and PFF's enterprise data platform. That's an extraordinary concentration of football analytics infrastructure under one $1B+ company. For NFL franchises, this could be a net positive — one integrated platform for film, data, and operations instead of a pa
AI Can't Bet on Sports? Wrong AI, Right Problem.
General Reasoning just released KellyBench — a study that tasked frontier LLMs with betting a full Premier League season. The results made headlines everywhere from Ars Technica to Benzinga: every model lost money, Grok went bankrupt, and Claude — the best performer — still finished down 11%. The headline is fun. But the lesson isn't "AI can't bet on sports." It's that general-purpose language models aren't built for this. What KellyBench Actually Tested The study placed eight frontier AI sy
We Found a Bug That Made Us Look Better Than We Are. That's the Point.
A few months ago, our NFL simulator was showing 72% ATS accuracy. That's an extraordinary number. If it were real, we'd be printing money. We almost believed it. We didn't ship it. We pressure-tested it. And we found a bug — a subtle data leakage issue that was inflating our results. Once we isolated and fixed it, our validated accuracy came down to 65%. Still strong. Still well above market break-even. But seven points lower than what a less rigorous process would have let us claim. That mome
The Lines Are Blurring
Six months ago, prediction markets were a niche conversation among quant traders and crypto-native speculators. Today, MLB has an official prediction market partner. The NHL gave both Kalshi and Polymarket access to league marks. MLS built an entire authorized operator framework around the category. And the CFTC is suing states for trying to regulate what it considers its exclusive jurisdiction. The old taxonomy — DFS in one bucket, sportsbooks in another, prediction markets somewhere closer to
The Problem With Expert Picks (And What We Built Instead)
Sports media runs on experts. Former players, analysts, journalists with decades of game experience — they break down film, assess matchups, and deliver picks with authority. And they are wrong about as often as they are right. This is not a knock on expertise. Film study and pattern recognition are real and valuable. The problem is that human experts — no matter how knowledgeable — systematically struggle with two things that determine outcomes in sports: quantifying uncertainty and processin
65% ATS — What It Means, and What It Does Not
💡 Editorial update — May 5, 2026: This post was originally published when our headline NFL ATS figure was 65%. After identifying and correcting a model bug during the 2025–2026 season (described in https://victory-ar.com/insights/we-found-a-bug-that-made-us-look-better-than-we-are-thats-the-point), our canonical multi-season figure is 63.2% across six validated NFL seasons. The reasoning here about how to interpret a high ATS rate is unchanged — only the headline number has been When we tell
Why Single-Point Predictions Are Lying to You
Every sports analytics tool you have ever used gives you a number. Team A wins 64% of the time. Player B scores 22.4 points per game. The spread is -3.5. These numbers feel precise. They are not. They are averages — and averages hide the thing that actually determines outcomes in sports: variance. A team that wins 64% of the time still loses 36% of the time. The question is not whether they win — it is under what conditions they lose, and whether those conditions are present this week. This