The Z Files: Revamping the Weekly Hitting Rankings

Recently, there's been some constructive criticism in the comments. As a result, I'm in process of adding some features which I hope to unveil this weekend. The following will review the current and impending factors included each week. If you're not a reader of the piece, look at this as my process to decide between fringe hitters in my weekly lineups or as a means to evaluate potential free agent acquisitions in leagues requiring all pick-ups to be active that week, like LABR and Tout Wars.

CURRENT FACTORS

Home/Away

Home field advantage is real. Globally, batters exhibit better skills in their regular venue as compared to on the road. Here's a look at some of the numbers:

Across the board, playing at home helps. Note even BABIP (batting average on balls in play) is consistently better at home. This may be counter-intuitive to those still insisting the outcome of a batted ball is left entirely to fate. Yes, there's more variance than with other stats, but over a large enough sample, it can be shown BABIP is more than just happenstance.

Platoon Matchups

There's a reason why managers set lineups based on the opposing pitchers and bring in relievers of the same handedness as the batter at the plate in later innings. To wit:

The RHB vs. LHP and LHB vs. RHB may flip-flop from year to year when it comes to which set is better, but overall, the pattern is clear: Hitters with the platoon edge are favored. Right-on-right is next with left-on-left consistently yielding the poorest results. Hopefully, this isn't something you didn't already know. However, it's always nice to see how much the numbers support intuition, as well as reminding us just how much of an advantage a southpaw hurler has over a lefty swinger.

Obviously, not every hitter falls perfectly in line with the league-wide numbers. However, they're usually close.

Chances are, you're wondering about hitters exhibiting reverse splits, opposite of the league norm. This is important, especially in DFS since many will cite a reverse split as actionable when it's not. According to the research published in The Book: Playing the Percentages in Baseball, by Tom M Tango, Mitchel G. Lichtman and Andrew E. Dolphin, a RHB needs 2000 plate appearances against a LHP before his splits are reliable, while a LHB requires 1000.

Putting that in perspective, between 26 and 27 percent of trips to the dish are with a southpaw on the hill. Using 650 plate appearances per season, a LHB needs 6.5 seasons before he owns his splits versus LHP, while a RHB needs 13 years.

We're getting beyond the scope of this discussion, but the closer a hitter is to the threshold, the more reliable the splits are. For those mathematically inclined, you can think of their splits as the weighted average of their career mark and the league average for the plate appearances necessary to reach the threshold.

The important point is ignoring small-sample platoon splits, especially those anti to the norm and/or the player's baseline. This isn't to say you can't find instances to leverage abnormal splits, just that usually they're more fluke than fact.

Home Run Park Indices

This one is obvious; no fancy charts are necessary to explain why some parks favor power hitters. However, it is worth reiterating a venue's home run factor and run factor aren't always in sync. Some parks embellish power but suppress runs, and vice versa. Back in April, I wrote this piece detailing how all 30 parks play in terms of homers and runs.

FACTORS TO BE ADDED

Speed Index

This is especially apropos after the firestorm that just occurred with the Chicago Cubs, resulting in the release of Miguel Montero after he threw Jake Arrieta and the rest of the rotation under the team bus when asked about the club's inability to control the running game. Some batteries are just easier to steal against than others. If you're chasing steals and deciding between a speedster and a slugger, knowing how many steals the scheduled starters give up can be integral information.

Currently, this data is not provided. The plan is to add it in index form, where 100 represent league average ability to throw out would-be base-stealers. The calculation will consider the starting pitchers scheduled for the week, how long they're projected to pitch and an overall bullpen factor. While it's true the better starters are likely to be followed by better relievers, programming on that granular level is too cumbersome.

Presently, catchers are throwing out runners at a 28 percent clip. Teams matching that will get an index of 100. Say based on the scheduled starters for the week, a team projects to throw out 31 percent. Their factor would be 100 x 28/31 or 90. To keep things consistent with park factors etc., below 100 will be detrimental for the hitter, over 100 will be advantageous.

In situations like this, it helps to have some framework. To this point of the season, the easiest team to run on is in Chicago except it's the White Sox, successfully nabbing only 14 percent, equating to a 200 index. On the flip side, Indians receivers are thwarting 44 percent, a 64 index.

Some adjustments will be made for teams with new backstops, such as the Milwaukee Brewers recent acquisition of Stephen Vogt, replace the jettisoned Jett Bandy.

Quality of Opposing Pitchers

This could be the single most important factor. Unfortunately, it's also the most intricate to properly represent via index. As a surrogate, the site's probable pitching chart is transposed to show the starters each team is slated to face each week. As such, it's possible to eyeball who's on the docket to get a feel for how difficult the upcoming week may be. But, there's no objective determination.

The first dilemma is deciding what to use as the metric. Keeping in mind to make it as useful as possible, a separate index should apply to RHB and LHB. Further, whatever is chosen, it's going to be different home and away. Not only does a hitter display superior skills at home, so does a pitcher. Putting this together, we're looking at parsing the metric into four parts: home and away, with those broken into versus LHB and RHB. The data is available. The programming is more tedious than difficult. The issue is the more you slice the data, the more variance is introduced.

Weighted on base average (wOBA) is the clubhouse leader as the measuring stick. ERA is also in the mix. For those not familiar, wOBA is akin to standard OBP, with different coefficients for the individual components. These coefficients are empirically determined using the expected run matrix. The result is a number serving as an excellent proxy to estimate potential run production.

An adjustment will need to be made for venue, since wOBA isn't park adjusted. You may be familiar with wRC+, which is essentially park-adjusted wOBA, expressed as an index. Using it helps, but I still need to account for the venues.

The final consideration is while hitters can be in a hot or cold stretch, how they're hitting now isn't predictive of how they'll perform the next time they step in the box. However, a pitcher with a string of solid outings has a better than even chance of continuing to perform at an elevated level. In other words, the wOBA index must factor in recent performance as well as seasonal or career levels.

While I hope to have something ready for this weekend, it will be a first iteration, a work in progress. There's also a chance I don't add it for another week.

Team Context

Players on teams scoring more runs are obviously in a better spot for runs and RBI. As part of the Pitching Ranks piece, I calculate the score of each game, used to project win probability. This can double as a vehicle for a team scoring index.

At first blush, determining average runs per game then comparing that to the league average renders a nice index. The problem is teams play differing amounts of games each week. So, upon further review, the run-scoring index will use total runs projected for the week. What matters is how many runs and RBI a batter may accrue in the given week. A lower scoring team with seven games could score more runs than a higher scoring club with five or six games.

Facilitating this addition is all the pertinent factors are already baked in. The run projection already considers home/away, venue and bullpen.

That completes the data to be included in the table from here on out. For me, there's not a one-size-fits-all hierarchy of factors. It's contextual with each player. That said, if only allowed to use one, I'd want to know the quality of the opposing hurlers. Of course, that's the most arduous factor to quantify.

Even with this imminent improvement, there will still be a big point of contention. From the comments of previous installments, there's a strong desire for an overall team score, which can be used to rank the 30 squads.

Mathematically, this isn't a problem. My sticking point is more practical. Call me obstinate, but I fail to see the utility of a team rank. Player decisions must be contextual. An elevated ranking because the Angels have a four-game set with the White Sox on tap helps Cameron Maybin more than it does Kole Calhoun. Yes, for those short on time, using an aggregate index may be more efficient, but it's not more effective. Still, I learned a long time ago

1. The customer is always right
2. When the customer is wrong, see #1

So, I'll include an overall team score. But, teams will still be listed alphabetically. Based on feedback, public and private, most users prefer an alphabetical listing so they know exactly where to find each team quickly.

Before rushing to post your dissatisfaction in the comments, the final planned upgrade is sortable columns. If you want the teams ranked by aggregate index, sort that way. The catch is we may not be able to add the necessary html this week. I assure you we're working on it. Something else I've learned (but often have trouble following) is under-promise and over-deliver.

I'll now open the floor to comments. Fire away.