Ron Yurko profile photo

Ron Yurko

Assistant Teaching Professor Carnegie Mellon University

  • Pittsburgh PA

Research focuses on developing methods at the interface of inference and machine learning, oriented towards problems in sports analytics.

Contact
Carnegie Mellon University logo

Carnegie Mellon University

View more experts managed by Carnegie Mellon University

Biography

Ron is an Assistant Teaching Professor in the Department of Statistics & Data Science at Carnegie Mellon University, and the Director of the Carnegie Mellon Sports Analytics Center (CMSAC). CMSAC’s initiatives include an active sports analytics research lab, a summer undergraduate research experience in sports analytics (aka CMSACamp), and the annual Carnegie Mellon Sports Analytics Conference.

His research focuses on developing methods at the interface of inference and machine learning, oriented towards problems in sports analytics and natural language processing.

Areas of Expertise

Sports Analytics
Data Science
Language Processing
Machine Learning & Artificial Intelligence

Media Appearances

Why only 7 home runs at PNC Park have reached the Allegheny River on the fly

Pittsburgh Post-Gazette  online

2026-06-29

“It’s the combination of having to hit it at a certain angle as a left-handed hitter, being pulled in this way, which makes it a more rare occurrence,” said Ron Yurko, a Carnegie Mellon statistics and data science professor.

View More

Should Hitters Adjust Their Swing According to the Count? Rice Study Provides Insight

Bioengineer.org  online

2026-06-08

New research led by Scott Powers and Ron Yurko delves deeper into this strategic facet, challenging some of these well-entrenched notions by leveraging an unprecedented trove of Major League Baseball swing-tracking data.

Intriguingly, this empirical investigation largely confirmed traditional baseball lore, which posits that batters “choke up” on the bat or shorten their swing in two-strike counts to optimize contact probability. “Batters can reduce their strikeout rate by changing their swing length based on the count,” Yurko explained. This affirmation lends credence to decades of coaching and player experience, signaling that some aspects of conventional wisdom have an underpinning in measurable biomechanical and performance data.

View More

Carnegie Mellon's 'Future of Sport' showcases AI, analytics possibilities to NFL executives

WTAE ABC Pittsburgh  tv

2026-04-22

“It’s a wonderful showcase for the broader audience to really see this,” said Ron Yurko, director of CMU’s Sport Analytics Center. “When people think of sports analytics, they think of Carnegie Mellon University. They think of robotics. They think of this amazing amount of technology that’s here.”

Yurko was one of several people who showcased their inventions and studies to NFL officials like Commissioner Roger Goodell. Alongside Goodell were Steelers Hall of Famer Jerome Bettis and famous entrepreneur and Mt. Lebanon native Mark Cuban.

View More

Spotlight

4 min

Professional sports have always embraced innovation, but today's competitive advantage increasingly comes from science. Researchers are applying advances in neuroscience, artificial intelligence, biomechanics, data analytics, and human performance to better understand how athletes make decisions, respond under pressure, recover from injury, and maximize performance. What once relied heavily on intuition and experience is now being informed by sophisticated research that can measure, predict, and improve outcomes at every level of competition. Recent studies from Carnegie Mellon University highlight the growing role science is playing across the sports landscape. Whether examining decision-making in high-pressure situations, analyzing performance strategies, or using artificial intelligence to improve health outcomes, researchers are uncovering insights that can help athletes perform at their best while extending careers and reducing injury risk. Ron Yurko is an Assistant Teaching Professor in the Department of Statistics & Data Science at Carnegie Mellon University, and the Director of the Carnegie Mellon Sports Analytics Center (CMSAC).  View his profile Scott Powers, an assistant professor at Rice University with vast front-office experience in Major League Baseball—including stints with the Los Angeles Dodgers and the Houston Astros—joined forces with Ron Yurko, a director at the Carnegie Mellon Sports Analytics Center, to analyze this cutting-edge data. Their study, published in The American Statistician in 2026, marks a significant advancement in the quantitative understanding of batting dynamics. It uses high-resolution measurements of bat speed and swing length, metrics that were publicly released for the first time in 2024, to explore how hitters modulate their swings under different pitch counts, particularly when facing two strikes. Eric Yttri is an Associate Professor at Carnegie Mellon University where his research goal is to establish how neural circuits lead to these action selection decisions.  View his profile As a neuroscientist, I have been working to uncover how the brain decides when to act and when to wait. Recent research from my team and me helps explain why this split-second pause happens, offering insight not only into elite athletic performance, but also how people make everyday decisions when the potential outcome isn't clear. We found that the key to hesitation is a response to uncertainty. This could be where a dropped hockey puck will land, when a race starts, or placing your order at a new restaurant. Eni Halilaj is an Associate Professor at Carnegie Mellon University where she directs the CMU Musculoskeletal Biomechanics Lab, an interdisciplinary group of engineers seeking to understand and optimize human movement mechanics. View her profile According to Eni Halilaj, an assistant professor in mechanical engineering at Carnegie Mellon University and biomechanist who specializes in orthopedic rehabilitation, 60 percent of those who suffer this common knee injury also develop osteoarthritis early in life. The degenerative joint disease, which affects an estimated 32.5 million individuals in the U.S., is especially problematic for younger patients because of the longer time span during which the chronic condition can cause debilitating pain, stiffness and limited mobility. "How can we make the 60 percent have the same long-term outcome as the 40 percent?" asked Halilaj, who is working to understand the difference between those who do and those who do not develop osteoarthritis following knee trauma. Matthew Walker is a Professor, Astrophysics & Cosmology at Carnegie Mellon University. His research focuses on the astrophysical properties of dark matter, but he is also a former collegiate D1 baseball player and lifelong, passionate fan staying apprised of advancements in the game. View his profile Carnegie Mellon University physics professor Matthew Walker said the system still has limitations, especially on pitches that are extremely close to the edge of the strike zone. "Every measurement device has a margin of error," Walker said. "ABS is, from what I can tell, somewhere around half an inch -which means if the ABS call says that the pitch was within half an inch of the border between a ball and a strike, whether it says it’s a ball or a strike is really no better than a guess." Walker said that in those situations, the umpire’s original call should remain in place rather than letting the automated system make the final decision. The influence of science in sports extends far beyond professional athletics. Research developed for elite competitors often finds applications in healthcare, rehabilitation, education, workplace performance, and everyday decision-making. As teams continue to invest in analytics, wearable technology, artificial intelligence, and performance science, the relationship between research and sports is expected to grow even stronger. The result is a deeper understanding of how humans learn, adapt, compete, and perform under pressure. If you're covering or looking to know more, we can help! Carnegie Mellon University experts can discuss: The growing role of science and technology in sports Performance optimization and decision-making under pressure Artificial intelligence and data analytics in athletics Injury prevention, rehabilitation, and athlete health The future of sports research and innovation

Ron YurkoEni HalilajEric Yttri

Media

Social

Industry Expertise

Sport - Professional
Financial Services

Accomplishments

Media and Education

2026-04-06

My work has been featured in popular media outlets such as The Athletic, FiveThirtyEight, The Wall Street Journal, and The Washington Post. I am a three-time degree holder from Carnegie Mellon: with a bachelors, masters, and PhD in Statistics. I also have industry experience in both finance and professional sports.

Statistical Methods in Sports Analytics

2026-04-06

I am actively working on a textbook titled Statistical Methods in Sports Analytics (to be released in 2027), and occasionally write a newsletter called Statistical Thinking in Sports Analytics.

Education

Carnegie Mellon University

Ph.D.

Statistics

Carnegie Mellon University

Masters

Statistics

Carnegie Mellon University

Bachelors

Statistics

Articles

Explainability and Analysis of Large Language Models via Evolutionary Methods

arXiv preprint arXiv:2605.02930

Shannon K Gallagher, Swati Rallapalli, Tyler Brooks, Chuck Loughin, Michele Sezgin, Ronald Yurko

2026-04-27

Evolutionary methods have long been useful for analysis and explanation in genetics, biology, ecology, and related fields. In this work, we extend these methods to neural networks, specifically large language models (LLMs), to better analyze and explain relationships among models. We show how relating weights to genotypes and output text to phenotypes can improve our understanding of model lineage, important datasets, the roles of different model layers, and visualization of model relationships. We demonstrate this in a controlled experiment, where our estimated evolutionary trees reliably recover the topology of the ground-truth training tree. We further identify the most important weight layers according to weight differences and show through phenotypic experiments that one training dataset appears to contribute more useful information than the others. Finally, we generate an unsupervised evolutionary tree of black-box foundation models. Throughout, we provide visualizations that support a clearer understanding of evolutionary relationships among LLMs.

View more

Swinging, Fast and Slow: Interpreting Variation in Baseball Swing Tracking Metrics

The American Statistician

Scott Powers, Ronald Yurko

2026-04-14

In 2024, Major League Baseball released new bat tracking data, reporting swing-by-swing bat speed and
swing length measured at the point of contact. While exciting, the data present challenges for their interpretation. The timing of the batter’s swing relative to the pitch determines the point of measurement relative to the full swing path. The relationship between swing metrics and swing outcomes is confounded by the batter’s pitch recognition. We introduce a framework for interpreting bat tracking data in which we first estimate the batter’s intention conditional on ball-strike count and pitch location using a Bayesian hierarchical skew-normal model with random intercept and random slopes for batter. This yields batter-specific effects of count on swing metrics, which we leverage via instrumental variables regression to estimate causal effects of bat speed and swing length on contact and power outcomes. Finally, we evaluate the tradeoff between contact and power due to bat speed by modeling a plate appearance as a Markov chain. We conclude that batters can reduce their strikeout rate by reducing bat speed as strikes increase, but the tradeoff in reduced power approximately counteracts the benefit to the average batter.

View more

Exploring the difficulty of estimating win probability: a simulation study

Journal of Quantitative Analysis in Sports

Ryan S Brill, Ronald Yurko, Abraham J Wyner

2026-03-26

Estimating win probability is one of the classic modeling tasks of sports analytics. Many widely used win probability estimators use machine learning to fit the relationship between a binary win/loss outcome variable and certain game-state variables. To illustrate just how difficult it is to accurately fit such a model from noisy and highly correlated observational data, in this paper we conduct a simulation study. We create a simplified random walk version of football in which true win probability at each game-state is known, and we see how well a model recovers it. We find that the dependence structure of observational play-by-play data substantially inflates the bias and variance of estimators and lowers the effective sample size. Further, to achieve approximately valid marginal coverage, win probability confidence intervals need to be substantially wide. Concisely, these are high variance estimators subject to substantial uncertainty. Our findings are not unique to the particular application of estimating win probability; they are broadly applicable across sports analytics, as myriad other sports datasets are clustered into groups of observations that share the same outcome.

View more