The bottom of the eighth inning of last Wednesday’s Brewers-Angels game was, at first glance, fairly uneventful. Down 4-2, the Brewers called on converted starter Junior Guerra to keep the game in reach. He delivered — two strikeouts sandwiched a groundout, and the team went to the ninth down only two. Guerra is the fourth or fifth option out of the Brewers bullpen; he’s also a perfect embodiment of modern pitching. He threw 16 pitches in the inning, and less than half were heaters — six fastballs, four breaking balls, and six splitters. When he did throw fastballs, though, he put some mustard on them — two hit 96 on the gun, and he’s averaging about 95 mph so far this year.

These two trends — fewer and faster fastballs — are spreading like wildfire across the game. Sometimes it happens in jumps, like the Twins hiring a progressive pitching coach this offseason. Sometimes it happens organically, like Junior Guerra leaning on his splitter and slider a little more out of the bullpen. It’s a game-wide trend, though, and it seems likely to continue. This year, starters and relievers are both throwing their lowest share of fastballs since we’ve had pitch-level data. When they do throw fastballs, though, both groups are throwing them harder than ever before.

Without looking at a single piece of data, you could have convinced me that those two trends were likely true, but I wanted to look into the numbers to know for certain. First things first — we’ll need a consistent sample across years. Taking this year’s stats and comparing them to previous full-year averages won’t work, because pitchers consistently throw at lower velocities in March and April than they do in the year as a whole. In 2018, for example, March and April four-seamers were .2 mph slower than the year as a whole. Thus, we’re going to use data only through April 10th for every season to properly account for this systematic bias. Let’s take a look at that data, shall we? A quick methodological note: I’m excluding cut fastballs, as classification systems have real trouble differentiating them from sliders:

Read the rest of this entry »

In November, I published the results of my research attempting to put a value on minor league prospects. It seems only natural that a similar study on draft picks should follow.

As with prospect valuations, considerable work has preceded mine in the area of valuing draft picks. Sky Andrechuk, Victor Wang, Matthew Murphy, Jeff Zimmerman, and Anthony Rescan and Martin Alonso have all done similar studies.

The work below is less a replacement of the work already done and is more of a continuation of, and addition to, the study of the subject matter. As to why we might want to know this information, creating an expected value for a draft pick helps us to understand and manage our expectations of draftees’ performance. More practically, teams regularly give up draft picks to sign free agents, receive extra draft picks when they lose free agents or reside in a smaller media market, and drop slots when they exceed the highest competitive balance tax payroll threshold, not to mention that some picks can be traded. Determining a value for these picks helps us better understand the decisions teams make regarding those picks.

In some ways, determining draft pick value is a little more complicated than figuring out prospect value. When determining prospect value, players are placed within the constraints of the current CBA, which provides for a minimum salary for roughly three seasons and suppressed arbitration salaries for another three years after that before a player reaches free agency. Draft picks are confined to the same system, but there is also a signing bonus to consider, not to mention slotting rules that are often manipulated in order to move money around to different picks.

Due to signing bonuses and bonus slots, to arrive at an appropriate value for a draft pick, it isn’t enough to determine the present value of players’ WAR in the majors without getting to a dollar figure. We also have to account for the present value in dollars and then subtract the expected bonus.

Before explaining the methodology for draft picks, we can look at the very similar framework used to get to the present value of minor league prospects. From my “Update to Prospect Valuation”:

To determine surplus value for players, I used WAR produced over the first nine seasons of a career, including the season in which a prospect was ranked. Why nine years? In today’s game, most players don’t hit free agency until after their seventh major-league season. By examining nine seasons, it’s possible to account for prospects who were still a couple years away from the majors when they appeared on a top-100 list — as well as late-bloomers who might have bounced up and down between the majors and minors for a full season.

Of course, not all prospects continue to develop in the minor leagues after appearing on a top-100 list. Some debut in the majors right away. Due to the methodology outlined above, such players might be in a position to receive greater credit for their first nine seasons simply because they were closer to the majors when they were ranked. To accommodate this issue, I’ve spread out a player’s WAR over the final seven seasons of the period in question, distributing 10% of it to years three and four before slightly gradually increasing that figure up to 20% by year nine. To calculate surplus value, I’ve discounted WAR by 3% in years No. 3 through 5 (to approximate the impact of the league-minimum salary) and then 15% in year six, 32% in year seven, 48% in year eight, and 72% in year nine. Spreading out the WAR in this way not only mimics a sort of generic “development curve” but also ensures that arbitration discounts aren’t too heavy.

After that, I applied an 8% discount rate for present value. For players immediately ready to play, the extra value they get from the eighth and ninth year is minimized by removing value they actually provided from the first two years and spreading into later seasons. This similarly ensures that the controllable years of players who take longer to develop or reach the majors aren’t treated the same way as those produced by players who contribute right away. A two-win season in 2019 is more valuable than a two-win season in 2021; and this method helps to strike that balance.

Draft picks aren’t as close to the majors as most minor league prospects are. To combat this problem, I used 10 years for college draftees and 11 years for those drafted out of high school, but kept the rest the same as above.

The other difficult issue for draft picks is one of sample size. When I looked at 15 years of prospect lists, it meant we were looking at hundreds of prospects at nearly every single prospect grade. If we did the same for draft picks over 15 years, we only have 15 players at every pick, which isn’t much of a sample. To compensate for this issue, I took a large percentage of the pick in question, and then a smaller percentage on a sliding scale of the next 12 picks. After all, having the third pick in the draft isn’t just an opportunity to take the third-best player; it is the opportunity to choose between a whole host of players. The Astros taking Mark Appel ahead of Kris Bryant doesn’t make the second pick in the draft better than the first. The Astros could have had Kris Bryant, and factoring in the picks that follow helps represent that challenge.

Smoothing things out a bit helps make sure a small sample doesn’t create a bias around a pick. For example, in the years I studied (1993-2007), the third overall pick often performed poorly, but Eric Hosmer, Manny Machado, and Trevor Bauer were taken with the third pick in the three of the four drafts that followed. It wasn’t bad to have the third pick from 1993-2007. It just happened that those picks didn’t work out well.

First round picks were then adjusted upwards slightly so that the actual WAR of the picks and the adjusted value using the method above matched. The values were then smoothed out to ensure the value of the picks moved downward. The smoothing stopped mattering after the second round. After finding the present-value WAR for each pick (I used $9M/WAR), I then subtracted the slot amount for each pick to come up with a current value.

This is what the first 70 picks look like:

The values at the very top of the draft are going to be context heavy. Sometimes, the top pick is a solid 55, like Casey Mize was a season ago. Other years, it might be Bryce Harper. For context, here is how the first round played out last season in terms of bonuses and slots for the pick.

The draft reveals just how important it is for teams to receive a compensation pick the following season when they fail to sign a pick in the current year. While there is certainly lost developmental time and opportunity in losing a pick for one year, losing that pick permanently would be a major loss, and provide considerably more leverage to the players when negotiating contracts.

Moving down, this is what the picks in the third round and below are worth. For the 11th round and below, the median value is used instead of the average given the potential for a few really good picks out of thousands to distort the value beyond what would be a reasonable expectation for that pick.

In practical terms, that means that for the picks in round 20 or later, you might come up with one average player every three years. For picks in rounds 11-20, a team can expect an average player every two or three seasons. The same is true for rounds three and four combined. It’s hard to find good players in the draft after the first round. There’s as much value in the first 100 picks as in the entire rest of the draft. Teams might opt to pay a third round pick a $3,000 bonus to save money and use it elsewhere. That doesn’t mean that we should expect the same performance from that pick as we would a typical third rounder, but we should expect that the slot money the team uses elsewhere will have a value somewhere close to $4 million.

When considering how teams sometimes shift money around from the second or third round to the sixth and seventh round (and vice versa) or use money to sign players above $125,000 after the 10th round, it helps to know how to properly value every dollar spent. For the first 100 picks, where the bonuses are the highest, every dollar spent generally yields five dollars in value. In rounds 4-5, every dollar should yield about six dollars in value, and in rounds 6-10, every dollar spent should yield 10 dollars in value due to the talent available and the small signing bonuses. Given this information, it appears teams might be better off paying slightly less money in the first few rounds while still getting good talent, and shifting some of that money elsewhere in the first 10 rounds. If teams are shifting money from the first 10 rounds to the back of the draft, they need to feel pretty confident in that player’s ability.

In terms of comp picks in this year’s draft, the Arizona Diamondbacks will receive a pick at the end of the first round for losing Patrick Corbin to the Nationals. That pick is worth something close to $10 million. The six small-market teams will receive picks between rounds one and two that are worth $8 million to $9 million each. The other eight small-market picks after the second round are worth around $4 million each, and the same is true for the free agent compensation picks like the one the Dodgers will receive for losing Yasmani Grandal.

Teams signing free agents who have received a qualifying offer generally lose their second pick, and that pick is worth somewhere between $4 million and $10 million depending on where in the draft the team is picking. The Red Sox’s top pick drops down 10 spots this year because they were more than $40 million over the competitive balance tax. That penalty is only worth around $2 million.

There’s further analysis to be done based on whether a player is coming out of high school or college, as well as whether he is a position player or pitcher, but that work will be left to a later date. For now, I hope this is a useful starting point for further study, and for gaining a greater understanding of draftees’ expected production and teams’ decision making.

After Manny Machado and Bryce Harper signed their gargantuan free agent deals, dominos began to fall left and right across baseball — if you’re of sound body and mind, you probably recently signed a multi-year extension with a major league franchise. When a star signs a new contract of any type, articles analyzing the contract’s value are never far behind, and this recent extension spree has been no exception. I wanted to get in on the action, but the analysis has already been done for the most part. Search for a player who recently signed a contract, and you’ll find FanGraphs analysis of it, likely with some dollars-per-WAR analysis. Chris Sale? Jay Jaffe’s got you. Kiley McDaniel covered Eloy Jimenez’s extension. Justin Verlander? Jaffe again. You get the idea. Craig Edwards even wrote about Harper vs. Machado in an exhaustive level of depth, down to figuring out state taxes.

What’s an author to do? Well, there’s one angle that hasn’t been covered for a while, believe it or not. More accurately, it’s been covered by a combination of shrugs and mathematical hand waves: the value of the opt-out in Machado’s (and Nolan Arenado’s) contract. The reason these haven’t been sufficiently covered is simple — they’re difficult to value. If we want to figure out how many wins a player projects for, a methodology exists for that exercise. Sprinkle in a little of the aging curve and the dollar value of the contract, and there’s one level of analysis. If you want to put everything in present-day dollars, it’s just more arithmetic, but the basic shape remains similar. Introducing opt-outs, however, is a step in a wholly different direction.
Read the rest of this entry »

Earlier this week, I focused on creating objective measures by which to examine and value individual prospects and farm systems. Inherent in those objective measures is the knowledge that the rankings themselves are not. Prospect writers like our own Eric Longenhagen and Kiley McDaniel combine in-person scouting with their knowledge and experience of the game, information from industry sources, and statistical data to arrive at well-informed but still subjective rankings and grades. What follows is one study attempting to determine if there has been any historical bias based on the position of a player.

As with the prior studies, I’m using the Baseball America Top 100 rankings from 1996 to 2010. To get a sense of how players were ranked by position, here are the raw numbers for the number of players listed at any given position, with multi-position players listed at both positions.

It should come as no surprise to find pitchers and outfielders first given that they have more starting positions available to them. Generally, a pitcher isn’t going to make a prospect list if the person believes he will be a reliever because the value is less. That might be changing some now, but for the vast majority of Top-100 pitching prospects, the hope is that they will be starters. If we were to divide the pitchers by the five starting rotation slots and outfielders by the three starting spots, shortstops would then have the highest representation on prospect lists. After shortstops, we have pitchers and third basemen, with outfielders slotting in ahead of first basemen and catchers with second basemen way down the list. Conventional wisdom holds that ranking so many shortstops is acceptable because many will eventually slide down the defensive spectrum, taking up slots at third base or over at second, which makes up for the lack of prospects there. Read the rest of this entry »

There are 6,000 or so minor-league baseball players at any given moment. By definition, meanwhile, there are only 100 minor-league ballplayers on any given top-100 prospect list. That means there are also around 6,000 minor leaguers not on top-100 lists — all 6,000 of them still intent on reaching the major leagues.

And many of them do reach the majors. For half-a-dozen years, Carson Cistulli has highlighted a number of prospects who failed to make a top-100 list by means of his Fringe Five series, and some of those players — like Mookie Betts and Jose Ramirez — have gone on to become stars. There should be little doubt that prospects outside the standard top-100 lists have value. Determining how much value, however, is a different and more involved question.

When I attempted to determine a value for prospects who’d appeared on top-100 lists, I was working with a relatively small pool of players. Even 15 years’ worth of lists equates to 1,500 players at most. Attempting to determine the value for every prospect, meanwhile, would appear to be a much larger task. Does one look at the roughly 90,000 minor-league seasons over the same period? That seems daunting. Looking at Baseball America’s team-level prospects lists, which feature 10 players per organization, would provide a more manageable 200 prospects per season outside the top-100 list, but that wouldn’t quite get us where we need to be, either.

And yet, as I’ve noted, these prospects have value. On THE BOARD, for example, there are currently 689 prospects with grades (a) of at least 40 but (b) less than 50 (the lowest grade earned by players on a top-100 lists, typically). It’s these prospects in whom I’m interested. What follows represents my attempt to place a value on them, as well.

Read the rest of this entry »

Over the years, a good deal of effort has been put into determining the value of prospects. Victor Wang, Scott McKinney (updated here), Kevin Creagh and Steve DiMiceli together, and Jeff Zimmerman have all published work on the subject, roughly in that order.

The reasoning behind such efforts is fairly obvious: teams trade prospects for proven players all the time. Finding an objective way to evaluate those trades is useful to better understanding how the sport operates. Indeed, FanGraphs has benefited from those prospect-valuation studies on multiple occasions.

With another year having passed, I’ve attempted to build on the work of others and produce updated valuations of my own. Previous efforts have been very helpful in the process, while the input of prospect analysts Eric Longenhagen and Kiley McDaniel has helped me find results that would be most useful.

In building this study, I set out with the following aims:

• To separate players into as many useful tiers as possible without creating unnecessary distinctions.
• To use as much data as possible so long as it was useful and likely still relevant today.
• To make the valuations as forward-looking as possible.
• To recognize that player development is not linear and that players appearing on prospect lists vary from major-league-ready to raw, Rookie-level talents.

To those various ends, here are some of the parameters of this study:

1. Baseball America’s top-100 lists from 1996 to 2010 serve as the foundation for prospect grades.
When I started the study, I looked at the lists dating back to 1990, separating out position players from pitchers and organizing by year. I found that the evaluations from the earlier part of the 90s — especially those for pitchers — had considerably worse outcomes than those that came after. I debated whether or not to throw out the data. Eventually, though, I decided that since 15 years of prospect numbers were showing decidedly different results, and that there was considerable turmoil occurring within the sport during that time — expansion, a strike, and a lockout — it seemed reasonable to toss the earlier years and go with the assumption that the 1996-2010 lists more accurately represented prospect evaluation today and going forward than the rankings of 25 years ago.

Read the rest of this entry »

The first month of the season was marked by cold weather throughout much of the country. It seemed to have an adverse affect on offense, with power numbers particularly affected. MLB players put up an isolated-power figure of .156 this March and April, which was the lowest mark since April of 2016. Rob Arthur, who has performed extensive research on the juiced ball, noticed the ball wasn’t traveling quite as far in early April even after accounting for weather — this despite a barrage of homers in the spring. Alex Chamberlain conducted some research of his own and determined hard-hit balls and barrels weren’t doing as much damage as in previous seasons, and he wondered if baseballs had been de-juiced. It’s an interesting question that deserves further research.

Chamberlain speculated that MLB had taken the juice out of the ball, potentially through the use of humidors. He found that hitters had to hit the ball harder to get it out of the park. He also observed that, when controlling for exit velocity and launch angle, batted balls weren’t quite doing the same damage as in years past. He concluded that, since we are now well past the cold-weather days of April, the change in batted balls this season is meaningful.

Read the rest of this entry »

Last week, I discussed the consequences of an expanded strike zone on the game, finding that it leads to more strikeouts and fewer balls in play. While some have suggested that a larger zone — by inviting more swings from batters — might actually result in an uptick in batted balls, the observed results don’t support that hypothesis. Whatever gains a larger zone creates in terms of swing rate, they’re negated by an increase both in whiff rate and called strikes, leading to more strikeouts overall.

What that post addressed was what would happen if the strike zone got bigger. This post attempts to answer a similar question — namely, what would happen if the strike zone got smaller?

In order to test the effects of a shrinking strike zone, it’s necessary first to identify an actual instance in which the strike zone has gotten smaller. Fortunately, such an instance exists, thanks again to the research of Jon Roegele, who produced this visual in his piece on the strike zone last year.

That’s the 2007 strike zone on the left and 2017 zone on the right. As you can see, the outside edge to lefties used to be called a lot more frequently than it is now. The bottom of the zone has gotten larger for both lefties and righties (a point addressed in my last post), and the result has been a smaller strike zone for lefties than their right-handed counterparts.

Read the rest of this entry »

There are a lot of strikeouts in today’s game. The most ever, in fact. If the season were to end today, the league’s 22.4% strikeout rate would represent an all-time high, eclipsing the record set in 2017. That record from 2017 surpassed the one set in 2016, which itself surpassed the one set in 2015, which surpassed the one set in 2014. Ever since 2008, actually, baseball has produced a new strikeout record, and there doesn’t seem to be an obvious end in sight.

With all those strikeouts come a lot of opinions on how to reduce strikeouts. The latest set of proposals come from Tom Veducci at Sports Illustrated. Verducci correctly places blame/credit for the strikeouts with the pitchers, where it belongs, and he suggests a few solutions: lowering the mound, limiting the number of pitchers on an active roster, and introducing a pitch clock.

I find it curious that Verducci omits any mention of the strike zone itself. I have previously proposed raising the bottom of the strike zone to put more balls in play, but there are others — including at least one MLB manager — who believe that a larger strike zone might increase the number of balls in play.

The possibility of this effect is one I’ve heard mentioned on broadcasts before, so it isn’t without precedent. The theory goes like this: an expanded strike zone will force batters to exercise less patience and, as a result, swing at more pitches. More swings, and perhaps more emphasis on contact, means more balls in play.

Fortuitously, this is a theory we can test, because the strike zone actually hasn’t remained static in recent years. In fact, thanks to great research by Jon Roegele, we know exactly where the strike zone has gotten bigger. The very bottom of the strike zone has increased considerably over the last decade, and although it got slightly smaller the last couple seasons, the trend has reversed itself this year. Even if there wasn’t an increase this year, the strike zone would still be substantially larger at the bottom of the zone than it was a decade ago.

There is considerable teeth-gnashing going on around the game due to a lack of action on the field. Those criticisms are not unfounded. All things being equal, the game is better with more and not less action. A walk might be nearly as good as a hit when it comes to scoring runs, but it is considerably less exciting. A strikeout does have some excitement of its own, but on a large field that ranges out to 400 feet in most parks, concentrating much of the action to the first 60 feet has some drawbacks when it comes to demanding and retaining the attention of fans.

In any given confrontation, both the pitcher and batter exert considerable influence over the outcome of an at-bat. Because of that, it might seem reasonable to place equal blame on the hitters and pitchers for the increase in strikeouts. In an era defined by greater velocity and more frequent shifts, one argument goes, batters are failing to adjust. If they would just take the ball the other way, they might strike out less, get more hits, etc.

That might be true. It is also possible, however, that changing their approaches might lead hitters to produce less valuable outcomes or, worse, abandon the very strengths that allowed them to become major leaguers in the first place. That isn’t fair to hitters. What I’d like to posit here is a much simpler explanation for the rise in strikeouts — namely, that pitchers are too good.

Fastball velocity has increased at a steady rate, some of that due to the rise of relief innings around the league and some of it probably to dramatic improvements in training and development. That’s not really the point of this post, though. The point of this post is to discuss one particular cause of the increase in strikeouts that likely has little to do with launch angle or players trying to hit home runs, but rather the talent level of the pitchers and a change in philosophy.

Below is a scatter plot of MLB strikeout percentage and average fastball velocity.

Read the rest of this entry »