A Few Questions
Here at FanGraphs, we attempt to answer a lot of questions. How good is this guy compared to that guy? Who throws the hardest? Which pitches are effective against different hitters? What should I expect from that rookie?
For the next few days, however, I’m going to take a slightly different approach. As baseball analysis has begun to explode on the web, there is no shortage of answers to common questions. However, I think that we may have a shortage of questions. So, rather than attempting to figure out something that we want to know, I’d like to spend a few days trying to figure out what else we should want to know.
I’m not going to have any answers on this. My hope is that this would turn into a discussion where we can stimulate some ideas for what things may be worth answering in the future, even if we can’t answer them now. What don’t we know that may be important, and that we maybe haven’t even attempted to answer yet?
The main area in baseball that still strikes me as something of a mystery is pitching. We understand some things about the art of pitching. We know that, in general, more velocity is good, and it helps to be able to throw the ball in the strike zone with regularity. But do we have any idea why some guys have good command and other’s don’t?
Really, this seems like a fairly basic thing, but I don’t know if we have an answer. We could throw out a word like “mechanics,” but what does that tell us, really? For all the talk about good and bad mechanics, there doesn’t seem to be a “do-this-thing-and-you’ll-succeed” blueprint. Different stuff works for different guys. So what is it that drives a pitcher’s command? Arm angle? Muscle memory? Practice?
I don’t know. Maybe you do – if so, great, let’s hear it (with proof, please). But this seems like something we’d like to know, right? So, this is my question – what else should we want to know that we don’t? What parts of baseball have we just not given attention to?
3
Dave –
Would working on this also give some insight into any measurable effects from Catcher game calling? It seems like off all the position players, Catchers are the most difficult to measure defensively and much of that stems from the debate over their ability to affect the pitcher’s performance.
Things like WP & PB, CS vs SB, errors, and the like can all be measured and give some insight into a Catcher’s defensive performance. Without being able to measure things like framing, game calling, any effect of pitcher comfort, etc., measuring only what we currently can for catchers probably leaves out some undetermined amount of their value (or cost) behind the plate.
One of the great difficulties in measuring a catcher’s ability to call a game is the difference in starting pitchers.
Last year, the guys at DRaysBay tried to compare Dioner Navarro and Michel Hernandez, but the comparison lacked significant sample sizes for Michel, the back up. To avoid the sample size issue, we have to compare full-time catchers — but this returns to the original problem of pitching ability. A Braves catcher (best team FIP) is invariably going to appear better than any Orioles catcher (worst team FIP). How can we normalize this? Does it come down to pitch sequencing, like longghandi mentions below, or is there some other way to examine this?
Similar to measuring park factors, I wonder if there is enough data to use umpires as the ‘control.’ While they have individual zones, who gets more called strikes than average against ‘x’ ump?
For that matter, I’d like to know more about umpire influence. You hear “this ump has a small zone”–not that size matters–but how much does that influence a pitcher’s results? It happens every day where a guy gets squeezed and then gives up runs out of frustration, or just the knowledge that they have to get more of the plate. And hitters know that too. On the flip side, your Edwin Jacksons can surprise with a 10-whiff gem with a bit of help.
I don’t think it is a quality-of-game issue, since both teams have the same umpire, but start to start I bet it does influence pitchers. Get enough ‘small-zone’ vs ‘large-zone’ umpires throughout the season and I’d be willing to bet there’s correlation to ERA/FIP variation.
Similar to hittracker’s “just enough” info, it would be nice to know which pitchers were helped/hindered the most by umpire-effect.
Building on what Jimbo was saying, what about stats for home plate umpires? This could be relevant for betting and fantasy purposes if a guy with a notoriously large strike zone is gonig to be calling a game, you may want to take the under for example. Does anyone know if UMP FIP or anything like that being tracked anywhere? There may be some serious issues, but I don’t think sample size would be one of them.
Maybe this is something that’s already been dealt with in the sabermetric literature, but I’ve always wondered how much the quality of the opposition affects performance. It seems to me that most metrics give the pitcher the same evaluation if he gets the same result in the same park — a strikeout and two flyouts count the same regardless of whether the batters were Teixera, A-Rod, and Posada, or Neil Walker, Brian Bixler, and Luis Cruz. Yet the expected performance of an average pitcher against the two lineups is different, so shouldn’t our evaluation of the pitcher reflect that? Or does it all wash out in the sample size. (One reason to think it might not: Pirates pitchers never got to pitch to Walker, Bixler, and Cruz.)
Partly this is inspired by looking at Football Outsiders’ stats, where perhaps the biggest thing they do is adjusting for quality of opposition.
Great question, matt w. I often wonder the same thing, too. The way I usually phrase it in my mind is, “Do hitters accumulate their majority of their great numbers against bad pitching?” Perhaps more important from a predictive standpoint, “Do some guys struggle more than we would expect against elite hurlers, and compile most of their stats against bad pitchers?”
We could even break this down by individual hitters. We know A-Rod’s L/R splits. I could even check his performance against individual pitchers. But I don’t know how he does against the top 20% of major league pitchers, and I’d like to. (Maybe that kind of info is already available and calculated for me, and I just don’t know where).
I think that kind of info would be valuable.
Thanks, Jacob. I did a quick search (sabremetrics + “quality of opposition”) and found a post by Eric M. Van, linked in my name, which asked a lot of questions like yours about people’s splits. He has some very interesting speculations about LH/RH splits, for instance that the best LHBs’ value against LHP may be dampened, because managers will leave them in against the toughest LHP where they’d pinch-hit for a weaker LHB, so that better LHBs face better LHPs overall.
I was even thinking of a more basic question — are there any players who are better/worse than their stats in a given season because they faced worse/better opponents overall? I guess the overall consensus is that it washes out over a season, but I’m curious (and also wonder whether that’s really true of, for instance, relievers).
I’ve always heard its better to build around pitching in a pitchers park, and hitting in a hitters park, but I’ve never seen any studies done to prove it…
Bill James once wrote “I think that teams built around pitching generally don’t do well in post-season play, and never have.” But I’ve never seen a study to back this up.
I’ve heard a lot of people say the opposite, which makes more sense to me because you can go with a four-man rotation in the playoffs, giving those top starters more chances to impact the series.
I guess I’m looking more if there is a benefit to, for example, the Padres focusing more attention on improving their pitching since they are in a low run environment, or if the Rockies should do the opposite, etc.
To me it seems the park effect would affect both run scoring and prevention equally, and equally for both teams, but I’m not sure if its one of the many myths of baseball that have or need to be busted?
The rockies have focused in recent years on ground ball pitchers which is a result of their park helping flyballs more than groundballs. Accordingly, they have focused on infield defense more and are willing to play glove less big bats like Hawpe in the OF.
I understand there are certain ways to get a home field advantage, but what I have never seen asked or answered is does a great pitching team have a park based advantage in a pitchers park over an equally great offensive team? Assuming the overall value of the teams is exactly equal, and in a neutral park they would both be .500 against each other.
Looking back at Texas, opponent hitters LOVE going into Arlington to spread the ball around. While the Rangers’ hitters get 81 games to pad stats, the pitchers are subjected to 81 games of hell… as a GM, you have to find a few ultra-competitive masochists to anchor the rotation.
I think a good idea for someone is to prove the REAL value of relief pitchers / closers, as they continue to be the most coveted “position” in baseball in terms of $/marginal win.
Aka whether there are certain pitchers who pitch better in late/hi lev scenarios and command a premium.
I doubt there is one, but who knows.
It ought to be pretty easy to compare projections made in previous years to the actual performance of the player. Given that, can there be some type of error bars so that one can quickly and easily evaluate how reliable a projection is. Maybe its easier to project a players avg, but the SLG% is far more variable year to year. You would be able to group players together based on their skill set in order to more accurately determine the potential error of your projections (sluggers vs singles hitters). In essence, instead of pinning a player to a single number for any given stat for the year where we all know is that there will be ups and downs, projections should estimate the statistical range in which their talent, development, and skills will land them.
I second this. Error bars are absolutely essential for any projection. It would probably also do away with a lot of the “games aren’t played on computers!” crap that we here if projections weren’t presented in a form that implies precision (even if the author does explain them, the numbers speak louder than anything s/he says).
Not only do I agree, but I think the error bars are MORE important than the mean projection itself. Winning fantasy baseball is not about using the most accurate projections, it’s managing risk better than your opponent. Using RMSE, most sabermetric projections are similar in their accuracy. But very few report the variance in the projections. (PECOTA is one exception.) I want to know which players are likely to bust out, which are likely to regress, etc.
Related to this is my strong belief that the “values” calculated for fantasy are all wrong because they are just a simple function of the mean projections and don’t incorporate risk (or at least don’t do so systematically). I think I know the correct way to “value” players for a fantasy auction, in theory, but it requires a projection system that has a distribution, not just a mean.
Dave-
I was wondering about base-running metrics. I have never fully trusted SB%, SB/SBO%, or even XBT as possible indicators. I have a feeling “smart” base-running might be the next undervalued skill, like defense is now. So my question is if there is some statistic that you have a lot of faith in, or do you think there is room for exploration in how this skill is measured?
As far as an alternative to SB%, I’ve always thought a WPA stat exclusively on SB and SB attempts might be interesting to see.
I don’t know if the data is available, but something I’ve never seen is an attempt to weight pitcher skills with respect to holding runners or picking them off. For example what is more effective in limiting SB’s (or generating outs), quick delivery to home or a quick pick off move?
Also, could a pitcher’s skill in holding a runner provide added value by reducing the chances that the runner moves from first to third on a base hit?
Greg Maddux had a knack for out-thinking hitters, or at least the tendency to throw one pitch when they seem to be looking for another. Maybe knowing when to throw a pitch might be as important as the location or the content of the pitch. This might be too granular for here, but pitch sequencing for individual pitchers and hitters could be revealing.
I have been wondering for a while how defense effects pitcher durability. I’ve had a theory for a while that, because good defense yields fewer men on base, the strain on pitchers should be less. Starters throw fewer high-stress innings and relievers don’t get overworked because they aren’t need as much. That’s the theory, at least. I have no idea if it works out or not and I’ve never seen anything done on it (that doesn’t mean I’m not missing something really obvious, I just haven’t seen it.)
I was just thinking about this over the weekend with respect to how defense affects starting pitchers for my various fantasy baseball leagues. I have not seen where you can pull up UZR numbers on a game by game basis – this might help with seeing if certain pitchers are unluckier Or luckier) than others with respect to how the defense plays behind them. Call it DS – for Defense Support…I wonder if DS is like Run Support which can vary wildly on the same team.
I have this sneaky feeling that the current defense metrics have not completely factored in luck, weather, ground conditions – BUT I have not studied UZR fully to see if or how these things factor in the ratings.
I think team defense has a huge impact on the development of young pitchers and their control. Guys who pitch for bad defensive teams are much more likely to nibble and less likely to trust their stuff. This could lead to more pitches per at bat and less chance of success. In other words, I think if a youngster spends his early years in the game pitching with a bad defense behind him then they are more likely to develop habits that result in less control and less success over the long term. So good team defenses may increase the likelihood of a talented but unproven pitchers success in the long term (less so for big strike out guys), and bad defenses increase injury and less success long term.
Do the results of one start predict the next start? Is there such thing as “streakiness” statistically?
Actually, yes. Pitchers who have been “hot” for four straight starts can be expected to remain hot for at least the next start, according to The Book. I think that’s the only real predictability from streaks, though.
Do you have a link? Would love to read..
Like I said, it was in The Book. I have no idea if it’s available online, though.
Buy The Book! It’s well worth it.
Dave, thanks for this — I was going to email you asking or an open call for questions that require further inspection. Here’s mine: The Armando Benitez rule.
If you’re a Mets fan (well, not *you*), you’ve seen that look Benitez used to get on his face where his head just went poof and exited reality. Wild pitches and grooved fastballs were coming next.
It’s clear, and this applies to more than just Benitez, that some subjective inputs affect a player’s performance. We use Leverage Index but LI doesn’t really matter (for say, clutch performances or choke jobs) if the players involved do not realize the gravity of the situation. Certainly, there’s some way to measure how a player is reacting, be it increased heartbeat or blinking or something. What biometrics should we be looking at to see if a player is in the groove, about to choke, or somewhere in between?
1. Using the available statistics right now, is there a way for us to finally make catcher defense quantifiable?
2. This is something I’m doing currently with the ’09 Twins on my blog behind the scenes, and hopefully can expand to include all 30 teams in 2010. We often talk about players “doing the little things,” but the arguments always seem to be based off of anecdotal evidence. One of these things is advancing baserunners. So, I’ve been looking at a player’s ability to advance and score these runners. I first started with a player’s success rate of driving in runners (both when in RISP and all runners), success rate of advancing runners, and average number of bases the runners advanced, whether or not the hitter made an out. Now, to hopefully better correlate to something like WPA, I’m saying that if a hitter makes an out but still advances a runner, he no longer receives credit for that advancement.* In doing this second part, I also then can find the difference in successes for each player to see how often they advance runners but still make an out. I have yet to finish this second part, but should be done within the next week or two. So now that I’ve gone through a long-winded intro, my question is: What else can we do to understand or demonstrate how valuable it is to advance a baserunner? If we remove the ability of driving in runs from the equation, can we determine how players rate in terms of advancing runners?
* In relation to the previous question, this is another question that I’ve been having difficulty answering. The second part of my research is focusing on advancing runners without making an out. However, the fact of sacrifice flies, RBI groundouts, etc. are difficult because on one hand, the hitter made an out. On the other hand, he drove in a run. An additional problem though is that we can’t guarantee what will happen after a certain event. Sure, the sac fly cost the team an out, but how do we know that the next hitter(s) would have driven in that runner? By putting this all together, does that mean that the hitter gets a small reward for driving in a run, despite making an out, or is this a black and white issue, where we say “No, he made an out, even with a runner scoring, he hurt his team’s chances of winning”?
Correction: “Sure, the sac fly cost the team an out, but how do we know that the next hitter(s) would have driven in that runner?”
Should read as…
“Sure, the sac fly cost the team an out, but how do we know that the next hitter(s) would have advanced or driven in the new runner?”
This is something that RE tables could handle fairly well, actually. Basically, if REafter+1>REbefore, then the run-scoring out is generally beneficial, otherwise not.
Gains:
xx3, 1 out, run scores
1×3, 1 out, run scores, trailer advances
123, 0 out, run scores, trailers advance
Washes:
1×3, 1 out, run scores, trailer remains
123, 1 out, run scores, trailers advance
Losses:
xx3, 0 out, run scores
1×3, 0 out, run scores, trailer advances
1×3, 0 out, run scores, trailer remains
x23, 0 out, run scores, trailer advances
x23, 0 out, run scores, trailer remains
x23, 1 out, run scores, trailer advances
x23, 1 out, run scores, trailer remains
123, 0 out, run scores, lead trailer advances
123, 0 out, run scores, trailers remain
123, 1 out, run scores, lead trailer advances
123, 1 out, run scores, trailers remain
Where the first string refers to the base occupancy at the beginning of the PA, the second the initial number of outs, and the remaining column(s) what the baserunners do on the play the batter makes an out on. Obviously, this is a general run-maximizing scenario, and ignores circumstances when the marginal value of the next run is disproportionately high.
How do individual skills age differently (measured both directly and indirectly) and what can this tell us about player development, performance projection, and the general aging curve.
Your answer is probably here:
http://www.tangotiger.net/
http://www.tangotiger.net/mgl/agecurve.pdf
Look through that whole list — there are several articles related to specific skills and aging. Also, one more very good one here:
http://www.hardballtimes.com/main/article/fielding-aging-curves/
Look around a bit, you should be able to find more. They can be kind of hard to find sometimes, but there have been quite a few studies done related to your question.
Thanks Nathaniel. This item by Tango starts to get at it (http://www.tangotiger.net/agepatterns.txt), but I’d still like to see it played out a bit more. Given an initial set of skills, how can we expect an individual player to age.
And furthermore, what do these curves tell us about replacement level. I suppose one of the items which particularly interests me are the late bloomer types who come up at 26/27, play well for a few years and are out the league by 32/33. These are guys who seem to only crack the surface of MLB during their peak. Brady Clark is a guy who comes immediately to mind.
If a player puts up certain component performances at a given age, can we project him using a generic component-skill development model — a PECOTA using skill based aging curves instead of actual players, if you will?
Tango also has this for pitchers (http://www.tangotiger.net/adjacentPitching.html).
As I think Tango points out, the issue of selection bias is a very real problem. That’s why I’d like to use direct measures of skill (reaction times, hand-eye coordination, strength, etc) rather than performance based ones, if possible. I’m sure certain teams, such as the Indians and Red Sox, are already headed down this road.
It may be a bridge too far to part-time analysts. Though it also suggests another topic, how can we better translate physical abilities to baseball performance. This seems like a holy grail type question for scouting…
I agree that pitching is still a gray area in many ways, particularly in relation to pitching injuries. There’s a lot of research on the correlation between the # of pitches thrown, # of innings pitched, overall usage, days between starts, etc. But – pitchers still throw – and throw a lot – when they’re not on the mound.
So – what I’d like to know is whether they’re worked too hard overall. Do they throw too much in winter ball or on their ‘throw days’? How much does overall, non-game usage contribute to dead-arm, tommy johns, rotator cuff injuries? How much would not throwing between starts affect command and stamina? Or, even better, not throwing at all from Nov-Feb? Do pitchers really need to throw so much in non-game situations to be effective?
Piggyback question: The magical 100 pitch mark (and pitch counts in general) and to a lesser extent, pitcher abuse points (well in the sabermetric community), are prominent in the game now. However, it feels like TV analysts and the fans completely ignore the number of warm up pitches, pitches in between starts/games, and pitches before games. I suppose that it’s tough to keep track of this unless you’re a coach, but does the number of non-game pitches affect performance and/or injury?
In the last few years, people have spent tons of time valuing players in a vacuum and evaluating teams by merely summing up individual’s values. However, I have a hard time believing that the effects of individuals playing together will always be additive. I think there are likely several instances of combinations being synergistic, or possibly disruptive. Nobody seems to have looked at what combinations might be like this or what types of magnitudes these effects have.
To give a broad example of what I’m talking about, it is common knowledge that a hitter’s stats will be a bit inflated if they have the best hitter in the game on deck when they’re batting. But how much, and do certain types of hitters (possibly those with poor plate discipline) benefit more? The previously mentioned question about what effect a catcher calling a game has on the pitcher would also fall into this broad categorical question. Does the catchers game calling matter differently when you have a young power pitcher or an older finesse pitcher? A lot of different permutations of this fundamental question are possible: Does good defense at one position lift that of neighboring positions? Can a good base stealer get his teammates better pitches to see by keeping the pitcher on edge? Is there some optimum ratio of slap hitting singles hitters to three true outcome players in a lineup?
I wondered the same thing recently (as I worked on a WAR team spreadsheet). Is it possible that there are synergies in terms of team composition? I think there is a strong case for suspecting that synergies exist among defensive players on the field. I’m not sure about hitters. But I sometimes wonder if a balanced lineup of different types of hitters (e.g., some contact hitters, some 3 True Outcome Hitters, some high OBP hitters) will outproduce a lineup of very similar hitters (all contact hitters or all high OBP hitters, etc.). This seems like an important question because many teams are attracted to certain types of hitters and they end up with lineups of identical type players. But would a portfolio of different type hitters produce better results?
Dave,
I have always been interested in deliveries and arm angles. Why does a submarine who throws a fastball ten miles per hour slower than other hurlers have success? Also, why do certain certain delivery styles benefit certain types of pitches; such as did Randy Johnson’s angle give his slider more movement or velocity?
In the most recent podcast, the two Daves muse about this same thing. It would be awesome to have some sort of database that detailed how different pitchers change their arm angles or deliveries in different situations.
I think, according to what Cameron said (IRC), that pitchers who change their deliveries/arm angles (of the 5 he examined) still had pretty rough splits. Maybe pitchers who do so are attempting to cover an even worse split? Or maybe they do so against their own benefit? I don’t know.
I think that anxiety disorders will eventually become a very strong topic when discussing pitching mechanics and command. See Dontrelle Willis, Zack Greinke.
Facts about anxiety disorders
—–
The National Institute of Mental Health states that over 40 million Americans, or 18.1 percent of people 18 and older, will suffer from an anxiety disorder this year.
Anxiety disorders represent the largest percentage of mental health problems in the United States.
Although we may think of anxiety as a mental condition, it also usually involves mild to severe physical suffering.
The natural neuro-physiological response to anxiety, commonly called “fight or flight,” makes your body enter a heightened state of physical and emotional awareness. Prolonged periods of this elevated state can cause:
Tremors and twitches
Muscle tension
Headaches
Fatigue
Insomnia
Pounding heart
Sweating
Nausea
Dizziness
Frequent urination or diarrhea
Shortness of breath
What is the value of an excellent pinch-runner? How many wins does an excellent pinch-runner produce over the course of a season?
I’m not a pro at this, and maybe there’s research out on this already, but I’m curious if pitching staffs could be used better. Obviously, you want to get as much use as possible out of your best pitchers in the highest leverage situations as possible. Is running them out there every fifth day and telling them to get you six good innings the answer? What if your best pitcher could give you four good innings every third day instead? Are there advantages to only seeing a pitcher once, and pulling him before facing the lineup a second time through? In little league, it would always through a team off if you pitched a fireballer for three innings them threw a soft-tosser at them for three or vice versa. Could the same be true in the pros? Is it possible that power arms are really effective in the bullpen because the starters may have lost some zip by the time they get pulled? Even outside of an individual game, would a team have struggled to hit Jamie Moyer a day after facing Randy Johnson or vice versa? How long could a day of hitting Tim Wakefield throw off batter’s timings? Could you throw Wakefield in the first game of every series (with him being able to bounce back so easily) and then let your pitchers feast on the affected swings the rest of the series? Even better if the opposing hitters get their timing back on track just in time for their next series to start, so that your team reaps the benefit alone.
Interesting questions, Billy. On a related note, I’ve noticed (anecdotally) that most players have a higher OPS the 2nd time facing a starting pitcher than the 1st time (available from splits on Baseball Reference). That might seem obvious, but I’d like to see that extrapolated some more, and for all players in the league. Is it basically the same thing as the pinch hitter punishment? Do guys hit better on their first time against a starting pitcher when they are home (i.e. played defense first, assuming they aren’t a DH)? Do certain types of hitters (guess hitters vs. reaction hitters) do better their first or second time?
I’d also love to see some more “traditional baseball axioms” challenged (or proven): a leadoff guy that “sees lots of pitches” helps the other hitters, players can purposefully hit into the infield holes when runners are held on, a strong-armed RF is more important than a strong-armed LF in all situations, ability to turn the double play is verifiable and makes a difference, etc.
Perhaps some of those have been addressed but I have missed them.
I don’t think too many people feel that a good arm in RF is more important than a good arm in LF in all situations. It’s only when a runner might be able to advance from second to third on a single or long fly out that creates the desire to put the stronger arm in right. Perhaps a better question would be: “are there other differences in fielding chances between the two positions that might also influence the decision on which players to play there?”.
“Is running them out there every fifth day and telling them to get you six good innings the answer? What if your best pitcher could give you four good innings every third day instead?”
I think the 5 innings for a win rule has held managers back from trying something like this. It makes a lot of sense to me. You get the benefit of not letting hitters become too accustomed to one pitcher, you can better take advantage of R/L splits, and it might well reduce pitcher injuries. It’s something we might see done regularly 20 years from now.
I concur with the questions about injuries to pitchers but with more specific questions. Can pitchfx be used to predict injuries? Are there warning signs of impending doom? ie. would changes in release points, velocity, etc able to point to pitchers about to incur tendon damage.
Are some release points better than others in preventing injuries? Are certain pitch types more prone to result in injuries?
http://www.hardballtimes.com/main/article/the-injury-zone/
I’ve often wondered if mound differences have an impact on command. Beyond the normal home/road splits are there certain parks whose mounds cause more of a command problem to a visiting pitcher than others?
With the recent trend of teams delaying players’ free agency by avoiding super two’s, buying out players arbitration years, and first year of free agency maybe a retroactive study of how effective this strategy is?
I think this was a good move for the Yankees. At best, they get an experienced arm in the pen who can get you out of a tight spot, eat up some innings, or make a spot start. At worst, they are down $1.2 million, which isn’t so bad (in Yankee terms).
These sorts of guys have worked out great for the Yankees the past few years. They pick up 5 or 6 bullpen arms, 2 of them work out, and everyone is happy.
Excellent question!
I think you want the Chan Ho Park thread.
Impossible stuff I’d like to understand better:
managing/coaching – it doesn’t necessarily have to be measured statistically, but evaluated and ranked by whatever it is that makes one better than another.
Developmental Practices – What does each team try to accomplish and how do they go about it. Obviously not all systems are created equally.
Scouting – What methods do the best scouts use and how do they approach and find their targets. rankings for scouts
Building off of this: I’ve been thinking it’d be interesting not so much to rank scouts as to normalize their data. Does one scout consistently rank a fastball a 60, while another ranks the same fastball a 50? If we can see how different scouts rank the same players, over time, we should be able to get a better handle on what each individual scouts’ numbers mean. (This is assuming that data is even publicly available, though.)
I tried to do an analysis of this the other day. I pulled from Pitchfx 4 stats, 1) I called Deception, which was the average difference in release points for the pitcher’s different pitches, 2) was number of different types of pitches thrown (minimum 60 pitch count to qualify) 3) maximum average velocity of fastest pitch thrown and 4) differential between fastest average pitch and slowest average pitch. (I didn’t have time to do “movement” to try to capture the better breaking ball pitchers) I normalized all the data points against a normalized ERA and the inputs explained a whopping 4% of the variation in ERA. So, I just became discouraged and went back to figuring out if I wanted to draft Justin Upton or Carl Crawford. One other person mentioned looking at pitchers vs opposition rather than just straight stats. I’ve always like that idea and once did the analysis (I think I used something like wOBA as my all encompassing stats, but I’ve also done a Strikeout/GroundOut compilation that I like too).. Besides the obvious benefits of seeing who reduced their opponents manhood the most, the analysis shows that the NL West and AL East Pitchers face the hardest opposition with NL West facing the worse, but the differential isn’t much, nothing that makes you look down upon Lincecum. I’ll try to clean that analysis up at some point and actually put it on the World Wide Web.
Is it a good strategy to “try to win every game”? Star players rarely get days off, but when they do, it is typically set up to coincide with a particularly tough opposing starter. Is that smart? If pitching staffs weren’t so rigid with starters pitching every fifth day, could a team improve the odds of season-long (or at least series-long) success by having a lesser pitcher take one for the team against a particularly strong opposing pitcher, affording the team’s better pitchers a greater opportunity to win against lesser pitchers in subsequent games? Is there really no difference between having a 50% chance of winning each game of a three game series versus having a 30% chance of winning game one but a 60% chance of winning each of the remaining games? How much can the sequence of starting pitchers affect the expected win total?
I believe a pitchers ability to command the strike zone is derived from several (difficult to prove) factors. One major one is the ability to easily repeat their delivery. The arm angle, etc isn’t important, rather it’s just important to repeat it each time. I think conviction and confidence are essential as well. There really isn’t a wrong or a right pitch to throw in any situation. The best pitch is always the one that the pitcher believes in and throws with conviction. A common sequence is to throw a curveball down in the zone and then a fastball above the letters. Well, if the pitcher doesn’t believe in the pitch it’s the wrong one. It’s not usually best to throw the change after the curve but if the pitcher is more confident with the change in that situation and executes, it’s the right pitch.
It’s doubtful anyone can prove what results in excellent command beyond repeating delivery. Some pitchers just ‘can’ while others can’t. Confidence is key, in my opinion. Unfortunately, it’s likely quite unmeasurable. Overthrowing (instead of pitching) will also obviously be at the expense of command. Take a Kyle Farnsworth, for example.
Which comes first …
[1] The ability to throw the pitch where you want it, how you want it (command).
or
[2] The confidence to know you can.
*grin*
Are there any metrics that can help us predict playing time? Or a way of assigned a “certainty” score or some other kind of beta that can at least give us a guide about playing time?
Obviously, great ability can be stymied by a lack of playing time and merely above-average ability can seem more impressive if you get a guy who plays in 160 games (see Granderson, Curtis, whose rising HR totals seem to be a function of playing in more games in 2009, rather than actually having more power).
I know there are lots of variables — injuries, development of players in the minors, trades, attitudes of and personal relationships between players and managers, etc., but I recently heard Ron Shandler talking about assigning a score (through his “Mayberry” process) that helps project playing time.
I think what makes pitching so complicated is that there’s two strategy components: one from the catcher-pitcher battery and one from the batter; plus three execution components. Just looking at the basic factors, there seems to be the following:
Pitching Strategy: How and where to throw a pitch.
Batter Strategy: Conditions for swinging and type of swing.
Pitcher execution: Accuracy in hitting locations.with certain pitches
Batter execution: Making contact with the ball as desired, if swinging.
Catch execution: Catching, blocking balls, and fooling umps with glove position
All of these elements are important but there’s no easy way to gauge them. We can’t get direct information about the pitching strategies or batting strategies, so those are always going to be latent variables. All we get to see is the execution. So if a pitcher misses his mark by a foot but strikes a guy out, that’s seen as a good thing (I’m looking at you, Danny Cabrera 2007).
The issue is, we don’t really have a good measure on the intended location so it’s pretty difficult to differentiate between bad pitch calling and bad throwing. It’s not even like we can judge the intended position from where a catcher lines up their glove, because most smart catchers will try to fool batters or umps by lining up in a different spot than the intended one. Plus we don’t generally know who is calling pitches all the time. Sometimes it’s the pitcher, sometimes the catcher, and sometimes the manager.
So really, our issue is the data. We’re getting data that entangles all strategies, plus all executions. For example, assume on a certain start a pitcher throws and extra 20% fastballs. This could mean 5 things:
1. The fastball is working great
2. He can’t locate his secondary pitches well enough to use them
3. He has a bad catcher who can’t catch his secondary pitches well.
4. Strategically, given the opponents, they expect the fastball to work well.
5. The pitcher-catcher strategy is to use more fastballs than usual to throw off batters’ expectations.
The extent to which these matter depends a lot on the skill sets of the players involved. A guy like Pedro, even when his stuff and execution was bad, tended to be very strategic. He’d intentionally change his distribution of pitches over the course of some games (usually withholding a pitch or two) and then suddenly change his pitching style/selection to throw off batters. The better pitchers I’ve seen tend to share this trait. On the other hand, there are also a lot of throwers- especially ones with only a couple good pitches, seem to just go batter to batter and strategize only for the at bat they’re dealing with.
I haven’t usually seen bad catching affect too many games, with the exception of knuckleballers. A knuckler with a bad catcher pretty much makes his primary pitch useless and the whole battery fails.
So those are my thoughts. I would probably think that the key to understanding pitching better is to be able to get at as much of these latent variables as possible. Some of these variables are measurable:
1. Catcher blocking: Catcher success rate in catching/blocking certain pitch types
2. Catcher ump shading: Pairwise comparison of catchers in strike/ball calls across the zone (holds the ump constant, while seeing who extends the zone better).
3. Pitch Control: Pitch location vs. Catcher setup – Median distance per pitch type might be a good thing to look at. Would help show who is consistently wild with certain pitches.
4. Swinging Strikes: Something already measured, but it seems important in this.
There are probably other things as well to look at. If we could get catcher setup positions and eliminate the ones meant to trick people, we could get a very good idea of the pitching battery strategy. I think this would be very informative.
As a former pitcher, I have to ask …. “If the catcher’s mitt does not represent the intended target, then how does the pitcher know exactly where to place the pitch?” Obviously, the catcher’s mitt is the target. Now, some catcher’s may shift the target late, but the mitt is the target. No doubt.
Again, I have to ask, “How does the pitcher know where the catcher wants the pitch, if the catcher is holding the mitt in a different location.” MLB pitchers-catchers don’t have hidden signs where a “up and in target” really means “low and away”. That’s a great way to lead the league in wild pitches and having the ump call some of your strikes as balls.
A good catcher will reveal the location as late as possible, but he still reveals the target. The smartest catchers know to get their chest/body as close to behind the target as possible, since pitchers have this tendency to “throw to the catcher’s body” regardless of where the glove is.
Catchers catching pitches that are centered also allow them to slightly shift or frame the pitches and get marginal balls called as strikes. [I don't think the casual fan really understands how much skill goes into catching, specifically "receiving pitches". Each pitch is caught a slightly different way depending on height, type of break, location, etc. It really is tremendously impressive, and why dummies usually can't excel as a catcher.] Hitting the glove without the catcher moving it at all is a great way to create the illusion that it was a “great pitch”, even if it were off the plate 3 inches or so.
Really, I just wanted to point out that the catcher’s mitt IS the actual target, even though it may be “shown” late.
One of the things I’m most interested in is strategy. Every pitch could be seen as a little game where the pitcher and batter are the two opponents. There should be a lot of strategy involved in selecting which pitch to throw where, when.
A good example of this is throwing Pujols low and outside – but not doing it too much that he starts anticipating and getting better at those pitches.
My instinct is that some batters and some pitchers use more strategy than others. I want to find a way to examine and quantify this. And in fact I’ve been looking into basic components on my site so far. There is a lot to do but I think it’s a very untapped field which could actually be used by teams to improve performance.
Work smarter, not harder
How about player influence on umpires. You always hear the announcers talk about how “If Bonds/Jeter/Giambi/etc… didn’t swing it probably wasn’t a strike” , even after a called strike 3. Somewhat like how the top NBA players have their own set of rules like Jordan. Do umpire strike zones vary based on who the hitter is? Do Willie Bloomquist, Milton Bradley, Cliff Pennington, and Albert Pujols all have the SAME strike zone?
My observation of watching the game for 30 years is clearly “NO”. Not every batter or pitcher gets the same strike zone in the same game.
As I assumed, I think it would be interesting to see the strike zone difference on average between, for example: Andy Pettitte (vet known to throw strikes) and a Brian Bruney (young pitcher with control problems).
How about team cohesiveness? This is something else announcers talk about, “this team really plays well together, they gel”. It would be interesting to explore whether there is any truth to that. I could only really see this affecting defense though. This also raises the question whether a player (Milton Bradley) can negatively affect his teammates?
I would like to know if you could apply psychological principles/game theory (not sure what term I’m looking for exactly here) to pitch sequencing. By this I mean could you teach a pitcher about how hitters think in predicting what the next pitch will be in order to help them fool the pitcher.
Closest example to this I can think of is from Derren Brown. Derren Brown is, for want of a better word, a UK illusionist. In his Mind Control programmes he is well known for “mind-reading” which he achieves through using a combination of “magic, suggestion, psychology, misdirection and showmanship”. Using his knowledge and skill, he appears to be able to predict and influence people’s thoughts with subtle suggestion, manipulate the decision making process and read the subtle physical signs or body language that indicate what a person is thinking.
One of the basic games he will use to demonstrate this is Rock, Paper, Scissors where he is almost unbeatable thanks to his skills at predicting what the other person will do combined with the subtle power of suggestion.
My point is could you identify and teach pitchers body language signs from a hitter to help them predict what the hitter is expecting and thus select a pitch to fool them more easily. Also could you employ subtle suggestive techniques to help fool the hitter? This could be from the pitcher or something else inside teh stadium such as a large fake advertising banner sending across the subliminal suggestion.
This could also work the opposite way round i.e. a hitter reading a pitchers body language to help pre-determine the next pitch; a suggestion hidden within a hitters uniform perhaps?
I hope I’ve made my thoughts clear but it seems to me that research into the psychology of hitting/pitching could be very interesting.
Yes. If a coach wants his pitcher to walk a guy on 3-2 all he has to say is “Don’t walk him”. Lots of coaches do this and it drives me crazy. The last thing your pitcher hears is “walk him”, and well … it’s a negative comment anyway. Telling a batter “don’t strike out” often has the same effect.
If you want your pitcher to pitch down in the zone, don’t say “don’t hang one” or “don’t leave it up” … just say “get a groundball”.
Obviously it doesn’t work every time, nothing does. But, the mind is influenced somewhat by the spoken word.
Don’t believe me, what’s the first thing visualize when I say “Don’t think of zebras?”.
A moose.
As one that works with pitchers a lot, from ages 8 to 18. I would say that to a large degree, it is something that is “inborn”. There is some type of “mind-body” connection that some folks have that just allows them to “see a target, and hit it”, and it seems to the case whether we’re talking about 3-point shooters, marksmen, quarterbacks, dart throwers, etc.
They don’t “better mechanics” everyone, although that certainly can help, nor do they “out practice” everyone, although that certainly can help. It’s likely a mixture of “uncoachable attributes” just as it is why the ball seems to jump off some guys bat more than it does others.
Mechanics and practice can certainly improve a pitcher’s ability to “throw strikes”, and it can improve command to some degree … but the Saberhagens, Maddux’s, and Martinez’s of the world have likely always had outstanding feel and command …. while the Rodney’s, Chamberlain’s, etc likely never will … no matter how much practice, etc. I would find it very unlikely to learn that Bret Saberhagen or Greg Maddux was “wild” in Little League, high school, or any other level. They have likely always been able to throw the ball pretty much where they want to, how they want to.
I get a lot of credit for my 8yo son’s performance, but I readily admit that I haven’t “taught him” what makes him stand out. He’s one of those kids that has what some might call “the mental link” or “body awareness”, where he picks up on things quickly and can do some things (such as hit the target repeatedly) without being able to explain it or achieve it through large amounts of practice.
All I know is that the pitchers I see continually improve, but whether they display a very high level of command is something that does not appear to “coachable in big amounts”. If it were something they could improve significantly through good coaching and lots of practice, they certainly would do it. Lots of guys have better shooting mechanics than Larry Bird, more physical ability, and probably practiced as much … very few have his “mental link”, or whatever the physiological term would be, to where he can see the hoop, figure the angle, and have his body execute the shot, without thinking about it. My guess is it is every bit as natural or innate as is running speed, vertical leaping, arm speed, etc.
You could teach 100 guys Chase Utley’s size, his compact swing, and have them practice it 1000′s of times per day, and there’s a good chance none of them would be the same great hitter with power that Utley is. It’s likely one obvious thing, and it’s not likely something any had “coached in” or “practiced to achieve”.
I know we don;t like “unexplainable things”, but there’s a reason why so few pitchers, even at the elite level, display a dominant level of “command” (throwing what you want, where you want it).
I would like to know …
[1] What percent of the time does a pitcher throw a fastball after throwing a changeup?
[2] What percent of the time does a changeup come after a fastball?
[3] How many pitchers throw back-2-back changeups? What were the results?
Many pitchers only seem to throw a certain pitch after another specific pitch, and that would seem to be a predictable pattern or sequence.
Personally, nothing is more beautiful than a well-throw changeup in a fastball count.
Search John Walsh and Josk Kalks articles at THT.
Let’s get some distribution projections. Death to point estimates.
The question that most interests me is the optimal pitching usage pattern. I doubt very much that the commonly used current pattern (5 starters, 6-7 innings per start, 7 man pen) is optimal from either a pitching effectiveness or from a roster allocation perspective. What is optimal is however entirely unclear to me. It pretty clearly differs depending on the pitching talent available to a club, but even assuming typical talent distribution, there does not seem to me to be a clear answer.
Take the 2010 Blue Jays (who have a more even distribution of pitching talent than typical with the departure of Halladay). If you accept Rally’s pitching projections, you’ve got 10 pitchers within a narrow band of talent, and none of whom could be projected to throw anything like 200 innings in a season. The logic for the 5 starter system for this club seems to be non-existent. What would happen if the club went to a tandem starter 4 day rotation? What would happen if the club went to a 3 day rotation with two or three pitchers assigned to pitch? Or even more radically, what would happen if the club moved to a 2 day rotation with 4 pitchers assigned to pitch? We know that starting pitchers in general lose effectiveness after 50 pitches, and that relief pitchers (as Steve Treder’s articles of a few years ago in THT demonstrates) are more effective than starters. There has, to my knowledge, never been a study of the effect of pitching consistent and predictable shorter stints. Has anyone ever tried it for a reliever for month? Felipe Alou tried something like that with the Expos in the early 90s, but nobody has since, as far as I am aware.
Yeah, um, like how big is your thingy?
I’d be interested if anyone could take a more useful approach to scouting. I know that Baseball America is extremely respected, but there’s a choice quote on the Braves’ #2 prospect, Freddie Freeman, that just infuriates me:
“Freeman has been an RBI machine at every level, thriving with runners in scoring position….His attacking approach at the plate doesn’t lend itself to walks, but Atlanta gladly will sacrifice some on-base percentage for RBIs.”
How does that work, exactly? Also, how can it be that the one of the most respected scouting organizations in all of baseball can write than in a report?