Hey everyone gather around, I have a secret. You ready? Everyone is going to regress. Next time you hear or read that someone is going to regress, say “Duh, tell me something I don’t know.”
Here’s the deal. Perhaps you consider this picking semantic nits, but regression has become so cliché that its true meaning has been lost. It’s the fantasy equivalent of the telephone game. As the message is passed along, it gets a little more muddled to a point the original message is transformed into something completely different, often comical. I’m not saying we’re at that extreme, but the connotation of regression has changed.
Presently, regress is synonymous with “play worse.” When an analyst says “Kevin Slowey will regress”, most of the time they are simply saying, “Kevin Slowey will play worse.”
Contextually speaking, in its truest sense, regress means revert to or approach a mean. This is why every player will regress. There are distinct elements of player performance that are largely out of a player’s control. It’s these aspects of performance that involve regression. Every player will regress to certain performance means.
The vast majority of regression involves a player off to a fast start due to some good luck. It became conventional analysis to point out such instances, saying the player would regress. Then the telephone game kicked in and any time a player was performing better than expected, for whatever reason, he was in store for some regression and the innate meaning of the word was lost.
Again, perhaps this is more semantics than anything but regression can go in both directions. Players underperforming will also regress. More specifically, certain elements of their performance will regress to their mean. It’s just that underperforming players are likely unlucky. But you hardly ever hear, “Don’t worry, Billy Butler will regress and be fine.” All you hear is, “There’s no way Shin-Soo Choo keeps this up, he’s going to regress.” Both are true.
Another misperception is all luck evens out. Note the definition is revert to or approach a mean. More often than not good or bad luck does not even out but rather going forward, performance is at the expected mean which slowly moves the actual mean closer and closer to what is expected, but the initial good or bad luck prevents it from actually reaching it.
By this point, you are likely aware of the acronym BABIP, which stands for batting average on balls in play. You may not be as familiar with DIPS theory. DIPS is short for defense independent pitching statistics, which was discovered by Voros McCracken and revolutionized the manner we look at the numbers. DIPS theory revealed that the BABIP of pitchers is largely out of their control. When a round ball meets a round bat, happenstance prevails.
For those unaware, the actual calculation of BABIP is (hits-HR)/(AB-hits-strikeouts). Homers and strikeouts are usual elements of batting average, but since the defense is not involved in either, by removing them from the standard batting average calculation, what’s left are only the batted balls the defense had an opportunity to field. McCracken observed that regardless of the pitcher, their BABIP was remarkably similar and clustered close to .300. Since McCracken first made his ground-breaking discovery, data collection and analysis has been refined so we have a better handle on the luck versus skill element of BABIP, but the fact remains that there is still a ton of luck involved with batted balls.
Tying things together, regression almost always refers to BABIP. Recall that in its truest form, regression is a reversion to a mean. BABIP is the mean. As McCracken discovered, pitchers revert to a mean BABIP which is the league average. Hitters develop their own individual baseline so they don’t revert to the league average. They regress to their career averages.
For many years, BABIP was looked at as a singular metric. If a hitter was sporting a number above his career norm, he was considered lucky and was due for regression and a drop in average. If his BABIP was lower than usual, he was unlucky and was also due for regression leading to a higher average.
The first major refinement to BABIP was breaking batted balls into groundballs, fly balls and line drives. Each was discovered to have its own BABIP. About 72 percent of line drives go for hits, 23 percent of groundballs are hits while only 13 percent of non-home run fly balls land safely. One reason a batter could sport a BABIP higher than league average is if he hits more liners than average. Two players with the same line drive rate can have different BABIP’s if they hit different amounts of ground balls. The more grounders hit, the higher the BABIP.
Of course, the BABIP on grounders will be different for Juan Pierre and Jose Molina. Faster players usually sport a higher BABIP on ground balls.
Breaking infield pop-ups from fly ball data also helps define BABIP. The BABIP of a pop-up is almost .000. A player hitting a lot of pop-ups can have a lower BABIP.
The next refinement on batted ball data was to delineate how hard a ball was struck – breaking it into hard, medium and soft. As you might expect, hard hit balls, irrespective of type, have a higher BABIP than similarly classified medium or slowly struck batted balls. Perhaps curiously, slow hit balls have a slightly higher BABIP than medium, though this is fuzzy since we’re broaching on a very subjective area. On the other hand, slowly hit grounders have a better chance of being beat out while weakly hit fly balls may fall in front of the outfielder. Chances are the defensive player will be able to get to a medium hit grounder or fly and make the play so maybe the result is not curious after all. The take home lesson is how hard a batter strikes a ball will also impact his BABIP. The more hard hit balls, the higher the BABIP.
For what it’s worth, the next major improvement in data collection will be removing the subjectivity from both trajectory determination and how well the ball was hit. Soon, both of these will be measured electronically which will help to normalize the data by removing subjective bias.
Most of this discussion has revolved around hitters though BABIP is just as relevant in pitcher evaluation. You have probably already figured out that one way for a pitcher to have a lower BABIP is if he is a fly ball pitcher. Of course this leads to more home runs, but the hits on balls in play are fewer than for a groundball guy. The gray area is line drive rate. The reason all pitcher’s BABIP’s hover near .300 is they have limited control over line drive rate. This may be counter intuitive but it’s backed by, ironically, regression analysis.
Next week we’ll take a look at some hitters that are regression candidates and go through their component BABIP’s in an effort to project how they will perform going forward. We’ll do the same for pitchers the following week. We’ll conclude today with some quick observations.
If you play the daily games, the past couple of weeks have been gold for anyone paying attention. With all the postponements and bad weather, the amount of dead money has been borderline ridiculous as many lineups have dead spots. Also, early in the season is when these games are recruiting new players so the knowledge and experience level is below what it will be later in the season when those that are not winning cease to play. We may be past the weather issues but we’re still at the point you can take advantage of the lesser quality playing field before it gets weeded out.
Along those lines, if you play in a regular fantasy league with daily moves, this is the time of season to really pound up those counting stats by making sure you have as many active players on Monday and Thursday as possible. Most leagues have reserves and most reserves are dedicated to streaming pitchers, but I have personally found the best strategy is to have a rotation of a few good starters with the rest being relievers with exceptional strikeout rates and not worrying about streaming -- yet. Instead use the reserves to stash as many position players as possible to maximize at bats on travel days since the quality of these reserves is better than they will be later in the season when even more injuries force these reserves to be regulars. Then later in the year, you can start folding in more starting pitchers to make up for those lost starts.
This is especially effective in leagues with an innings cap since the quality of innings provided by the relievers is better than you’d get from streaming starters. Also, keep in mind that if your league does have an innings cap that you will readily attain, the strikeout category is better thought of as K/9. This renders the low strikeout hurlers as lesser streaming options since you’re eating up limited innings with an arm likely to not deliver whiffs.
Todd Zola has been with Mastersball since its inception in 1997, presently serving as Managing Partner in charge of the Platinum subscription content. Lord Zola, as he is affectionately known in the industry, also contributes to KFFL, the ESPN Insider and is a frequent guest on SiriusXM Fantasy Sports radio. He’s a veteran of Tout Wars and LABR and has won multiple National Fantasy Baseball Championship titles. Follow Todd on Twitter @ToddZola
Rounding The Bases - Todd Zola Looks At The First 10 Rounds of a Likely 2014 Fantasy Baseball Draft
Master Fantasy Baseball Notes From Todd Zola - Looking Ahead To The 2014 Fantasy Baseball Season
Rounding The Bases - Master Fantasy Baseball Notes From Todd Zola
The Farm Report: Minor League Baseball Coverage From A Fantasy Perspective (20 Prospects for '14)
2013 Fantasy Baseball Waiver Pickups: Week 24
The Farm Report: Minor League Baseball Coverage From A Fantasy Perspective (September Call-Ups)
2013 Fantasy Baseball Waiver Pickups: Week 23
Rounding the Bases: A Day For the Little Man
Rounding The Bases: Morse, Willingham On the Move?
The Farm Report: Minor League Baseball Coverage From A Fantasy Perspective (Roster Expansion)
2013 Fantasy Baseball Waiver Pickups: Week 22
Rounding the Bases: Five Surging Pitchers & Hitters
Rounding the Bases: Rookies, Veterans & Roster Movement
The Farm Report: Minor League Baseball Coverage From A Fantasy Perspective (Pitching Help On The Way!)
2013 Fantasy Baseball Waiver Pickups: Week 21
Master Notes From Todd Zola - Taking A Look at Starting Pitching In Keeper Leagues
Rounding the Bases: Small Names, Big Production?
Rounding the Bases: A First for Stephen Strasburg and the (Near) Final Round-Up
The Farm Report: Minor League Baseball Coverage From A Fantasy Perspective (Hitting Help On The Way!)
2013 Fantasy Baseball Waiver Pickups: Week Twenty