MLB FORUM

Comments

  • mwek

    As I learn MLB DFS I’m getting more and more in to batted ball profiles and how to leverage them. I have a couple question for anyone who’s kind enough to answer:

    Where is a good place to find data on what type of pitcher (ground ball/fly ball) a certain batter succeeds more often against?

    Finally, are there any ways you translate this into a winning lineup beyond the obvious (i.e.: using a fly ball hitter in a hitter-friendly park)?

  • BennyRamirez

    Like a lot of things, it depends. Batted ball data doesn’t really translate well to a model for a quick wOBA modifier, if that’s what you’re asking. It’s more of a starting place to simulate games. There are some general rules that good fly ball pitchers who don’t give up baserunners are good spots for contrarian 2-run and solo HRs from power hitters, but a bad fly ball pitcher is going to give up more line drives to ground ball hitters and that good pitcher can also get those power hitters to pop-up in the infield more often.

    There is an unspoken argument in DFS on the levels of contact allowed (soft, medium, hard). Pitchers’ contact degrees are often cited, despite some research in the SABR community on the correlation that and the pitchers’ ability being lower than the citations may imply. I don’t see this argument discussed (I think Derek Carty did this study at Baseball Prospectus years ago, but I may be wrong.) I don’t know where I stand on this, other than to oversimplify and say that bad pitchers give up more hard contact because the hitters’ are further from the bottom of their ranges due to the pitcher being bad, meaning that modifying a hitter’s expectation using hard contact data may end up being redundant, as it is already being calculated in wOBA and/or whatever else is being calculated.

    To better answer your question, batted ball data is more of a supplemental tool to look at matchup-by-matchup and can hurt you if you’re looking to modify the “what has happened” data to create “expected” data in a “sort by x” manner. You shouldn’t really be reducing analysis to a single number anyway, because baseball is more complex and has too many variables from slate to slate and game to game. Batted ball data will show you this, as it emphasizes that there are varying types of good and bad hitters and pitchers, and will give you better explanations of this.

  • FrancisBooth

    I don’t know which episode it was, but I highly recommend finding the thread posted about rotonomics. The stuff cheese and jm discuss on there is in line with what you are asking I am pretty sure. Fangraphs is a great resource but it can be a bit overwhelming if you are just learning all of the advanced statistics for MLB.

    Here it is….
    https://rotogrinders.com/threads/rotonomics-advanced-stats-disussion-1319837?page=1

    You will have to look through grinders live to find the rest. They are on almost every Thursday and Cheeseisgood and Jmtowin are the hosts

  • mwek

    Thanks, guys. The Rotonomics episode you linked is what really peaked my interest here. I found myself looking for more data but found it hard to come by. JM quotes data like “so and so hits better against ground ball pitchers” and I am not sure where to find that sort of data.

    Benny, thank you for replying in such detail. I love your point about not using batted ball data solely as predictive. Your first paragraph was also helpful.

  • mwek

    Also, I check fangraphs often and don’t find it too difficult to wrap my head around the statistics listed there, but I do find navigating the site to get it to tell me what I need to know a little more difficult.

  • BennyRamirez

    Once you get the hang of using the dropdowns for sorting and how stats are tabbed, it gets easier. Start making custom reports and you’ll get used to it, day-by-day.

  • BennyRamirez

    Once you get the hang of using the dropdowns for sorting and how stats are tabbed, it gets easier. Start making custom reports and you’ll get used to it, day-by-day.

  • KindGuy

    I was actually reading an article recently about this. I’ll link it if i can find it.

    Bascially:

    Ground Ball Pitcher vs Ground Ball Hitter: Good for pitcher

    Ground Ball Pitcher vs Fly Ball Hitter: Good for hitter

    Fly Ball Pitcher vs Ground Ball Hitter: Good for Hitter

    Fly Ball Pitcher vs Fly Ball Hitter: Good for Pitcher

  • KindGuy

    Here you go: http://www.fantasylabs.com/articles/predict-batted-ball-profiles/

  • thepacificocean

    The stats that they are referencing are from Baseball-Reference.com and they have to be searched individually. Go to a player page on B-R, and find the “splits” tab for a player and click that, you can sort by career or individual year, once inside the splits page, you can scroll down the sheet quite a ways, until you see the “versus flyball pitcher/versus ground ball pitcher” split, I think it’s right below the vs power/finesse pitcher splits.

    That said, as a general rule of thumb, most groundball hitters, i.e. hitters with GB% that is noticeably higher than their FB%, will perform better against FB pitchers and vice versa.

  • mwek

    Thank you you guys.

  • divusjulius

    • Blogger of the Month

    @mwek

    hey. very smart of you to look into this; it really helps make your probability models much more accurate; if one knows how to use BB profile and pitch f/x data, and mix it in with the new statcast stats and data, finding players in a spot for massive upside becomes much much easier. get it down and you’ll be amazed at how often you identify the players who hit hr’s that night.

    i wrote a series of blog posts here at RG on it, and have more in depth at my site, and there are some excellent articles at fangraphs and the hardball times. But honestly, i don’t think anyone has theorized this type of material as much or as thoroughly as Bales (in his books and course) and to be honest, the batted ball and statxcast data behind his paywall his worth the price of admission.

  • CheeseIsGood

    MLB Premium Director

    • 23

      RG Overall Ranking

    • Ranked #22

      RG Tiered Ranking

    • x2

      $1M Prize Winner

    • x2

      2015 DraftKings FFWC Finalist

    I wanted to jump in here with a couple thoughts. Batted ball data is incredibly complex, but absolutely fascinating once you get into it. The most basic batted ball data (ground balls or fly balls) and soft hit%/hard hit% that can all be found on fangraphs is the place to start. But, once you discover that it sends you down a rabbit hole of this type of pitcher vs this type of batter, etc.

    A word of caution on that type of data as it relates to sample size. We always talk about how we need large sample sizes for data to be relevant. Unfortunately, when it comes to batted balls in DFS, all we have are large sample sizes that are actually too large. All of this GB vs FB data that was first compiled in ‘The Book’ is tens of thousands of at bats from all different types of players. So when we say, Ground Ball Hitters do better against Fly Ball Pitchers and vice versa, that is only true in a big picture sense and not necessarily in any one matchup. A left-handed sinkerball pitcher with a 50% GB% cannot be expected to have the same results against Batter A as a right-handed pitcher with a 50% GB who throws a 2-seam fastball. But, all we have is all that data smashed together.

    BaseballReference.com splits page where they show a batters overall numbers vs GB/AVG/FB pitchers is a better next step than nothing when trying to determine how a matchup will go, but unless the numbers are fairly extreme, we still don’t know what type of ground ball pitcher that data came against.

    The large sample size data will have Noah Syndergaard and Mike Pelfrey as the same Ground Ball pitcher. If you really wanted to the most accurate info, you would break down every at bat and see who and when they were against, but of course this is impossible, at least it is for me and my notebook and calculator.

    None of this is to say the in depth batted ball data is not useful, but rather that you should be careful not overweighting things that we can’t fully understand. Definitely worth spending time looking at the GB vs FB data on baseball reference along with the hard hit data on fangraphs, but just beware of its shortcomings in any one day matchup.

  • thehazyone

    RG Contributor

    • Blogger of the Month

    Awesome response CheeseIsGood. You are indeed the GOAT of MLB.

  • BennyRamirez

    Cheese, how useful is the batted ball data in recent data? I will often hear that a guy is really squaring up the ball or not seeing the ball well lately. Giancarlo Stanton was a good recent example where his ABs were almost binary K-Hard results, despite everything just coming up bad in the standard counting numbers. Case-by-case, I form biases when I look at these, but bring myself to going with a grain of salt because I’m unsure of how far to take it.

    Sorry if the question is unclear, but I have wondered if there is some short-term utility in finding regression candidates when I think people will be afraid of bad recent results.

    Also, what are your thoughts on the argument I mentioned in my first reply ITT on the relevance of hard contact allowed?

    Thanks. Always appreciate your great work.

  • mwek

    Wow. Wasn’t expecting such a great response from one of my favorite analysts. Can’t thank you enough for weighing in. Your words of caution (and encouragement) are helpful.

    One of the reasons I got really interested in this after listening to your Rotonomics pod was because I’ve had some success targeting Chris Davis against GB pitchers. I’m not good enough at this to know if that matchup and batted ball profiles are the reasons I got a homer out of Davis those nights, but it sure was fun to be right. Again, just got me thinking and I appreciate everyone’s guidance on how to dig in further.

  • divusjulius

    • Blogger of the Month

    thanks for the input cheese….thanks for the brooksbaseball mention too, that’s been real fun to dive into. but you are right, the various discussions of BB data are like going down the rabbit hole, and at some point one has to stop and realize that what’s going on is an academic discussion among statisticains and baseball people about how to make sense of all this data for their game (player projection for scouting, front offices, competing analytic services etc) and not for our game of DFS.

  • CheeseIsGood

    MLB Premium Director

    • 23

      RG Overall Ranking

    • Ranked #22

      RG Tiered Ranking

    • x2

      $1M Prize Winner

    • x2

      2015 DraftKings FFWC Finalist

    @BennyRamirez said...

    Cheese, how useful is the batted ball data in recent data? I will often hear that a guy is really squaring up the ball or not seeing the ball well lately. Giancarlo Stanton was a good recent example where his ABs were almost binary K-Hard results, despite everything just coming up bad in the standard counting numbers. Case-by-case, I form biases when I look at these, but bring myself to going with a grain of salt because I’m unsure of how far to take it.

    Sorry if the question is unclear, but I have wondered if there is some short-term utility in finding regression candidates when I think people will be afraid of bad recent results.

    Also, what are your thoughts on the argument I mentioned in my first reply ITT on the relevance of hard contact allowed?

    Thanks. Always appreciate your great work.

    I think that Stanton case was very interesting. If you recall in the midst of Stanton’s cold streak, I was saying my opinion was that his problems were mental and not physical, so I assumed he’d be fine. The reasoning behind that was mostly due to the hard hit rate. He was striking out a lot, which would be the case if he was in his own head swinging at bad pitches. But, when he connected, he was still hitting the ball hard, they were just outs a lot of the time. If he was injured, I would’ve expected more soft contact and infield fly balls, etc.

  • CheeseIsGood

    MLB Premium Director

    • 23

      RG Overall Ranking

    • Ranked #22

      RG Tiered Ranking

    • x2

      $1M Prize Winner

    • x2

      2015 DraftKings FFWC Finalist

    @mwek said...

    Wow. Wasn’t expecting such a great response from one of my favorite analysts. Can’t thank you enough for weighing in. Your words of caution (and encouragement) are helpful.

    One of the reasons I got really interested in this after listening to your Rotonomics pod was because I’ve had some success targeting Chris Davis against GB pitchers. I’m not good enough at this to know if that matchup and batted ball profiles are the reasons I got a homer out of Davis those nights, but it sure was fun to be right. Again, just got me thinking and I appreciate everyone’s guidance on how to dig in further.

    The more extreme the data, the more you can rely on it. In Chris Davis case, his numbers vs GB or FB (on baseballreference) are so incredibly skewed in favor of him facing GB pitchers that we can take it to be real. Most guys fall somewhere in the middle, and their splits one way or the other aren’t extreme enough to assume they’ll hold up.
    So, yes, with someone like Davis, if you look at the data and see the clear numbers, you should use that in your play, but always check the data on anyone before you use him assuming that he’ll have a GB/FB advantage.

  • CheeseIsGood

    MLB Premium Director

    • 23

      RG Overall Ranking

    • Ranked #22

      RG Tiered Ranking

    • x2

      $1M Prize Winner

    • x2

      2015 DraftKings FFWC Finalist

    @BennyRamirez said...

    Also, what are your thoughts on the argument I mentioned in my first reply ITT on the relevance of hard contact allowed?

    If I understand your question based on your first response, what I think you’re talking about is the potential to double weight hard contact from pitchers. If you look at a pitcher and see he has a high ISO/wOBA/OPS against, and say he’s bad, and then you look and see the hard hits and say “he’s even worse”, then you’re doubled weighted the hard hits, because the reason his ISO/wOBA/OPS is high is simply because he does give up hard hits.

  • lonestar2016

    Cheese, I’ve been using a lot of data related to pitch type and how the hitters hit those pitches. For example, Tyler Duffey throws the Knuckle-Curve 38.5% of time, and Ian Desmond hits that pitch well. Is this type of data worth targeting, how can I improve on targeting this data, and is this data more useful for off-speed pitches than a fastball?

  • CheeseIsGood

    MLB Premium Director

    • 23

      RG Overall Ranking

    • Ranked #22

      RG Tiered Ranking

    • x2

      $1M Prize Winner

    • x2

      2015 DraftKings FFWC Finalist

    @lonestar2016 said...

    Cheese, I’ve been using a lot of data related to pitch type and how the hitters hit those pitches. For example, Tyler Duffey throws the Knuckle-Curve 38.5% of time, and Ian Desmond hits that pitch well. Is this type of data worth targeting, how can I improve on targeting this data, and is this data more useful for off-speed pitches than a fastball?

    If you have the time to look at specific pitch type, yes it is useful. You just need to make sure you’re checking the handedness, as some pitchers throw different arsenals vs R and L, and unless it’s an extreme percentage, you can’t overweight it, because in your example, if Duffey throws 38.5% knuckle cuves, and Desmond has a 7-pitch at bat, that means maybe he sees 3 knuckle curves. We’re really tempting sample size at that point. However, of course, even for 3 pitches, you would prefer a guy to see a pitch type that he hits well.

    It is generally accepted that type of data is more useful for knuckleballs, but I don’t have a definite answer for that.

  • BennyRamirez

    @CheeseIsGood said...

    If you have the time to look at specific pitch type, yes it is useful. You just need to make sure you’re checking the handedness, as some pitchers throw different arsenals vs R and L, and unless it’s an extreme percentage, you can’t overweight it, because in your example, if Duffey throws 38.5% knuckle cuves, and Desmond has a 7-pitch at bat, that means maybe he sees 3 knuckle curves. We’re really tempting sample size at that point. However, of course, even for 3 pitches, you would prefer a guy to see a pitch type that he hits well.

    It is generally accepted that type of data is more useful for knuckleballs, but I don’t have a definite answer for that.

    I think the fact the basic versus-handedness data is less reliable when looking at knuckleballers, the value of data specifically related to the knuckleball raises in relativity. Whether or not hitters can have enough of a sample size against knuckleballers to make data reliable is where the jury is out. Knuckleballer BvP Truthers prematurely conflate the lessened value of versus-handedness with BvP utility in these cases. They may be correct, but it is by accident, as the evidence is difficult to compile with any reasonably acceptable margin of error.

  • BennyRamirez

    @CheeseIsGood said...

    If I understand your question based on your first response, what I think you’re talking about is the potential to double weight hard contact from pitchers. If you look at a pitcher and see he has a high ISO/wOBA/OPS against, and say he’s bad, and then you look and see the hard hits and say “he’s even worse”, then you’re doubled weighted the hard hits, because the reason his ISO/wOBA/OPS is high is simply because he does give up hard hits.

    Yes, and (as opposed to “yes, but”) there is this sentence about which I frequently think:

    “A major-league pitcher does not only control whether he gets ground balls or fly balls; he also has a significant degree of control over how hard the ball is hit, though the batter has somewhat more control over the quality of contact than the pitcher.”

    (Source: http://www.baseballprospectus.com/article.php?articleid=15532)

    This leads me to wonder if hitters with below average hard hit rates are more prone to harder contact against pitchers with higher hard hit rates allowed or are the hitters who already have the higher hard hit rates just the one with the escalated expectation of harder contact? I would assume the correlation would have whatever the opposite of the long tail?

    Or is it more of a bell curve where those around the average are the only ones prone to escalations or reductions in hard hit rates, in accordance with the pitcher?

    Jarrod Salatlamacchia and Brian Dozier are two guys playing today with ISO rates of .243 and .246, respectively, against lefties since the start of 2015, but hard hit rates under 30% against lefties? Manny Machado and Corey Seager have equal ISOs of .244 against righties over that same span, but hard hit rates of 35.3 and 40.7%. Does the former group’s lower hard hit rates make them more matchup-dependent than the latter? Being a lesser quality of hitter, the answer is pretty clearly yes, that we ought to trust Machado or Seager against a good righty over Dozier or Salty against a good lefty, and the overall quality is emphasized in the hard hit rates. But how much ought the pitcher’s hard hit rate be factored into how likely we are to get the top of Dozier/Salty’s range versus other data?

  • MTro86

    RG Writer

    This is a fascinating discussion. My own thoughts on some of the pitch type and batted ball stuff is that if it’s something you have to search individually per batter for instead of being able to pull up league leaders (ie good hitters against ground ball or fly ball pitchers) there’s a great edge there if you are able to stop time for an hour or two and go through individually, otherwise it is even something players have enough time for or are they losing a bigger edge somewhere else by giving up the amount of time it costs to search individual hitters?

    As far as batter vs pitch type, I used to like to look at it a lot more, but I’m now assuming that if we know a guy murders sliders, doesn’t the other team know this as well? Unless it’s something there the guy only has two good pitches and he has to throw it, aren’t they going to avoid throwing pitches a crushes otherwise?

    For me a lot of this stuff comes down to time management. Even in writing articles, I wish I had more time to explain things better or go into more depth, but I’m already putting in a full day and always looking to be more efficient. Stuff like LD% and contact rates (Hard, Soft) aren’t as predictive as descriptive over short sample sizes (and contact rates seem to have very little correlation with BABIP, but more with power), but they’re quick ways to see and explain how someone has gotten to a certain point so far.

  • PGASplits101

    So how useful is the Power/Finesse split on Baseball Reference? They include BB as part of the Power equation which makes it misleading to me. There are plenty of soft tossing low strike out pitchers who have control issues. My opinion is that if you can’t break it down into Left/Right splits then it is not as accurate as it could be. For example, with the Fly Ball/Ground Ball split on Baseball Ref, you don’t know if the higher OPS against a FB pitcher was mostly due to success against RHP or LHP. Perhaps a batter has tons of success hitting for a high OPS against RH Fly Ball pitchers, but does poorly against LH Fly Ball Pitchers. I also don’t understand why you can’t search for multiple splits at once on there…and especially on Fangraphs. You still cannot search for multiple splits at the same time on their leaderboard pages. For example, it would be awesome to see a leaderboard of Hard% vs RHP over the last 14 days.

  • X Unread Thread
  • X Thread with New Replies*
  • *Jumps to your first unread reply

Subforum Index

RotoGrinders.com is the home of the daily fantasy sports community. Our content, rankings, member blogs, promotions and forum discussion all cater to the players that like to create a new fantasy team every day of the week.

If you or someone you know has a gambling problem, crisis counseling and referral services can be accessed by calling 1-800-GAMBLER (1-800-426-2537) (IL). Gambling problem? Call 1-800-GAMBLER (NJ/WV/PA/MI), 1-800-9-WITH-IT (IN), 1-800-522-4700 (CO), 1-800-BETS OFF (IA), 1-888-532-3500 (VA) or call/text TN REDLINE 1-800-889-9789 (TN).