Compatiblity matching

The baseball statistics available freely online have advanced greatly over the past few years. As fantasy players, it is important to keep up with the research going on in the “real” aspect of baseball and continually try to apply it to fantasy baseball.

One of the leaders of the movement, FanGraphs, now has Pitch Type Linear Weights that show the pitches individual pitchers are better at throwing, and which pitches individual batters are better at hitting. (Click here for the accompanying explanation article by Dave Allen.) Taking at quick glance at Mark Teixeira’s FanGraphs player page shows the following information:

+-------------------------------------------------------------------------------------+ | Mark Teixeira Pitch Type Values | +--------+----------------+------+------+------+-----++-------+-------+-------+-------+ | Season | Team | wFB | wSL | wCB | wCH || wFB/C | wSL/C | wCB/C | wCH/C | +--------+----------------+------+------+------+-----++-------+-------+-------+-------+ | 2006 | Rangers | 20.8 | -0.7 | 0.8 | 5.6 || 1.24 | -0.36 | 0.32 | 1.05 | | 2007 | Rangers/Braves | 31 | 4.5 | 0.9 | 0.4 || 2.43 | 2.36 | 0.48 | 0.1 | | 2008 | Braves/Angels | 39 | 1.3 | 3.1 | 5.3 || 2.46 | 0.43 | 1.54 | 1.49 | | 2009 | Yankees | 24.3 | -1.5 | -1.1 | 0 || 2.5 | -1.02 | -0.63 | -0.01 | +--------+----------------+------+------+------+-----++-------+-------+-------+-------+

The numbers on the left of the divide show in total how many runs Teixeira has earned hitting fastballs, which for 2009 has been 24.3 runs. On the right side, where the column headers have a “/C” after them, the numbers show how many runs a player earns on a certain pitch per 100 pitches. For Mark Teixeira in 2009, that number is 2.5 runs per 100 fastballs.

image
Casey Blake connecting for a home run on what was most likely either a fastball or change-up. (Icon/SMI)

For these stats I believe 0 is average, positive means a player is good at hitting that pitch, and negative means the batter struggles with the pitch. Messing around with the leaderboards will help give you a relative context of how good a 2.5 wFB/C is.

Although extremely interesting, keep in mind the numbers are not perfect. The pitch classifications FanGraphs uses are not manually adjusted and no tests have been done to my knowledge on the stability of the pitch value numbers. Also the methodology behind the values is still somewhat of a work in progress, but nevertheless they can still used for fantasy purposes.

Most simply, when deciding which batter to start of two, besides looking just at the skill and handedness of the opposing pitchers, you can also check out the pitches they are better at throwing and the pitches your batters hit better.

For example, let’s say you own Cody Ross who has pretty consistently hit change-ups well throughout his career. Some nights he starts for your team and other nights he sits one out. Let’s say tonight the Marlins are playing the Astros and Wandy Rodriguez is pitching. Taking a look at his Pitch Value numbers, he historically has a below-average change-up and still throws it somewhat often at 10 percent of the time.

Tonight, then, would be a time to make sure Ross is in your starting lineup because of the increased possibility of him pounding one of Wandy’s change-ups out of the yard, or at least what FanGraphs—as provided by BIS—is classifying as a change-up.

In terms of importance, I would rank this below matching up handedness and skill of the opposing pitcher, simply because I do not really know how effective mixing and matching batters to pitchers by individual pitch is. Unfortunately, I did not invest the time yet to find out, so for today that question will be left unanswered.

Instead I’ll leave you with what is possibly new idea and if it’s not new, feel free to tell me in the comments how you have been using it.


Print This Post
Sort by:   newest | oldest | most voted
Paul Singman
Guest
Paul Singman

Yea I tried to stress in the article not to weigh the matchups too heavily. This article just scratched the surface, and I would like to look into exactly how much better you can expect a fastball hitter to hit off a pitcher with a worse fastball compare to one with a better one.

Troy Patterson
Guest
Troy Patterson

I would worry about using this to much, but it is a good idea.  How many teams might use data like this and as part of the scouting report tell their pitchers not to use certain pitches as much versus certain hitters.

Jeremy Boyd
Guest
Jeremy Boyd
Its possible to do this with pitchers too, by looking at the pitch type linear weights on FanGraphs for the opposing team overall (batters), and comparing them to the pitch type linear weights for your pitcher. What I’m trying to figure out is how to weight (a) pitch type linear weights, relative to (b) park factors, (c) pitcher skill, (d) opposing team skill. A good case study is last night.  My league has two SP spots and I had to choose two of J. Vazquez, Y. Gallardo, M. Garza, and J. Santana.  Vazquez was a no-brainer pitching against San Diego… Read more »
Troy Patterson
Guest
Troy Patterson
Jeremy, I think you are looking at much to small of a sample size.  One month of pitch values is not enough to judge a team.  Park factors and better pitching should always come first.  As long as Santana has a K/BB over 3 and is pitching in the pitchers park of Citi Field he is the best choice by far. You can say I was wrong since he gave up 5 runs and 2 homers, but that’s the danger of trying to project single game results.  At this stage there is no way you can’t pick Vazquez and Santana… Read more »
Jeremy Boyd
Guest
Jeremy Boyd

Troy,

Good point about K/BB, but the evidence shows Citi is a neutral park, not a pitcher’s park. 

K/BB selects Santana over Garza or Gallardo, but Park Factors selects Dodger Stadium, a pitcher’s park, over Tropicanna Field, and Citi Field, which are both neutral parks.

Which brings me back to my original question about which factors to weight highest.  Using K/BB and Park Factors, I eliminate Garza, but I still can’t decide between Santana and Gallardo based on K/BB and Park Factors alone, unless I weight one higher than the other.

digglahhh
Guest
digglahhh
Perhaps this is an example of paralysis by analysis, or of people overthinking things. These are very real dangers of the virtually infinite tool kit we have at our fingertips nowadays. We must always remember that we have to use these tools for our benefit and not to simply create new, more esoteric, dilemas. (Don’t mean to sound like Joe Morgan here, but sometimes I draw the line.) Quite simply, when you draft or buy Johan Santana you are paying a premium. Part of the reason for that cost is that you are going to start him every start. You’re… Read more »
digglahhh
Guest
digglahhh

Sorry for the horrific grammar/typos in the post above – sometimes a casualty of posting from work…

Jeremy Boyd
Guest
Jeremy Boyd
Thanks digglahhh, You and Troy both made good points, IMO.  I’m a statistical programmer by trade, so my interest in this is partially motiviated by trying to develop a statistical model to make these decisions for me.  Because you are right, I did hit the point of analysis paralysis regarding which of the 4 SP to start yesterday.  Statistical models are great panaceas for analysis paralysis, because of their ability to objectively sift through a lot of information.  And I oppose SP RP distinctions too. On a different note, I don’t want to focus on price, investment, premiums, etc.., exclusively,… Read more »
Paul Singman
Guest
Paul Singman
I’ll jump in here seeing how it is my article. Jeremy, picking between players of that caliber should not happen often, I envisioned it being used more for fring-ish players like Cody Ross (who is getting hot right now, btw). As I say in the article, I would weigh these pitch type values the least because no one has really tested how much better you can expect a slider pitcher to do against a team that ranks one of the worst at hitting sliders. That’s not even mentioning the pitch classification issue with this data. In the future, though, this… Read more »
digglahhh
Guest
digglahhh
Jeremy, Yes, but once the investment is made (the roster is filled) ROI becomes meaningless. You are accruing actual production, not production per dollar, or per ADP. So, you want to put in the best guys you have. The more mid-level assets you have, the more difficult decisions you will have to make, and the greater proportion of your initial investment you likely leave on the table over the course of season (regardless of whether than investment was above or below fair market value). In other words, this didn’t become a tough decision because you have Johan Santana, a tier… Read more »
Jeremy Boyd
Guest
Jeremy Boyd
digglahhh, Thanks for the great advice.  In retrospect, Gallardo vs. the best team in baseball was a bad bet.  The pitch type linear weights appear to have corroborated this. Garza vs. a team he’s pitched well against all year was a probably a good bet.  It looks like pitch type linear weight matching gave a similar answer.  I agree that Johan was the best pitcher in baseball, for the first two months of the season, but I exhausted my failure comfort with Santana back in June.  His earned runs per game are highly variable since June, pitching shutouts or blowouts. … Read more »
wpDiscuz