Paul Smith: A Data Architect’s Journey into Sports Betting and Racing Analytics
In the realm of sports betting and racing, there’s a man whose two-decade-long expertise in software design, architecture, development, and testing has propelled him to new heights.
Meet Paul Smith, a seasoned data architect. Before this, he had a successful career as a consultant in the data industry.
However, his latest venture involves applying his knowledge on sports and racing, and he’s sharing his insights with the world through his personal YouTube channel.
DATA MODELLING
Paul’s journey begins with the meticulous process of data collection and data modelling.
Data modelling in the context of betting involves organising historical data in a format to make it efficient and useful for constructing predictive models, allowing the forecasting of future outcomes.
As he puts it, “I find the best sources of data, statistics and ratings and start collecting it for analysis.”
This foundational step fuels his approach to predictive modelling.
“From there, I build and use models to determine what is the probability of an outcome happening. I’d have about 900 models in production.” Smith says.
Smith’s modelling is not merely rooted in historical data; it extends to the key specifics that can influence racing outcomes or sporting events.
“I have models that will try to predict the weather, who might get scratched in races, where horses will be placed, or team rosters” he explains.
But what sets him apart is his keen eye for spotting the quirks in the Betfair Starting Price (BSP), where personal bias often lurks among recreational punters.
“I am someone who hits the BSP (Betfair Starting Price) a lot,” he asserts. “In most markets, you can find edges based on personal bias from recreational punters.”
The Keys to Approaching Races and Sporting Events: Timing and Analysis.
Smith has a carefully calibrated approach when it comes to analysing racing and sporting events. He’s all about seizing the right moment.
“I’ll look at all races about five minutes out (greyhound, harness, and thoroughbred), it doesn’t matter what it is or what country, even if it doesn’t have great odds” Smith said.
For sports, Smith’s process starts three hours prior to the event, where he meticulously watches the overrounds.
(“Overrounds” is a term to describe when the betting percentage of a market is above 100% and therefore provides an advantage to the bookmaker and the opposite to the bettor.)
“(I) See what tips are being put out, put some analysis on it, and work out what might be an over bet or an under bet.”
He thrives on uncovering opportunities driven by human bias and market misalignment.
Smith also points out certain sports where the historical context often gets neglected when framing the market. He highlights Test Cricket as an example.
“Often the odds don’t reflect the magnitude of a challenge.” Smith said.
For instance, India needed a world record score to beat Australia in the World Test Championship, yet the odds were around $4.”
Smith’s models are designed to uncover such anomalies.
Recognising Bias: A Fundamental Aspect
For Smith, a key aspect is understanding human psychology and the biases it entails.
“Human psychology is the key. Everything I do is based on human bias from others.”
He knows that people tend to bet on certain teams more than others and are influenced by recency bias. These insights help him fine-tune his approach and exploit BSP opportunities.
Scale and Risk Management: The Mental Game
Managing risk is a crucial aspect of Smith’s strategy, and his background in risk assessment is instrumental. He emphasises the significance of mental fortitude when dealing with sums of money and unpredictable outcomes.
“For me, it’s all mental,” he reveals.
“When you have a larger bet size, and you have cricket matches that run for five days, protests in racing, and other delays, you can be sitting there waiting on results and this can be stressful. It’s all mental, and that’s important.”
Smith doesn’t rely on massive bets; instead, he endeavours to accumulate small margins across a multitude of markets in all types of racing and sports.
Unlike recreational punters seeking big wins on small bets, Smith takes a measured approach, consistently identifying opportunities over time.
Paul Smith’s journey is a testament to the fusion of data science and the art of betting. His meticulous data modelling, awareness of human biases, and disciplined risk management provide a fresh perspective on the intricate world of sports betting and racing analytics.
Paul will also be participating in Betfair’s 2023 Spring Racing Data Modelling Competition on Victorian Spring Racing.
Betfair is offering a $50,000 prize pool in their competition, where entrants are supplied with historic thoroughbred data from Punting Form.
Participants are required to create their own predictive model to estimate the winning probability for selections in Saturday Melbourne Feature race meetings starting on Cox Plate Day (October 28, 2023) and concluding on November 18.
Registration closes at 4:59pm AEDT on October 23, 2023. Enter here (https://www.betfair.com.au/hub/datathon) Terms and Conditions Apply.