October 9, 2020
When we talk about serious futures investments here at ScoreMetrics, we always end up discussing models when assessing opportunities. So, what do predictive models mean in practice for sports traders and why should you care about them?
Glad you asked, because that’s the topic of today’s article!
What are predictive models in sports?
The name kind of says it all – predictive models use various factors to determine probabilities for future events. Often in the case of the major sports leagues, this means predicting championship or playoff win probabilities, or the outcomes of single matchups.
Many serious sports investors develop their own proprietary models in search for profits in the market, and we do this as well here at the ScoreMetrics Lab. But there are also many publicly available models that serve as great examples of what the whole topic is all about.
To make things concrete, let’s explore what FiveThirtyEight and the Ringer have to say about their popular NBA playoff predictions.
Starting with FiveThirtyEight, whose models are well respected and have been historically pretty accurate:
”These forecasts are based on 100,000 simulations of the rest of the season. Our player-based RAPTOR forecast doesn’t account for wins and losses; it is based entirely on our NBA player projections, which estimate each player’s future performance based on the trajectory of similar NBA players. These are combined with up-to-date depth charts — tracking injuries, trades and other player transactions — to generate talent estimates for each team. A team’s full-strength rating assumes all of its key players are in the lineup. A team’s current rating reflects any injuries and rest days in effect at the moment of the team’s next game. Elo ratings — which power the pure Elo forecast — are a measure of team strength based on head-to-head results, margin of victory and quality of opponent.”
This is what the Ringer say about their NBA playoff model:
”Based on team strength as measured by point differential—a historically strong predictor of future success—with adjustments for individual wrinkles like player absences and injuries, the Restart Odds include four main features: single-game odds for each day’s games, an explanation of the importance of each individual game, postseason odds for all 16 playoff teams leading up to the championship round, and series possibilities that show how many games each matchup is likely to run.”
Neither is revealing the exact recipe for the sauce, but we can get a pretty decent overview from these descriptions on what the models are all about.
A concrete example from the NBA playoffs
Let’s take a simple example to demonstrate how an investor can utilize a predictive model to find value gaps in the market – in other words, opportunities where the odds are favorable to us according to a model.
During this year’s NBA Conference Finals, by the time when the Miami Heat were up 3-1 against the Boston Celtics and the Los Angeles Lakers led their series against the Denver Nuggets 2-1, the Heat were listed at +350 odds to win the NBA championship.
That +350 means a 22.2% implied probability from the bookmakers’ side for the Heat to win it all.
At that point in time, the Ringer’s predictive NBA playoff model gave Miami a 35% probability to win the championship, while FiveThirtyEight’s Elo forecast gave the Heat a 41% chance to lift the trophy, and their RAPTOR player rating forecast had them all the way at 63% (Miami was the favorite for the title according to both FiveThirtyEight predictions).
If you found these models to be trustworthy, the value gaps were pretty clear and significant:
Implied probability (+350 odds)
According to these models, Miami to win the title at +350 would be a pretty neat investment opportunity.
Why rely on models?
When assessing futures investment opportunities, a smart investor has to have a hypothesis in place for finding value that’s based in facts, not a gut feeling. That’s what separates us from gamblers in the market.
When our head trader John Todora gave out his recommendations for investing in the 2020 NFL Draft, this is what he wrote:
”The idea isn’t to gamble, it’s to invest in patterns that give you a high probability of achieving profits.
So what I did was build a system around the draft. I did so by developing ten models, based on backtested historical data, and then I used those systems (which were based on patterns we found) to invest.”
Those elaborate predictive models ended up in John recommending his newsletter readers to invest in four high-value prop opportunities for the draft that ended up returning a 65% profit on the investment.
Sure, you can find odds in all futures markets where a single investment would return 1000% or more. But for smart investors, it’s all about assessing the expected value of the moves one can make and to pick up the ones that fit their strategy the best. And that’s exactly what John did for the NFL Draft.
To take advantage of the return of sports and to learn everything you need to know about smart investing in the sports betting market, check out our head trader John Todora’s new book – “Zero Correlation Investing – The Score Metrics Secret”. It’s currently on sale for a limited time, so go get yours now!
The ScoreMetrics Lab is the engine that runs the Sports Trading System operation, consisting of a team of researchers and writers who are constantly testing and retesting algorithms. They work hand in hand with our Head Trader and Creator of ScoreMetrics, John Todora to help find new breakthroughs and develop new systems.