When spring comes into view, our focus splits between increasingly heated NCAA basketball action and the prospects of a profitable new MLB season. Spring training gives us insight into each team’s season outlook while building excitement for over 2,400 games and even more wagering opportunities. Wait, over 2,400 games?!? This massive number of games spanning over half a year presents more than just plentiful opportunities for profit; it is a test of discipline, sound handicapping skills, and mastering repetition.

Our off-season preparation generally consists of making team adjustments and refining/streamlining our handicapping model. Rating adjustments allow us to adapt last year’s team profiles for Opening Day so we can hit the ground running. But what do you adjust and how? With an overwhelming array of baseball statistics available at your fingertips it can be difficult to sort through the mess and pinpoint what can help you reliably pick winners.

Statistics and Modeling

For the start of the MLB 2014 season, we ran a rudimentary model that used player and aggregate WAR (wins above replacement) factors to determine which team was stronger for a given game. This approach was fueled by Joe Peta’s book Trading Bases. Peta intermixes a personal tale of reinventing himself after tragedy with insights into his unique handicapping angle (if you have no interest in the personal tale, those chapters can easily be skipped). Although his approach ultimately did not work with our skill set and knowledge base, our inaugural baseball betting season gave us plenty of feedback to build upon.

Instead of relying solely on WAR from sources like Fangraphs.com and Baseball-Reference.com, we dug deep into the internet for insight. It didn’t take long to see that baseball has been the subject of many research papers for years, resulting in a vast pool of available statistics. Years later, we sit on a model that uses key predictive statistics but still requires a handicapper’s feel for how any given game could play out (e.g. how long the starting pitcher will last, bullpen usage, etc.).

DIY Resources

Two true powerhouses of MLB statistics are the aforementioned Fangraphs.com and Baseball-Reference.com. Sites like ESPN.com and MLB.com are stuck in the minor leagues when it comes to the insane amount of statistics collected for this game. You can easily argue that baseball is one of the most over-analyzable sports out there, in part because of the regimented segments of gameplay (pitch, AB, and inning) and how statistics can be looked at situationally. The strength of these websites lies in the breadth of available information and the ability to manipulate and customize the data.

The insane array of raw data and prepared metrics (i.e. data mashed with and/or compared to other data) requires you to focus on the factors that may determine a game’s outcome. Since 2015, our plan of attack has been to run a model that blends these predictive data with game-flow handicapping to project a probable outcome. Handicapping models can take many forms and  encompass as many or as few metrics as you desire…or, most importantly, can effectively manage. For MLB handicapper-bettors, a strong model that works for you is the “secret sauce” allows achieving long-term wagering success and bankroll growth. At BetCrushers, our methodology is based on how effectively teams get runners on base and how efficiently they are converted to runs. Conversely, we account for how well pitching and defense keeps runners off base and runs from scoring.

Need Reps? No Problem

Managing data and staying up-to-date with the current composition and trajectory of 30 MLB clubs demands a disciplined daily process to be successful. Because we’re talking about an everyday sport whose season stretches over six months, this becomes a test of mental and emotional stamina. All seasoned bettors know there are plenty of peaks, valleys, hot streaks, cold streaks, and give-some/take-some periods throughout a season. It takes mental focus and a strong stomach to keep coming back for more.

Creating and sticking to a disciplined, repetitive process is the foundation of our MLB work. How often do you need to refresh the data inputs that comprise team profiles? How do you integrate recency into a much larger sample size (think how diluted season-long data can be in mid-August) to reach the perfect blend? Experience and a little trial and error balance these and other practical considerations with your daily life in a short 24 hours.

Good habits are essential to making it in the long game regardless of how robust your data-crunching capability is, the amount of time you have in a day to dedicate to handicapping, or what bet size your bankroll can handle. Develop a routine in which you update data and metrics at regular intervals, analyze the slate of games during the same part of each day, identify lines that present betting opportunities, and track daily results. This routine helps break up the workload into manageable chunks, making the MLB grind less daunting and more a part of your daily life.

Feedback is Key to Success

One of the toughest things to get under control is how to bet the outcomes produced by the model. A common issue that aspiring handicappers face is the challenge of turning information into profitable action. It requires a lot of feedback – like maintaining a log of your projections alongside game results – and scouring this information to find underlying angles that can guide your betting approach.

We spent our first few MLB seasons getting a feel for the right handicapping style to suit our strengths. Peta’s WAR and cluster luck approach was not for us, so we shifted toward the approach we use today. That transition was possible because of the game results and handicapping comparison feedback we gathered in those early seasons. Model building should be an iterative process (especially in the early development stage) that builds on strengths and eliminates or minimizes weaknesses.

Once a handicapping model/process is established, analyzing the results can identify betting advantages. Different handicapping approaches can exploit the betting markets in different ways. Feedback from large samples can make this evident even if the intended purpose of the model/process is initially unclear. Totals, money lines, run lines, and first 5 innings are all viable ways to produce long-term profit. Rarely can you have it all though, so seek out strengths and exploit them. For example, our model’s outputs correlate to money line underdogs and run line favorites in a range of scenarios. This works for us, but only you can determine what will help you achieve success through hard work and determination.

Conclusions

• MLB is one of the most grueling seasons to successfully handicap.
• Create or find a model that uses predictive statistics suiting the way you understand the game of baseball.
• Stick to a disciplined process that keeps your data fresh and provides important performance feedback.
• Align your wagers to the strengths of your process.
• Visit our MLB page to access free daily picks, analysis, and insights.