2018 MLB Draft Guide: Stat Splits That Matter
Steve Pimental takes an insightful look into which splits matter most with regard to MLB players and how fantasy baseball owners should use them.
There is nothing worse than watching a baseball game and seeing a graphic on the screen or hearing the broadcaster quote a stat that is completely meaningless. “The Twins have scored in at least three different innings in 62.2% of their games (46/74) since the All-Star Break.” Who cares? It doesn't tell us anything useful. Even worse are the stats that appear to tell us something useful but in fact could be completely misleading, like “Lonnie Chisenhall had a .967 OPS against lefties in 2017 and a .857 OPS against righties.” That is technically true, and it could be part of a larger trend, but it is far more likely just noise. If you start making fantasy decisions based on the idea Lonnie Chisenhall is better against left-handed pitchers, you are likely setting yourself up for failure.
That is why understanding stat splits is so important. Breaking down stats into smaller parts is an important way to find value that might otherwise be hidden. Not all .300 hitters are created equal, and splits can uncover important differences. At the same time, the more you split the numbers, the more you run into problems of small sample size. Drawing any conclusions based on a single year can be dangerous in and of itself, but when you start slicing and dicing a individual seasons into smaller parts, the results become far less reliable.
One of the most important ideas for evaluating any sort of statistic is to put it in its proper context. This is especially true for splits. How big is the sample? What else has changed in addition to the split? If we are looking at splits for a specific hitter, we need to know if he changed teams, or leagues, or if he changed his swing. All of those factors can influence how we evaluate a given split.
One of the most important rules for evaluating splits is to give more weight to multi-year splits. To go back to the Lonnie Chisenhall example, 2017 was the first year since 2011 that he had a higher wOBA against lefties than righties. His career wOBA is still better against righties even after his eye-popping 2017 season. If Chisenhall is better against lefties in 2018 we will probably have to reevaluate his fantasy value, but until then, it is probably best to continue to regard him as a platoon player.
When you are looking at splits, the stats that you use matter. In the example above, there is a decent chance Lonnie Chisenhall’s OPS or wOBA against lefties and righties last season were heavily influenced by luck. That being said, he had a career-high 15.0 percent walk rate against lefties last season, exactly double his 7.5 percent walk rate against lefties for his career. The increased walk rate could still be a small sample fluke, considering he only had 60 plate appearances against lefties in 2017, but I feel more optimistic his 15.0 percent walk rate is sustainable compared to his .400 BABIP against lefties.
One of the nice things about splits is they apply to both hitters and pitchers. Just as we examined Lonnie Chisenhall’s splits against lefties and righties, we can look at how pitchers fare against right and left-handed batters. If a pitcher really struggles against left-handed hitters, or in a certain ballpark, or in day games, that knowledge can be just as useful as it would be for a hitter.
With that in mind, let’s take a look at the nine most important types of splits for fantasy players. Each set of splits can offer insights that can help astute fantasy players gain an advantage on draft night and throughout the season, but they all have to be put in the proper context. After that, we will look at a few splits than can simply be ignored outright.
When most people think about splits in baseball, the first thought is usually lefty-righty splits, like our Lonnie Chisenhall example from earlier. Lefty-righty splits inform decisions not just for seasonal and daily fantasy players but also for MLB front offices and managers. If you don’t think splits against righties and lefties matter, just look at the way teams are valuing platoon players or relievers who are effective against batters from either side of the plate. A good rule of thumb is that if MLB teams think something is important, fantasy players probably should as well.
This is pretty much common knowledge at this point, but it bears repeating: Most pitchers have an advantage against same-handed batters, and vice versa. This is mostly due to the way a breaking ball breaks away from same-handed batters. For rookies or other batters with limited experience, it is often useful to assume they will be better against opposite-handed pitchers until proven otherwise.
Don’t forget to look at minor league splits as well, especially when it comes to lefty-righty issues. If someone like Michael Conforto demonstrates he can hit lefties in the minors (at least adequately), he probably can in the majors, too, even if his manager is reluctant to give him the opportunity. Identifying players who are currently in a platoon but could hold their own in a full-time gig is a great way to find fantasy value.
Another place to find value is in reverse splits, though they come with significant pitfalls. As we discussed above, most right-handed batters are better against left-handed pitchers, sometimes dramatically slow. A right-handed batter who is actually better against a right-handed pitcher is said to have reverse splits.
The problem is that even over multiple seasons, we are still talking about relatively small sample sizes, especially for players who do not play every day. Even a veteran batter who appears to have reverse splits could just be getting lucky. I’m looking at you, Justin Turner.
Coming into 2017, Justin Turner was the poster child for batters with reverse splits. After his first two seasons, when he had 40 total plate appearances, Turner had a better wOBA against righties than lefties in five of his six MLB seasons. Then, in 2017, he had a .484 wOBA against lefties and a .360 wOBA against righties. His career numbers are still slightly better against righties, but they are obviously far closer than they were at this time a year ago. If you only are a DFS player who only uses Turner against right-handed pitchers, you could be making a big mistake.
Reverse splits are far more reliable for pitchers, because there is more they can do to conteract the platoon advantage. Breaking balls from right-handed pitchers will always break away from Justin Turner, but pitchers like Lance McCullers and Charlie Morton can develop pitches like changeups, splitters and cutters to help them get left-handed batters out. Again, it is important to look at multi-year data whenever possible, but there are several prominent pitchers with noteworthy reverse splits.
First Half/Second Half Splits
The other most prominent splits cited both by fantasy players as well as the baseball media are first and second half splits. Similarly to lefty-righty splits, first and second half splits can be both useful and dangerous. Let’s address the dangers first.
One common narrative you will hear throughout Spring Training is “he finished last season strong, he is poised to have a big season this year.” On its face, the narrative makes sense. If a player is doing something differently that allowed him to have success at the end of 2017, and that player continues to do things the same way in 2018, he will continue to have success.
The problem is, there is little statistical correlation between second-half stats in one season and full-season stats the next. It is just as likely that player’s good second half was a fluke, and his “true value” is closer to what he displayed over the course of the entire season. If we know a player changed his approach or his mechanics in the second half, then maybe we can use those stats to make our projections for the next season.
When I look at first and second half stats, I am far more interested in players who display consistent trends year after year. Who are the slow starters who are usually good buy-low candidates in May? Who are the guys like Salvador Perez who usually start well but wear down after the All-Star break? It may be worth moving Perez up your draft board a bit if you think you will be able to sell high in June, while the slow starter may be someone you avoid drafting with the intention of trying to trade for him during the season.
Everything we said about first and second half splits also applies to monthly splits. The sample sizes are even smaller, so we have to view monthly splits with a discerning eye even for veteran players. That being said, there can be meaningful insights to glean from monthly stats. For example, let us look at Madison Bumgarner.
Madison Bumgarner’s first and second half splits are nearly identical (his first half splits are actually better), but he does have a 3.52 FIP for his career in March/April, higher than any other month. Bumgarner was a consensus top-six starting pitcher for fantasy heading into 2017, where he finished with an ERA over 3.00 for the first time since 2012 and made his fewest starts since 2009. If he struggles in April, his fantasy owners may panic. Discerning fantasy players may look at his career numbers in April and determine he is a buy-low candidate, if his velocity is okay and he appears to be healthy. You shouldn’t blindly assume a player with poor April numbers will automatically bounce back from another slow start, but it is probably more likely, given his history. That is a good way to approach all of these splits. Using any splits to make sweeping assertions regarding fantasy value is probably a mistake, but they can nudge the range of likely outcomes in one direction or another.
In general, players—both hitters and pitchers—perform better at home. They get to sleep in their own beds, get dressed in a nicer locker room and enjoy the backing of the home fans. Perhaps most importantly, they don’t have to deal with any of the ill-effects of travel.
That being said, not all players respond the same way to playing at home or on the road. Unless we see a large amount of data, we should probably assume a modest boost in performance at home, but not enough to completely change our fantasy strategies. This is especially true for players who have played for multiple teams. How can you evaluate if a player has a home field advantage if he has called several different fields home?
A starting pitcher who plays for the Rockies probably doesn’t have much of a home field advantage, because Coors Field is so much better for hitters. This is an extreme example, but it illustrates both the dangers of home/road splits and the potential importance of ballpark splits.
Ballpark splits are most useful for DFS players, and even then only for veterans who have spent several years with the same team. We probably shouldn’t care about Marcell Ozuna’s 1.009 OPS at Miller Park considering he has played 13 games there. That being said, it can be useful to take note if Joey Votto has great numbers in Miller Park, since he has a much longer career and has spent all of it in the NL Central, thus getting even more opportunities in Milwaukee.
Batting Order Splits
Andrew McCutchen famously turned his 2017 season around after he was dropped to sixth in the batting order, batting .380/.471/.690 from May 25 through June 25. Maybe he should have stayed there, because while McCutchen stayed hot in July after returning to the third spot in the order, he finished the season with a .775 OPS in the second half compared to a .909 OPS in the first half.
Batting order splits are McCutchen’s are tricky. Did he really pull out of his early-season slump because he moved down in the order? It is certainly possible he would have turned things around anyway, given his track record. The move down the order followed two games in which he was benched, so perhaps the benching was more important than the spot in the order. We may never know, but you can rest assured McCutchen’s spot in the batting order will be a topic of concern in 2018 now that McCutchen has been traded to the Giants.
McCutchen asked to move back to the No. 3 spot in the order back on June 26, so he either didn’t think he was better off lower in the order, or he thought it was better for his team that he bat higher, regardless of how it affected his performance. That, in a nutshell, highlights the difficulties of drawing conclusions from batting order splits.
It is certainly possible some players are more comfortable in certain spots in the order, and their performance may suffer if they are asked to move, but how do we identify those players? McCutchen didn’t chose to move to sixth, but his numbers were better there.
The only thing we know for sure regarding batting order is if you bat higher in the order, you can expect more plate appearances. You might get more RBI opportunities higher in the order, but that isn’t always the case. In 2016, Addison Russell had 194 plate appearances with runners in scoring position, good for ninth in MLB. He batted seventh or lower nearly half the time. That being said, it might not be a bad thing if he moves back to ninth, where he has a .318 wOBA.
Splits Against the Shift
As defensive shifts have gained popularity throughout the decade, several pull-heavy hitters have seen their batting averages plummet. This is especially true for left-handed hitters, since it is easier to shift against them. Beating a shift for one season is probably just a matter of luck, but if you beat it often enough, teams will stop shifting against you. More importantly, if you were terrible against the shift, you can expect to see more of it until you prove you can beat it. Guys who struggled against the shift in 2017 like Joc Pederson, Brian McCann, Albert Pujols, Rougned Odor and George Springer could struggle in 2018 if they don’t adjust
Splits with men on base or men in scoring position
There is obviously a difference between having one runner on first or having the lead runner on second or third, but for our purposes, they are essentially the same. For the most part, these splits, perhaps more so than any others, simply come down to luck, especially in an individual season. Baseball commentators like to treat hitting with runners on as a skill, but there is little evidence to support that position. If a player or team excels or struggles in those situations, it is largely due to luck.
With that in mind, we can look at a players’ splits with men on base or in scoring position as a place for potential regression to the mean. Marcell Ozuna batted out of his mind with runners on base last season, which you pretty much have to do to record 124 RBI in a season. Even if we expected him to repeat his overall slash numbers, we probably don’t expect him to bat 64 points higher with runners in scoring position than with the bases empty, as he did in 2017. We should adjust our expectations accordingly.
Batted Ball Splits
Much like situational splits, batted ball splits tend to normalize over time. If we can identify outliers within the batted ball stats in one season, we may be able to predict a regression to the mean the following season.
For Instance, Javier Baez batted .319 with 23 home runs on 116 fly balls in 2017. In 2016, he batted .185 with 12 home runs on 108 fly balls. When we project Baez for 2018, we should probably expect his batting average and home runs on fly balls to decrease, and thus his overall numbers as well.
What in the Wide World of Sports happened to Masahiro Tanaka last season? He went 3-6 with a 6.99 ERA in 10 day starts and he went 10-6 with a 3.93 ERA in 20 starts at night. You may think it is part of a larger trend, but he had a lower ERA in day games in two of his previous three seasons. A disparity that great bears watching for 2018, but it could also offer hope for a bounce back. Maybe Tanaka actually was worse in days games in 2017, but perhaps that is an area he can improve going forward if he changes his routine leading up to day games.
Tanaka’s case is an extreme example from one season, but day/night splits can be impactful. They are especially useful to look at if a player goes to or leaves a team that plays an unusual number of day games, like the Cubs.
Finally, we have reached the portion of our show where we discuss stat splits to ignore. The most prominent of those are league-specific splits and splits against an individual team. We will deal with AL and NL splits first.
Because the American League uses the Worst Rule in the History of Sports, otherwise known as the designated hitter, we would expect pitchers to do better against National League lineups than American League lineups. If that is not the case, it is probably noise, but that doesn’t stop some people from citing stats for one league or another.
Intellectually, it is pretty obvious that just because a pitcher has had more success against the American League in the past, that is in no way predictive for the future. While the League may be the same, everything else is different. Having success against the White Sox three years ago is no indication of your ability to get out Angels in 2018. If you see anyone cite statistics against the American or National Leagues as anything other than mildly interesting, you should raise all of your eyebrows.
If you are a fan of a specific baseball team for long enough, you can probably identify a handful of players who seem to always do well against your team. As a Cubs fan, the number one “Cubs killer” might be Yadier Molina. It seems like every time the Cubs lose to the Cardinals, Molina is right there in the middle of it. The numbers seem to back that up too, as Molina has a .796 OPS in his career against the Cubs. He has a higher OPS against just three teams. They are all in the American League, and he has fewer than 65 at-bats against each of those teams. That being said, over the last three seasons, Molina’s OPS against the Cubs is just .738. It makes sense, considering the Cubs have been much better over the last three seasons, but it proves my point, which is:
Unlike ballpark splits, team-specific splits are just about laundry. They don’t really tell us anything. Yadier Molina is entering his 15th season. There is no reason to believe how he did against the Cubs in 2006 tells us anything meaningful about how he will do against them in 2018. Maybe he is more confident against the Cubs, and maybe that results in better numbers against them the next time around, but how much stock are you really willing to put into that?
Perhaps the most important takeaway from all of this talk about splits is to view any split with a healthy amount of skepticism. Many splits (but not all) can be useful given the correct context, but we have to understand the context before we use them to draw any conclusions or move players around your fantasy draft board.