The world of sports betting requires handicappers to constantly adapt and evaluate. Things that worked 10 years ago don’t work as well today. Things that worked five years ago don’t work as well today. There are some concepts and philosophies that will always work, but have also been accounted for by the oddsmakers and by other bettors that probably get to the market faster.
Fortunately, we live in an era when there is a lot of new data and people are coming up with new ways of breaking down the games and the players that play them. It creates so many opportunities for fans and bettors to look through different lenses in order to see things with a unique perspective.
When it comes to baseball, specifically, that perspective for me comes in the form of sabermetrics. Over a decade ago, I wanted to know how my favorite team, the Cleveland Indians, makes decisions and values players. So, I researched. And I studied. I won’t say that I was early to the party, but I certainly didn’t arrive all that late. It is something that has been a huge part of my fandom and my handicapping ever since.
There continues to be a pretty sharp divide between old school and new school when it comes to player evaluation. A lot of fans cling to stats like batting average and RBI, while others proselytize about wRC+ and wOBA. While I certainly believe in the advanced metrics far more than the traditional ones, I hate the holier-than-thou attitude some in the stats community have. The best way to induce change and induce a new way of thinking is to reach common ground. I feel like that is part of my skill set when it comes to talking about baseball and I’ll try to do that today with some of the stats that I will reference heavily in my MLB season win totals.
First, we’ll start with some team-oriented metrics:
Pythagorean Win-Loss – Most people don’t care about what should have happened. They only care about what actually happened. Pythagorean Win-Loss records can be a good barometer for teams in line for regression. It is a standings metric based on run differential. Teams that excelled in one-run games may regress the following season. Similarly, teams that got blown out a lot should have better personnel the next year.
As an aside, one-run games are a big deal when it comes to Pyth W-L. Generally, teams are within a couple of games above or below .500 in one-run games. Significant outliers are likely to regress the following season.
BaseRuns – I like to refer to the BaseRuns standings from Fangraphs as well. BaseRuns takes all of a team’s outcomes for and against and removes the context. Think of it like this: If a team goes HR, 1B, 1B, K, K, K, that team scores one run that innings. If a team goes 1B, 1B, HR, K, K, K, that team scores three runs that inning. The same six individual outcomes produced two very different outcomes overall. BaseRuns eliminates that element of randomness and spits out a measure of runs per game and runs allowed per game given all of the individual outcomes. Then, it produces a win-loss record.
3rd Order Win% – An alternate standings metric at Baseball Prospectus that looks at a team’s strength of schedule and “underlying statistics” to come up with a win percentage based on the team’s performance.
Now, some individual stats for hitters and pitchers:
wOBA – wOBA is my favorite statistic of all. It stands for Weighted On-Base Average. Unlike its predecessor, on-base percentage, wOBA assigns a weighted value to each way of reaching base. When it comes to on-base percentage, there is no distinction between a single or a home run. wOBA has changed that. The weights of the outcomes are assigned based on the offensive climate around Major League Baseball. For example, the weight of a home run was 1.98 runs, the lowest mark since 2007. Walks, however, were the highest since 2011 at .693 runs.
Quite simply, wOBA actually distinguishes between the value of ways of getting on base, thus making it better than most every other offensive metric.
wRC+ – Weighted Runs Created Plus is another popular statistic. The + simply means relative to league average, wherein league average is 100. A player with a 110 wRC+ was 10 percent better than league average when adjustments are made for park factors and the current run environment. Now that we’re in the Juiced Ball Era, or so they say, the run environment is higher, so the baseline is higher. Not that it’s relevant to my win totals or overall handicapping, but you can use wRC+ to compare hitters from previous eras because the stat is adjusted for park factors, leagues, and run environments.
K% & BB% – These seem pretty obvious, but are worth mentioning. These are (Strikeouts / Plate Appearances) and (Walks / Plate Appearances). Those are important stats for hitters, but I will use them more frequently with pitchers.
BABIP – BABIP stands for Batting Average on Balls in Play. Keep in mind that traditional batting average factors strikeouts into the equation because those are at bats. BABIP is a good measure of luck. Hitters with a high BABIP are either making terrific contact, are fast, or are getting lucky. Hitters with a low BABIP either have poor contact quality or are getting unlucky. The same can be said about pitchers in terms of contact quality and luck. Home runs are NOT factored into BABIP because they are not balls in play. Traditionally, the “average” range for BABIP is between .290 and .310, but extreme fly ball pitchers and hitters need to be graded accordingly. Last year, the league average BABIP for hitters was exactly .300.
FIP – I hate ERA. It is such a tremendously flawed statistic. Let’s remember that a pitcher that allows three runs over six innings has a 4.50 ERA, which is widely considered to be bad, but it is a “quality start”. FIP is a better metric and one that I use often. FIP stands for Fielding Independent Pitching. It is a run metric derived from things that a pitcher can “control” – strikeouts, walks, home runs, and hit by pitches.
It takes the defense out of the equation. Bad defenders can really hurt a pitcher’s ERA by not catching balls that should be caught. FIP takes that element out of the equation. Once the ball leaves the pitcher’s hand, he has no control over what happens. It is all subject to variance, especially once it is put in play. This is a far better way to assess a pitcher’s performance.
xFIP – A derivative of FIP is xFIP, which stands for eXpected Fielding Independent Pitching. The difference between FIP and xFIP is that it recalculates the home run portion of FIP by assuming a league average home run to fly ball rate. Last season, obviously, we had a significant number of home runs hit. The HR/FB% league-wide rose from 12.8 percent to 13.7 percent. It was just 11.4 percent in 2015. The relevance of that will become clearer as we move forward.
The important takeaway here is that we can use xFIP in a similar context to BABIP. Sometimes pitchers are getting unlucky with fly balls that hit a jet stream or just keep carrying. Other times, they are simply making bad pitches. Like any statistic, we have to dig deeper to find out the root cause, but xFIP is a good predictor of future performance. It eliminates some of the noise of small sample sizes.
Pitchers with high ERAs that have lower FIP and xFIP marks are generally pitchers to circle for positive regression. Pitchers with low ERAs that have higher FIP and xFIP marks are likely to regress negatively. There are always outliers, and I’ll discuss them in the season previews and on a day-to-day basis, but keep that in mind.
SIERA – SIERA stands for Skill Interactive Earned Run Average. This is how we get a little bit deeper. If you find FIP to be too oversimplified because it doesn’t take into account whether a pitcher is more of a ground ball guy or a fly ball guy, this is for you. Think about it. Ground ball pitchers will have a higher HR/FB% because they have a smaller sample of fly balls. Fly ball pitchers will generally allow more home runs, but they also allow more fly balls, which will cut into the HR/FB%. Pitchers that allow a lot of line drives are going to give up more hits. Hard contact is a bad thing, no matter how good the defense is. SIERA is probably the best ERA estimator we have, though there are some very good ones at Baseball Prospectus.
The big thing about these stats is that they carry predictive value. That’s what we’re looking for. We’re not looking for what happened in the past. We’re looking for what will happen in the future.
DRS – DRS stands for Defensive Runs Saved. One of the last great frontiers to explore for baseball stat geeks like me is defense. Errors are a poor stat. They only count if the fielder gets to the ball and are based on subjective discretion by the official scorer. Because fielding percentage uses errors, it is also a poor measure of defensive ability.
DRS is somewhat complex. Players are graded on a plus/minus scale, where zero is average. It is measured on location of a batted ball, type of batted ball, and a general description of the speed of the ball. All of these plays are catalogued and a baseline is set. If a ball has a 70 percent catch probability and the fielder fails to make the play, that accounts for -0.7 defensive runs saved. If that play is made, the fielder gets +0.3 defensive runs saved.
Now that we have Statcast data, we’re getting a much clearer picture of defensive metrics based on Catch Probability, Sprint Speed, and a variety of other metrics. Fans and writers have merely the tip of the iceberg when it comes to Statcast data, but it is very important.
Statcast is remarkable. Those that really want to go down a rabbit hole of baseball statistics are going to fall in love with the data. It is something that I have studied a lot and can be found at BaseballSavant.com.
Here are some of the Statcast metrics I’ll be using:
Exit Velocity –We think of pitchers with high BABIPs as “unlucky”. We think of pitchers with high ERAs and low FIP and xFIP marks as “unlucky”. Well, now that we have exit velocities to factor in, we can see if pitchers are simply getting hit really hard. Balls that are hit harder are more likely to go for doubles, triples, and home runs and are also likely to be tougher plays for fielders to make.
I often refer to the “percentile” that the player fell into with regards to Exit Velocity or Hard Hit%. A 97th percentile Hard Hit% means that the player ranked in the top 3% in Hard Hit%, which is percentage of batted balls with an exit velocity of 95+ mph.
xwOBA – wOBA – Statcast does its calculations based on batted ball distance, launch angle, and exit velocity. Using that data, it can estimate hit probabilities, including whether or not balls should be home runs, doubles, singles, etc. xwOBA stands for eXpected wOBA. xwOBA – wOBA is a good indication of pitcher luck. This is a stat I will be utilizing a lot more this season. It is a measure of the gap between a pitcher’s expected wOBA and actual wOBA against. It can work for hitters as well and will be something I utilize in my DFS pieces.
xBA – BA – This is a similar stat. This is eXpected Batting Average minus actual batting average. Pretty simple and straightforward.
Barrels & Barrel% & Barrels/PA% – Barreled balls are a recent development at Statcast to suggest batted balls with a very high likelihood of positive outcomes. A “barreled ball” as defined by Statcast using primarily launch angle and exit velocity is a ball with a xBA of .500 and a xSLG of 1.500. More often than not, those would be doubles, triples, or home runs.
Pitchers that allow a lot of barrels or a high Barrel% have bad command. Hitters that have a lot of barrels or a high Barrel% would grade well in exit velocity, launch angle, and contact quality as a whole.
As far as my MLB analysis goes, these will be the most popular terms and statistics, so I certainly encourage readers to familiarize themselves with these concepts. I will also use additional PITCHf/x and Statcast data.
I would encourage readers to check out more detailed definitions at FanGraphs with the Offensive Stats and the Pitching Stats.