Top 27 Prospects: Texas Rangers
Below is an analysis of the prospects in the farm system of the Texas Rangers. Scouting reports are compiled with information provided by industry sources as well as from our own (both Eric Longenhagen’s and Kiley McDaniel’s) observations. For more information on the 20-80 scouting scale by which all of our prospect content is governed you can click here. For further explanation of the merits and drawbacks of Future Value, read this.
| Rk | Name | Age | High Level | Position | ETA | FV |
|---|---|---|---|---|---|---|
| 1 | Willie Calhoun | 23 | MLB | DH | 2018 | 50 |
| 2 | Leody Taveras | 19 | A | CF | 2020 | 50 |
| 3 | Cole Ragans | 20 | A- | LHP | 2020 | 50 |
| 4 | Yohander Mendez | 23 | MLB | LHP | 2018 | 50 |
| 5 | Bubba Thompson | 19 | R | CF | 2022 | 45 |
| 6 | Pedro Gonzalez | 20 | R | CF | 2021 | 45 |
| 7 | Hans Crouse | 19 | R | RHP | 2021 | 45 |
| 8 | Ronald Guzman | 23 | AAA | 1B | 2018 | 45 |
| 9 | Chris Seise | 19 | A- | SS | 2022 | 40 |
| 10 | Kyle Cody | 23 | A+ | RHP | 2019 | 40 |
| 11 | Brendon Davis | 20 | A+ | 3B | 2022 | 40 |
| 12 | Mike Matuella | 23 | A | RHP | 2019 | 40 |
| 13 | Isiah Kiner-Falefa | 22 | AA | UTIL | 2019 | 40 |
| 14 | Josh Morgan | 22 | R | INF | 2020 | 40 |
| 15 | Jonathan Hernandez | 21 | A+ | RHP | 2020 | 40 |
| 16 | Anderson Tejeda | 19 | R | SS | 2021 | 40 |
| 17 | Brett Martin | 22 | A+ | LHP | 2020 | 40 |
| 18 | Joe Palumbo | 23 | A+ | LHP | 2020 | 40 |
| 19 | Carlos Tocci | 22 | AAA | CF | 2018 | 40 |
| 20 | Jose Trevino | 25 | AA | C | 2018 | 40 |
| 21 | Matt Whatley | 22 | A- | C | 2021 | 40 |
| 22 | Connor Sadzeck | 26 | AA | RHP | 2018 | 40 |
| 23 | Tyler Phillips | 20 | A | RHP | 2022 | 40 |
| 24 | Jean Casanova | 20 | R | RHP | 2021 | 40 |
| 25 | Alex Speas | 20 | A- | RHP | 2022 | 40 |
| 26 | A.J. Alexy | 19 | A | RHP | 2022 | 40 |
| 27 | Miguel Aparicio | 18 | A | CF | 2020 | 40 |
50 FV Prospects
| Age | 22 | Height | 5’8 | Weight | 187 | Bat/Throw | L/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 50/60 | 65/65 | 50/60 | 30/30 | 40/40 | 45/45 |
Calhoun doesn’t have a position (he’s been tried at third, second, and in the outfield since college), but he’s going to rake. Scouts have him projected for plus hit and power. He takes huge, beer-league-softball hacks but has the hand-eye coordination and bat control to make it work. He could yank out 30 or more homers as soon as he’s given regular at-bats. The corner-outfield and DH situation in Texas is pretty crowded, but he should start seeing regular big-league time this year. There’s some risk that Calhoun’s aggression is exploited the way Rougie Odor’s has been, but otherwise Calhoun looks like a stable mid-order slugger.
| Age | 18 | Height | 6’1 | Weight | 170 | Bat/Throw | S/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 30/55 | 45/50 | 20/45 | 60/60 | 45/55 | 55/55 |
Taveras didn’t have a spectacular 2017, but he was just a teenager in full-season ball, so his .250/.312/.360 line is pretty palatable. He still projects as a plus hitter with plus defense in center field and a plus arm, but questions about his power potential pervade. He’s somewhat projectable, physically, but unless a change in approach is made, his power output and overall ceiling are capped. He still profiles as an above-average regular, but he’s a great developmental distance away from that.
| Age | 19 | Height | 6’4 | Weight | 190 | Bat/Throw | L/L |
|---|
| Fastball | Curveball | Changeup | Command |
|---|---|---|---|
| 45/50 | 40/50 | 60/65 | 40/50 |
Ragans emerged as a sophomore on a loaded Tallahasse-area high-school team with five pro players on it. He isn’t a sexy-upside type of prospect, but Ragans has a clear separator that will get him to the big leagues: a knockout 65-grade changeup for which he has plus feel. His fastball is sneaky productive despite average velocity and his curveball has always been fringy, though scouts assume there will be an average breaking ball of some sort eventually due to Ragans’ ability to get the most out of his other pitches. His upside is a mid-rotation starter and more likely a back-end innings eater, but Ragans doesn’t need to make many more adjustments before he’s ready, and scouts rave about his makeup.
| Age | 22 | Height | 6’5 | Weight | 200 | Bat/Throw | L/L |
|---|
| Fastball | Slider | Curveball | Changeup | Command |
|---|---|---|---|---|
| 55/55 | 45/50 | 45/45 | 55/60 | 45/50 |
Mendez is still up this high because he has above-average velocity — especially for a left-handed starter — and a difference-making changeup. He also had a solid year at Double-A, where he posted 22% and 7% strikeout and walk rates, respectively, over 24 starts. Scouts have grown a little bit concerned about Mendez’s body and athleticism, and the way they impact his command. He struggles to locate his fastball to his glove side, which makes it hard to set up his fringey slider off the plate, away from lefties. Mendez has a loopy, get-me-over curveball that works situationally, but it isn’t going to miss many bats and he shelved it during his brief big-league stint late in 2017. There’s still enough here to project Mendez as a near-ready No. 4 starter, but he’s trending down.
45 FV Prospects
| Age | 19 | Height | 6’2 | Weight | 180 | Bat/Throw | R/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 20/50 | 45/55 | 20/50 | 70/70 | 40/55 | 50/50 |
Thompson was a two-sport star in high school. He threw 38 touchdowns and led McGill-Toolen football to the 7A state title game in Alabama as a senior quarterback. Please enjoy this highlight reel of Thompson dropping 40- to 50-yard dimes.
On the baseball field, Thompson made pretty significant progress as a senior and moved toward the front of the high-school outfield class because he started hitting and was a much better bet to stay in center field than the other top high-school bats. He signed for $2 million and got his first taste of pro ball in the AZL. There, Thompson looked as you might expect a raw, two-sport athlete to look. He generated impressive power on contact but had pitch-recognition issues typical of players focusing on baseball for the first time. Due to knee tendinitis, he didn’t run well in the summer, but the condition required only maintenance and rest, never a complete shut down. As long as that’s remedied, he projects as a plus-plus runner who is at least above average in center field, and he has the present bat speed and frame to merit power projection.
Thompson is a high-variance prospect because it’s tough to project his bat, but he has premium athleticism and makeup. (He arrived at camp early and has been attending college games in Surprise just to drink up more baseball.) Those prospects tend to cure their own ills.
| Age | 20 | Height | 6’5 | Weight | 190 | Bat/Throw | R/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 30/50 | 50/60 | 30/55 | 45/40 | 40/45 | 50/55 |
Gonzalez was tough to scout while he was with Colorado because the Rockies don’t have an AZL team, nor do they play many games during extended spring training or fall instructional league. It meant that Gonzalez spent more time in the Dominican than is typical for a prospect of his age and caliber. He was sent to Texas as a high-end Player to be Named Later in the Jonathan Lucroy deal after dominating a repeat trip to the Pioneer League.
Gonzalez has big power projection because his well-constructed frame has room for another 30 pounds or so. He’s a long-levered 6-foot-5, which causes some swing and miss, but if Gonzalez stays in center field and hits for power, then there’s room for him to have contact issues and still profile as a regular. At Gonzalez’s size, or the size he projects to be, that might be difficult. He’s only a 45 runner from home to first because it takes him a few steps to really get moving, but Gonzalez is a 55 runner underway and has enough range to stay in center field for now.
This is a classic high-variance outfield prospect. Gonzalez might stay in center field and have a middle-of-the-order bat, or he might end up a strikeout-prone corner guy. The light looked like it was starting to come on for him late last year as he dominated instructional league.
| Age | 18 | Height | 6’4 | Weight | 180 | Bat/Throw | L/R |
|---|
| Fastball | Curveball | Changeup | Command |
|---|---|---|---|
| 60/60 | 60/70 | 45/55 | 35/45 |
Crouse has premium stuff, but teams were concerned about his ability to start when he was in high school because he has a strange delivery and well-below-average command. Pro scouts were drooling over his arm after he signed: he was up to 99 in the AZL, sitting 95-97, and mixing in his plus curveball and promising changeup. Crouse’s fastball plays down a little because he’s a short-strider and doesn’t generate good extension, but he has No. 3 starter stuff if he can improve his command enough to pitch every fifth day. If he can’t, his stuff is nasty enough to pitch late, high-leverage innings.
| Age | 22 | Height | 6’5 | Weight | 205 | Bat/Throw | L/L |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 50/55 | 60/60 | 45/50 | 30/30 | 50/55 | 50/50 |
Guzman is remarkably difficult to strike out for a hitter his size and only K’d at a 16% clip last year at Triple-A. His defense has improved dramatically since he first entered pro ball but is still fringey. The same is true for his in-game power, which is average overall but mediocre at first base. His .298/.372/.434 line last year at Triple-A was good for a 112 wRC+. MLB average at first was 113 last year, and Guzman’s defense rounds his profile down. He has great makeup and is nearly big-league ready. He’ll probably have a long MLB career as a second-division regular.
40 FV Prospects
| Age | 18 | Height | 6’2 | Weight | 175 | Bat/Throw | R/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 20/40 | 45/55 | 20/45 | 50/45 | 40/50 | 55/60 |
The 2017 draft was thin at shortstop, and some argued Seise has more upside than any of the other prospects available. Those scouts thought he’d stay at short and grow into big raw power. Even those optimistic about Seise developing pop had concerns about his ability to hit, though. Scouts have knocked his timing, plate coverage, and bat control.
While Seise’s hands and actions need some polish, he’s very graceful and smooth for his size, and has plenty of arm for short. His chances of staying there are pretty good, especially as big shortstops without traditionally viable range are becoming more widely accepted throughout baseball. If he’s a 50 defender at short at maturity and his broad-shouldered, 6-foot-2 frame has grown into above-average raw power, as scouts currently project, he’ll be an everyday player. There’s also a chance he grows off of short, at which point it will be necessary for the bat to develop.
| Age | 22 | Height | 6’7 | Weight | 245 | Bat/Throw | R/R |
|---|
| Fastball | Slider | Changeup | Command |
|---|---|---|---|
| 60/70 | 60/60 | 45/50 | 40/45 |
Cody had a disappointing junior year at Kentucky, but the Twins, who were drafting a lot of relief-only prospects at the time, made him their second-rounder in 2015. He didn’t sign, in part due to post-draft medicals, and went back to Lexington for his senior year, then fell all the way to the sixth round in 2016.
In a year and a half of pro ball, Cody has been healthy and thrown strikes while retaining the stuff that made him a potential first-rounder as a junior. His sinking fastball averaged 95 in 2017 and touched 99. He has a plus, vertically breaking slider that misses bats when buried in the dirt and he’s flashed a 50 changeup. There are scouts who still think Cody ends up pitching in relief, but throwing exclusively from the stretch as a pro has helped him throw more strikes. His stock is climbing, but we’re still talking about a 23-year-old who mostly pitched at Low-A last year and who had some medical red flags as an amateur. Another healthy year of strike-throwing might put those concerns to rest.
| Age | 19 | Height | 6’4 | Weight | 185 | Bat/Throw | R/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 30/40 | 45/55 | 30/50 | 50/40 | 40/55 | 55/60 |
Davis, who is built like an NBA shooting guard, was one of the younger prospects in the 2015 draft class and still has a lot of room on his frame, which means he’s very likely to move to third base, full-time, at some point and he’s already seeing a lot of time there. That likely corner defensive profile is a little concerning when coupled with the strikeout issues caused by the length of Davis’s levers, but he might also grow into enough power to make up for it. Davis doubled his walk rate last season, and if that holds, he could develop into a three-true-outcomes hitter who is also an above-average defender at third base.
| Age | 23 | Height | 6’6 | Weight | 220 | Bat/Throw | R/R |
|---|
| Fastball | Curveball | Changeup | Command |
|---|---|---|---|
| 60/60 | 55/60 | 45/50 | 40/50 |
Matuella was a high-profile college prospect who was in contention to go first overall in 2015 before injuries derailed his stock and arguably his career. He was diagnosed with spondylosis and later blew out his elbow, requiring TJ ahead of the draft. He first stepped onto a pro mound during extended spring training in 2016 and his stuff was electric (93-96, plus breaking balls), but he made just one affiliated start before dealing with more elbow discomfort and was shut down for the rest of the season.
The 2017 campaign was Matuella’s first healthy one in nearly five years (he never threw more than 58 innings in a season at Duke). He made 21 appearances that averaged about three-and-a-half innings per outing, logging 70 frames total. It appears Matuella has scrapped his slider and now works with a three-pitch mix. His fastball is plus. It averages 94 and has heavy sink. His curveball is a 55 and projects to plus as he continues to accrue reps he’s missed due to injury. His changeup, meanwhile, will flash 55 but is generally a 45 or 50. It’s solid No. 3/4 starter stuff depending on how aggressively you project on the changeup, but Matuella has extreme injury risk is far behind the age curve, from a workload standpoint, which dilutes his overall grade.
The standard, annual innings increase for developing pitchers is 20 to 30 innings, which reasonably puts Matuella on track to throw 110-130 big-league innings in 2019. Texas could also just ‘pen him and move him through the minors more quickly, which would increase their chances of extracting big-league innings from Matuella before something else goes wrong. Matuella is a consensus elite makeup guy, and everyone in baseball is rooting for him.
| Age | 22 | Height | 5’10 | Weight | 176 | Bat/Throw | R/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 45/55 | 40/40 | 20/30 | 50/45 | 50/50 | 45/45 |
Kiner-Falefa catches once a week but spends most of his time playing either second or third base. His hands are terrific, and he’s an athletic receiver and ball-blocker with a fringey arm, but he’s a good enough defensive catcher now to be a legitimate part-time option back there. He also sees the ball well and has terrific bat-to-ball skills that allow him to spray relatively soft contact to all fields. He doesn’t have the power to play everyday at second or third base, but he seems likely to be a versatile and valuable bench option.
| Age | 22 | Height | 5’11 | Weight | 200 | Bat/Throw | R/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 45/55 | 45/45 | 30/40 | 40/40 | 45/50 | 55/55 |
Though Texas had been experimenting with Morgan behind the plate on the Surprise backfields for a little while, he only debuted as a catcher in affiliated ball in 2017. He played 36 games at catcher (and seven more in the AFL, where he did nothing but catch) and the rest at shortstop, but mostly abandoned second and third base, where he had seen semi-frequent reps dating back to 2014. Morgan has become a fringey, but passable, receiver with a 45 arm and a solid ground game. He’s a 30 runner (at least, he was in the AFL) and doesn’t have the lateral quickness nor arm strength to play short. He has good defensive footwork, average hands, and an average arm, and scouts considered him a future 50 defender at third, a 45 at second, and anything from a 40 to unplayable at shortstop. He projects as a contact-only backup catcher who can moonlight at other positions.
| Age | 20 | Height | 6’2 | Weight | 175 | Bat/Throw | R/R |
|---|
| Fastball | Slider | Curveball | Changeup | Command |
|---|---|---|---|---|
| 60/60 | 60/60 | 50/50 | 50/55 | 35/45 |
Hernandez looked like a pitchability righty during his first pro season in the U.S., exhibiting impressive command and an advanced understanding of pitching relative to his age. In the two seasons since then, however, he has become an enigmatic fireballer. Hernandez will touch 98 and sits 94-95 with a fastball that has a lot of arm-side tail because of Hernandez’s lower arm slot. That arm slot means his two-plane slider, which is already plus, plays above that against right-handed hitters. It also means scouts are questioning Hernandez’s ability to start because of how early left-handed hitter see the ball out of his hand — a fact that is compounded by his below-average control. He also has an average changeup and curveball, so there’s a viable offspeed pitch and sufficient repertoire depth here to help mitigate Hernandez’s platoon issues, at least. He’ll still have to develop a full grade of control to profile as an inefficient No. 4 starter.
| Age | 19 | Height | 5’11 | Weight | 160 | Bat/Throw | L/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 20/40 | 55/60 | 20/50 | 55/55 | 30/40 | 60/60 |
Tejeda has cacophonous tools and absolutely no polish on either side of the ball. His natural, uphill swing and bat speed could one day allow him to hit for considerable power, but the plane of that stroke causes Tejeda to swing and miss thru pitches in the strike zone. His poor ball/strike recognition doesn’t help in this regard.
At shortstop, Tejeda is error-prone. Some scouts don’t think he has the hands to stay on the dirt and want to see him tried in center field. He runs well enough to give that a try if Texas is inclined. Tejeda’s power and speed clearly represent premium tools, but a lot has to happen for them to play that way on the field.
| Age | 22 | Height | 6’4 | Weight | 190 | Bat/Throw | L/L |
|---|
| Fastball | Curveball | Changeup | Cutter | Command |
|---|---|---|---|---|
| 50/55 | 50/55 | 40/50 | 40/45 | 40/50 |
Martin has missed significant time with injury during each of the last three years. Between 2015 and -17, he had hip, elbow, and finally back issues that shelved him for much of May and June. He has mid-rotation upside if he can stay healthy and more consistently locate a fully developed changeup. Reports from the season have Martin sitting 91-93 with some sink, an above-average curveball, average cutter, and fringe command and changeup. When I saw him in the fall, he was up to 95, sitting 92-94 with no curveball (probably for developmental reasons) but a cutter and changeup that lined up with reports from the summer.
Maritn is long-limbed and has lost many reps due to injury, so there’s reason to continue projecting on his secondaries and command. He’s unlikely to develop a knockout secondary offering but should end up with a deep coffer of 50s and 55s. If he can stay healthy, he could be a No. 4 starter, but he hasn’t shown he’s able to yet.
| Age | 22 | Height | 6’1 | Weight | 168 | Bat/Throw | L/L |
|---|
| Fastball | Curveball | Changeup | Command |
|---|---|---|---|
| 60/60 | 55/60 | 40/45 | 40/45 |
Palumbo’s velocity ticked up in 2016 and didn’t regress when Texas moved him into the rotation late in the year. Scouts thought Palumbo’s combination of velocity, deception, and plus breaking ball gave him mid-rotation potential, but they wanted to see him hold this newfound heat every fifth day for an extended period of time. He was 90-94 and touching 96 during 2017 spring training but blew out his elbow after three great starts and needed Tommy John. Texas added him to the 40-man this offseason and moved him to the 60-day DL to make space for Jesse Chavez. He’s on track to start throwing off a mound soon.
| Age | 21 | Height | 6’2 | Weight | 160 | Bat/Throw | R/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 45/50 | 30/30 | 20/30 | 60/60 | 50/55 | 60/60 |
Tocci, Texas’s Rule 5 pick from Philadelphia, slashed .307/.362/.398 as a 21-year-old at Double-A Reading last year before spending a fruitless final few weeks at Triple-A Lehigh Valley. He’s a skinny 6-foot-2, 160 pounds and has very little raw power, but Tocci has good bat-to-ball skills, is an above-average runner with good instincts in center field, and has a plus arm. His lack of power will likely prevent him from everyday duty; instead, he profiles as a competent fourth outfielder.
| Age | 24 | Height | 5’11 | Weight | 211 | Bat/Throw | R/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 40/50 | 45/45 | 20/30 | 30/30 | 50/55 | 55/55 |
Trevino had a bad offensive season away from the friendly confines of the Rangers’ former Cal League affiliate, and his projection has shaded toward backup rather than second-division starter. He remains a good receiver with an average arm and every intangible quality you can possibly imagine, which is probably why he’s come so far so quickly as a catcher after he spent his college career playing all over the infield. Trevino is one of four catchers currently on the Rangers’ 40-man — five if you count Isiah Kiner-Falefa. Robinson Chirinos is atop the depth chart and mashed lefties last year, but there’s no obvious platoon partner anywhere in sight. His contract expires after the 2018 season, so Trevino has a clear path to playing time in 2019.
| Age | 21 | Height | 5’10 | Weight | 200 | Bat/Throw | R/R |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 20/30 | 40/45 | 20/30 | 45/40 | 50/60 | 60/60 |
An excellent defender with high-end makeup, Whatley has a plus arm and quiet hands that steal strikes from umpires. He has an aggressive, pull-heavy swing but below average bat speed, so it’s unlikely he does any sort of damage with the bat. He profiles as a glove-first backup.
| Age | 25 | Height | 6’7 | Weight | 240 | Bat/Throw | R/R |
|---|
| Fastball | Slider | Curveball | Changeup | Command |
|---|---|---|---|---|
| 60/60 | 55/55 | 50/50 | 40/45 | 40/45 |
Texas developed Sadzeck as a starter until his age-26 season, but it appears he’s moving to the bullpen for the 2018 campaign, fulfilling the scouting prophecy laid forth for Sadzeck when he was drafted seven years ago. He has a plus fastball that will probably play up out of the bullpen (his fastball has been sitting in the upper 90s this spring, but he’s struggled in two appearances), and it could mean Sadzeck pares down a diverse but mediocre group of secondary pitches to a singular offspeed offering. Scouts have long preferred his slider, but Sadzeck has also utilized a curveball and changeup. He profiles in middle relief.
| Age | 19 | Height | 6’5 | Weight | 200 | Bat/Throw | R/R |
|---|
| Fastball | Curveball | Changeup | Command |
|---|---|---|---|
| 55/60 | 50/55 | 50/60 | 30/45 |
Texas sent Phillips back to extended spring training after seven tumultuous, early-season appearances at Low-A in 2017. After a month in Arizona, he went up to Spokane (where he pitched in 2016) and dominated, striking out 78 and walking 11 in 73 innings. Phillips has 40 command (and it looked worse than that when I saw him in the fall), but he’s a huge, northeastern prep arm who has made only 17 starts in three pro seasons, so there are legitimate reasons he has yet to polish his usage of very enticing stuff.
Phillips sits 90-93 and will touch 95 with heavy sink. His changeup flashes plus surprisingly often, and Phillips’ curveball is average but flashes above, and most scouts have it projected there. There’s a potential mid-rotation arm in here somewhere if the control/command really come and enough stuff here to project him as an inefficient No. 4/5 or multi-inning relief piece if it doesn’t.
| Age | 20 | Height | 6’3 | Weight | 155 | Bat/Throw | R/R |
|---|
| Fastball | Curveball | Changeup | Command |
|---|---|---|---|
| 50/55 | 45/50 | 45/50 | 45/60 |
Texas has scooped up a bunch of interesting size/athleticism/arm-strength lottery tickets in the late rounds of the draft. You’re about to meet several, but Phillips and Casanova are the best of them. He is a rail-thin 6-foot-3, 155 and is so slight of frame that it’s debatable whether or not he actually has physical projection. Casanova sits 89-92 and will touch 95. He has plus command projection and has already shown an ability to locate his fastball to all quadrants of the zone. He has a fringey 12-6 curveball and changeup. Both project to average and elicited a surprising amount of swing and misses last year, so perhaps there’s something going on here that allows his stuff to play up. He projects as a No. 4 starter with plus command.
| Age | 19 | Height | 6’4 | Weight | 180 | Bat/Throw | R/R |
|---|
| Fastball | Curveball | Changeup | Command |
|---|---|---|---|
| 60/70 | 50/60 | 40/45 | 30/45 |
Speas’ strike-throwing didn’t develop at all in 2017, and he pitched out of the Spokane bullpen for most of the summer. As a result, the amateur reports indicating Speas would likely become a reliever have turned into pro reports assuming he’s already on that path. His fastball has been all over the place the last two years. He was up to 99 as a high-school senior, then topped out at 95 that fall. He was 93-97 during 2017 spring training, and during the summer, I have him topping out at 99 but only sitting 92-94 at other times.
Speas went from throwing every sixth or seventh day as a starter to throwing every third day as a reliever, and it’s possibles this transition caused his velo to fluctuate. It’s not something about which to worry right now, and we still have Speas’ fastball projected to a 70 out of the bullpen, but it’s worth monitoring during the 2018 campaign. He also has a two-plane breaking ball (some scouts call it a curve, others a slider) that projects to plus, but he hasn’t had the opportunity to hone his changeup in games because of the late-season move to the bullpen. Speas’ needs to make significant strike-throwing progress to be any kind of big leaguer at all, but he has late-inning stuff if he can.
| Age | 19 | Height | 6’4 | Weight | 195 | Bat/Throw | R/R |
|---|
| Fastball | Curveball | Changeup | Command |
|---|---|---|---|
| 50/55 | 55/60 | 40/45 | 30/40 |
Alexy was an 11th-round pick out of a Pennsylvania high schol in 2016 and exhibited a pretty significant uptick in velocity under the Dodgers’ player-development program before moving to Texas as part of the package for Yu Darvish. Alexy was 88-92 after he signed but was sitting 90-92 and touching 95 before the trade. He has a tight, knee-buckling curveball that projects to plus, but his max-effort delivery and mediocre changeup feel make it likely that he winds up in the bullpen. Northeastern prep arms typically have more changeup/command projection because those are things they haven’t needed nor had time to develop as amateurs. Because of that, Alexy has some chance to start. To get there, though, he’ll have to reign in the control problems that became increasingly problematic as last year wore on.
| Age | 18 | Height | 6’0 | Weight | 165 | Bat/Throw | L/L |
|---|
| Hit | Raw Power | Game Power | Run | Fielding | Throw |
|---|---|---|---|---|---|
| 20/55 | 30/40 | 20/30 | 55/55 | 45/60 | 40/45 |
Aparicio is a polished defensive center fielder, but he’s not a plus or better runner, so his ceiling as a defender is somewhat limited. He also has terrific bat control and hand-eye coordination, but he’s not selective enough to attack only those pitches he can drive. He settles for lots of subpar contact as a result. That will have to be remedied if Aparicio is going to profile as an everyday player, because he doesn’t have much raw power and his frame doesn’t suggest he’ll grow into any. There’s a chance his hit tool maxes out as he refines his approach with age, which would allow to to profile as a soft 50. It’s more likely he develops into a fourth outfielder. His mean projection looks something like what Carlos Tocci is right now.
Other Prospects of Note
Ryan Dease, RHP – Texas’s fourth-rounder from 2017, Dease’s 90-93 mph fastball has some life, and he has impressive feel for a potential plus changeup. He can already run the latter pitch back over the outside corner against righties, something not many pitchers can do with their changeups at all, let alone at age 18. Dease has a 40 breaking ball, which is why he isn’t in the main section of the list, and looks like a potential No. 4/5 starter.
Ronald Herrera, RHP – The Yankees’ cup runneth over with 40-man-worthy talent, and Herrera was lost amid a surfeit of pitching in the upper reaches of the New York’s farm system. He was traded to Texas as part of what has become an annual exodus of 40-man spillover in exchange for LHP Reiver Sanmartin. Herrera has an above-average changeup and command. Every other aspect of the repertoire is fringey, but Herrera has had success up through Double-A and projects as, at least, a valuable depth arm and possible back-end starter.
Keyber Rodriguez, SS – A switch-hitting Venezuelan shortstop, Rodriguez signed for $1 million last July. He has 40 raw pull-only power as a right-handed hitter and a more conservative style of hitting from the left side that leads to low-lying contact. Rodriguez has a pretty good shot to stay at short. He’s an above-average athlete with a 55 arm, and his hands and range are fine for the middle infield.
Tyree Thompson, RHP – Thompson is a fastball/changeup/physical projection righty who was clearly asked to focus on fastball-command development in 2017 because he rarely threw anything else. His heater approaches the plate at a good angle and he’ll touch 94. His frame (6-foot-4, 165) has some room for mass yet at age 20. Thompson has shown glimpses of good fastball command, and he’ll show you a 50 or 55 changeup, though scouts don’t have a firm grasp on Thompson’s loopy rainbow curveball because he hasn’t used it much. Realistically, Thompson is back-end starter, but because he has arm strength, athleticism, and physical projection, there’s a chance he turns into more. Texas’s system is often full of guys like this.
Damian Mendoza, RHP – Mendoza was the best Mexican prospect eligible to sign last July. His delivery is graceful and controlled, he’ll top out around 91, and also shows some breaking-ball feel. He doesn’t have a prototypical big-league pitcher’s frame, but he is somewhat projectable and checks most of the other boxes.
Eric Jenkins, CF – Jenkins is a 70 runner with breathtaking range in center field, but he’s made no progress as a hitter and scouts are starting to wonder if he’ll even be a viable bench outfielder at this point. Last spring, he showed signs of limiting the scope of his offensive approach to something that took advantage of his speed, but that wasn’t the case during the regular season. He could still slap and slash his way into a big-league role but he needs to show progress soon.
David Garcia, C – Most high-profile internationl prospects play in either the GCL or AZL at some point during their first pro season, but Garcia — who signed out of Venezuela for $800K in 2016 — was badly overmatched by the other prospects during his first pro instructional league and spent his whole first year in the DSL. When he was back in the states for his second instrux, he had made some clear progress but still looks likely to be a multi-year Rookie-ball prospect. He has promising defensive attributes (he’s a 45 receiver with an average arm already at age 18), and he has a good left-handed cut. He’s an interesting long-term prospect at a premium position and is probably a half-decade away from the majors.
Yohel Pozo, C – Pozo is going to challenge Willians Astudillo for the Three False Incomes crown. He tracks pitches exceptionally well and has great bat control, which makes him difficult to strike out. But he’s also overly aggressive and doesn’t walk. Both his career walk and strikeout rates hover around 6% (Astudillo’s are each about 3%, which is mind-boggling), and Pozo actually has a pretty good chance to catch. He’s a 45 receiver with a 45 arm. It’s hard to project a prospect this weird, and catchers without power have a hard time being better than a 45, so it’s fair to say Pozo’s ceiling is limited despite his unique traits and then project him as a backup.
Ariel Jurado, RHP – Jurado looked like a pretty stable No. 4 starter prospect after 2016. The sink on his fastball evaporated last year, and he was working with a 30 breaking ball. His arm slot was higher in 2017 than it was in 2016, so perhaps there’s some relationship with that mechanical change and a dip in Jurado’s stuff. There’s a chance he bounces back and returns to form as a likely sinker/changeup/command No. 4 starter, but if he’s as hittable this year as he was last, he’ll project as a minor-league depth arm.
Adam Choplick, LHP – Choplick is a massive 6-foot-9 lefty reliever whose fastball plays like it’s in the mid-90s thanks to an extreme combination of deception and extension caused by the huge stride Choplick takes off of the rubber. He lands maybe a shoe’s length from the grass in front of the mound. Choplick is a below-average athlete with below-average control and an average curveball, but the ball is tough for hitters to pick up out of his hand one inning at a time, and he has a good chance to carve out a lefty relief role.
Yanio Perez, 1B/DH – Perez got off to a hot start at Low-A but cooled off after he was promoted and didn’t look good in the AFL. He has 55 raw power, but he’s huge, nowhere near his listed 205, and really only playable at first base. It’s highly unlikely he has the bat-to-ball skills or approach to profile there. Instead, he might max out as a corner bench bat.
Demarcus Evans, RHP – An imposing mound presence at 6-foot-4, 240, Evans struck out 81 hitters in just 60 innings last year but also walked 40. He sits 93-96 with nasty movement and has a plus, low-80s curveball. As you can probably guess based on his walk total, though, Evans has control issues. He projects in relief with his ultimate role likely dictated by how much his control develops.
Andy Ibanez, 2B – Ibanez repeated Double-A last year as a 24-year-old and produced a statline nearly identical to that of his 2016 campaign, which resulted in a 102 wRC+ in the Southern League. He also began seeing some time at third base. There are contact skills here but probably not enough power on contact for Ibanez to be more than a fringe bench bat.
Andretty Cordero, 1B/LF – It’s going to be tough for Cordero to profile because he’s an impatient right-right first baseman, but he does have above-average power, makes an average amount of contact, and will probably stick around for a while because he’s a good-bodied athlete, and those guys have longer shelf lives. If he develops a more patient approach during that time, maybe he finds his way to the majors.
Keithron Moss, INF – Texas signed Moss out of the Bahamas for $800,000 in December. He’s a twitchy, switch-hitting infielder with high-end speed. He’s small and is a long-term developmental project both physically and technically, but he’s an interesting teenage athlete from a location that has begun producing more and more interesting baseball prospects.
Joel Urena, LHP – This is a gigantic, 6-foot-5, 235-pound 18-year-old whom the Rangers drafted out of Monroe CC in New York in the seventh round. He has an average fastball/breaking-ball combo right now, and there might be a little more in here if he becomes more mechanically efficient. Realistically, he projects as a lefty reliever.
Seth Nordlin, RHP – Nordlin sits 86-91 and has an above-average curveball that’s likely to tear up the lower levels of the minors. If his fastball ticks up as a pro, he could become relevant pretty quickly.
Melvin Novoa, C – A strong-bodied 21-year-old Nicaraguan catcher, Novoa slashed .281/.338/.467 in the Northwest League last year. He’s a 40 receiver with a 45 arm and has a decent ground game. He’s got a shot to wind up with a 50 glove and arm, and he could at least be a backup who hits for strength-driven power with that kind of defensive profile.
Kole Enright, INF – One year ago, Enright looked like a multi-positional utility bat who had a shot to hit his way into an everyday job. Then he posted a disappointing .233/.314/.323 line in an aggressive assignment to Low-A. Scouts saw a drop-off in bat speed and general twitchiness. He’s still on the radar but needs to bounce back in 2018.
Jairo Beras, RHP – Beras, 23, moved to the mound in 2017 after struggling to make contact yet again at Low-A. He has little more than premium arm strength at this point, sitting 95-98 and touching 99 during my instructional-league look.
Chad Smith, OF – Smith is a three-year Rookie-ball guy who had on-paper success at Spokane last year. He has above-average bat speed, but his long, sweeping swing makes him strikeout prone and he’s a corner-only prospect.
Luis Yander La O, 2B – La O’s bat head drags through the zone, so he’s late on some stuff he shouldn’t be, but he’s got some pop when he catches one up and in and the hand-eye coordination to hit is here. He’s already 26, so La O has to come quickly. He signed for $110,000 in January of 2017 and has several middle infielders to pass on the org depth chart to reach the majors.
Jose Cardona, OF – Cardona has wonderful bat control and hand-eye coordination — and he lifted the ball in the air more often last year — but the overall quality of his contact took a step backward. He looked like a likely bench outfielder heading into 2017 and now needs a bit of a bounceback to recapture that projection.
Michael De Leon, SS – De Leon peaked, physically, at a pretty early age and reports on his already middling tools have begun to regress. He’s still a pretty advanced 21-year-old with the moxie to pass at shortstop, but he’s looking more like upper-level depth at this point.
Jeff Springs, LHP – As a 24-year-old at High-A, Springs struck out 146 hitters in 112.1 innings. He’s got a dandy changeup and sits 90-92, maxing out at 94.
Selected by Carson Cistulli from any player who received less than a 40 FV.
Jose Cardona, OF
Cardona appeared among the Fringe Five a couple times last year in that space designated as the Next Five — or, roughly the equivalent of an honorable-mention section. Cardona, like many of the players included in that weekly column, makes contact frequently. Unlike many of those players, however, Cardona also appears to be a legitimate defensive asset at a premium position.
Consider his fielding numbers from the last three seasons, per Baseball Prospectus:
| Year | Level | PA | Pos Adj | FRAA | Defense | Def/600 |
|---|---|---|---|---|---|---|
| 2015 | Low-A | 526 | 1.1 | 16.4 | 17.5 | 20.0 |
| 2016 | High-A | 453 | 0.9 | 2.4 | 3.3 | 4.4 |
| 2017 | Double-A | 465 | 0.5 | 9.3 | 9.8 | 12.6 |
| Average | — | 481 | 0.8 | 9.4 | 10.2 | 12.7 |
**FRAA = Fielding Runs Above Average.
***Def/600 = Total defensive run prorated to 600 PA (i.e. roughly full season).
Most relevant here, probably, is that figure in the bottom-right corner, where one finds that, prorated to 600 plate appearances, Cardona has recorded nearly 13 more runs than average defensively per season. By comparison, only 12 major leaguers (catchers excluded) have exceeded that mark during the same timeframe.
| Rank | Name | Team | PA | Pos | Fld | Def | Def/600 |
|---|---|---|---|---|---|---|---|
| 1 | Andrelton Simmons | – – – | 1713 | 19.0 | 48.3 | 67.3 | 23.6 |
| 2 | Kevin Kiermaier | Rays | 1370 | 4.9 | 45.2 | 50.1 | 21.9 |
| 3 | Brandon Crawford | Giants | 1754 | 19.1 | 39.3 | 58.4 | 20.0 |
| 4 | Adeiny Hechavarria | – – – | 1394 | 16.2 | 27.9 | 44.1 | 19.0 |
| 5 | Addison Russell | Cubs | 1506 | 14.4 | 33.2 | 47.6 | 19.0 |
| 6 | Francisco Lindor | Indians | 1845 | 18.2 | 37.2 | 55.4 | 18.0 |
| 7 | J.J. Hardy | Orioles | 1143 | 13.4 | 19.1 | 32.5 | 17.1 |
| 8 | Jose Iglesias | Tigers | 1456 | 17.0 | 22.6 | 39.6 | 16.3 |
| 9 | Billy Hamilton | Reds | 1547 | 5.3 | 35.4 | 40.7 | 15.8 |
| 10 | Kevin Pillar | Blue Jays | 1844 | 5.9 | 42.6 | 48.5 | 15.8 |
| 11 | Zack Cozart | Reds | 1229 | 12.0 | 15.5 | 27.5 | 13.4 |
| 12 | Michael Taylor | Nationals | 1180 | 2.1 | 23.9 | 26.0 | 13.2 |
**Fld = Fielding runs
***Def/600 = Total defensive run prorated to 600 PA (i.e. roughly full season).
Defensive data at even the major-league level is, of course, something to be handled cautiously. Nevertheless, Cardona’s numbers (which are similar by Clay Davenport’s methodology) point to what could be an elite outfielder, thus taking pressure of the bat.
You can learn a lot about a team’s talent preferences and biases by looking at what the lottery-ticket prospects in the system are like. Upper-crust prospects like Taveras and Thompson are sought by every organization, but not all teams would be interested in someone like Evans and Pelham in the draft or Alexy, Davis, and Gonzalez in trades. Texas fills their backfields with big, explosive athletes, some with more refined baseball skills than others, and then tries to develop them. They’ve been doing it this way, mostly with success, for a while now.
The organization also has a long track record of success on the international market and finds players from all over the world. They now have several late-round arms whom the industry considers prospects, and the system is full of up-the-middle talent. Texas is very likely to add two more good prospects in what looks to be a deep June draft (they pick 15th), and they’ve positioned themselves to sign Cuban OF Julio Pablo Martinez by acquiring international pool space.
Eric Longenhagen is from Catasauqua, PA and currently lives in Tempe, AZ. He spent four years working for the Phillies Triple-A affiliate, two with Baseball Info Solutions and two contributing to prospect coverage at ESPN.com. Previous work can also be found at Sports On Earth, CrashburnAlley and Prospect Insider.
Love reading these, thanks Eric! One question on Tyler Phillips (#23). On the chart you list his present Command at 30, but you say it’s 40 in the comments. Is one of them a typo?
There were so many errors in the Astros list, and now this? Hate to say it, but I’m starting to question the veracity of every prospect list here. The scouting numbers and writeups are why I visit these lists. If they’re wrong, or wrong enough that we can’t rely on them being right, there’s no reason to read them.
Lol
Any other self-proclaimed champions of evaluation want to chime in?
I think it is important to note that Kiley and Eric are passing along, at least in part, industry consensus. In that regard, it’s obvious that they are first class, comprehensive, probing, attentive to detail, and constructive with a flair for well balanced discussion.
But for the most part, eye test scouting isn’t bold- because ‘networking’ is such a key part of the professional remit.
I’ll give too examples, Robles is a pretty good one. He has very obvious plus-plus tools. Any 70 year old or 7 year old can hardly fail to pick up on his plus-plus tools. You can see him run, you can see the bat speed, you can see him avoid Ks, you can see him throw, and you can see him get to fly balls- all five tools are clearly visible to anyone watching. Nearly every traditional scout will like Robles, partially because ‘taking industry valuation’ into consideration is a part of the job.
KATOH, however, will be predictably lower on Robles because he lacks elite patience and comes a tad short of elite power, so he merely projects as a great prospect rather than an absolute top prospect. If Robles turns out to be below average in the field, as he briefly tested last year, the eye-testers will not really concede his defensive shortcomings (see Hosmer) because they have no epistemological means to actually quantify his performance. They will continue to see a fast player with a good jump and ‘instincts’ and therefore assume that he is plus in the field.
In this way, Robles is a bit like Andrew Wiggins- his tools are so super obvious that eye-testers have trouble looking past them. Wiggins, it turns out, is a horrible defender- but who seriously has an eye test sharp enough to realize why? I certainly do not. And even if watching a few Wiggins games could lead a top eye tester to see that he is a below average defender- the eye test is hopeless to measure exactly how far below average, and how his defensive performance matches up with his other contributions.
I’ll mention Leody Tavares also in a similar vein. He is a player who is very likely loved more by scouts than by data analysts across the board. He’s young with some pop, some speed and some strikezone control: so he won’t be disliked by analysts, but he is mostly a bundle of obvious tools with poor hitting performances (RC+s of 104, 59 and 96 aside from a fantastic 193 posted in 45 PAs in rookie ball), so his projections will be nowhere near Willie Calhoun’s. Tavares is simply a classic ‘scouts guy,’ and my point is that these players usually are overrated by scouts and more accurately rated by data analysts- with the same being true for classic ‘data guys’ like Calhoun.
As far as the errors mentioned above- trying to cover anything comprehensively with the eye test will lead to tons of errors. How can any person’s eye tests on ~900 players be current? Obviously, they cannot. Nor can a player’s eye test assessment of 900 players be comprehensive- who can watch 900 players 100+ times per season? No eyes can, but computers can.
I don’t love big data, but clearly we all know that sport is the one major high-grossing industry in the world that gives ‘performance assessment’ jobs to eye testers with limited scholarly training. There is nothing about sport that is not best measured with basic modern business methods- eye testing has a romanticism about it, but as with medicine, architecture, finance, and anything serious, it is not a serious approach for accuracy.
I think he did that because, as he said in the write up, Phillips control (from reports) is 40, but was worse when EL saw him.
I love a lefty with a plus changeup, and Texas has two possibilities here. I’m hoping Mendez and/or Ragans pull through.
These lists are actually pretty inaccurate year to year, just a regurgitation of common industry analysis plus some eye test.
Unlike KATOH or the fringe five, these eye test lists don’t have a sound methodology or provide any notable insight. Who is the last under the radar guy on these fangraphs eye-test lists to do far better than projections would expect? I cannot think of a single one and have been following them for years.
Unlike KATOH et al, there doesn’t seem to be any public accounting for the accuracy of the lists. The miss rates are huge, the insights are very rare, and the particular assessments are always tools-heavy, less concerned with plate discipline, and relatively conformist. It’s simply as if a top hospital also hired some smart witch doctors,- to show that they could do that too- just to appease the rabble. Given the task, the work done seems standard enough, but there simply isn’t anything that Kiley or Eric can pick up with their eye that has any value: neither is their networking providing insights that prove worthwhile either. It’s just standard industry eye test, romanticized but ultimately regressive and unproductive.
It’s fun to read the nearly comprehensive coverage, but the eye test simply cannot and does not pick up on anything of particular value compared with the error and bias involved.
These lists are a tip of the hat to the ‘make baseball great again’ crowd and the traditionalists, mimicking the popular features at BA (and – fair enough- getting lots of clicks) but the eye test lists don’t offer much as far as substance.
Olson was a 40 FV last year, Hoskins a 45, Judge a 55. The write ups are the same- ‘scouts are concerned with Hoskins tools…fringe player.’ Olson said to have 40 hit potential and 50 game power potential. With both players, KATOH was more optimistic and insightful.
To be fair, the job appears to be ‘just write a summary of what scouts think, plus add an observation or two from your eye test watching the player a couple of times or at least once-‘ so in that sense, mission accomplished, but this is a concession to regressive analysis.
Thom, I’m normally a big fan of yours, but think you’re way off base on this one. KATOH is great for finding gems based on “scouting the stat line” that FanGraphs’ and other publications’ scouting/”eye test” miss. KATOH though comes up with a lot of bizarre results like its Rule 5 predictions where several players had higher forecasts than a consensus top-25 prospect like Mitch Keller and yet only 4 guys were actually picked and these 4 lined up more with Eric Longenhagen’s rankings.
In fact, when trades are made, I think the prospect values implied in the trades line up much more closely with scouting rankings (and you can take an average of them if you’re concerned about variance from publication to publication) than with KATOH rankings.
I like to scan Eric’s lists of teams’ prospects just because I enjoy trades in general and like to see the value exchanged for a veteran player. I’m not sure how else you’d assess a trade of some major-league player you’ve heard of for some minor league player(s) you haven’t. I don’t think KATOH is good at this.
I also like Eric’s writing, Kiley’s too- and to be fair, they aren’t necessarily looking to come up with an under-the-radar scoop like Carson does with Fringe Five. I agree with you that these writers do a good job on the remit: delivering an assessment of how the industry sees various prospects based largely on eye test assessments.
I just don’t think the eye test has any use at all really. There isn’t an articulable thing that is best measured and assessed live and by the eye as opposed to via computer measurement and data processing. It’s simply folksy romanticism, jobs for the boys, and a cultural tug-of-war that fuels the concept of eye-test scouting.
Consider even how sloppy the concept of ‘command’ is- loosely or variously defined criteria often become a bit of a catch all. A lot of these assessments can turn into narrative- ‘does the guy trust his changeup against a good hitter in a big spot-‘ that isn’t anchored or contextualized sufficiently in order to permit accurate assessment.
In the end, these eye test approaches end up only marginally different than the ‘Yu Darvish sucks because he couldn’t handle the pressure in the World Series’ scouting reports from the general public- it’s just that when a scout makes that same sort of snap judgement off of a SP not using his changeup in a high leverage situation during a Fall League game, the reading audience tends to take it a bit more seriously (not otherwies knowing the player) and come away thinking, ‘ok makes sense, young pitcher but needs to trust his stuff more.’ These eye test narratives are largely intuitive and provide some color to the story, but aren’t really serious minded.
Thanks for the good points and civil disagreement, I agree with what you say- but I do wonder, does anyone remember a ‘big insight’ from these eye test lists like we’ve had so many from Carson and Chris? Just looking at the top 10 in the rookie of the year vote last year, the eye test at fangraphs didn’t have useful any insight that differed in anyway from other online sources on any of them. If anything, it is as you’ve said and I concede- the lists do a decent job of simply summarizing what is out there in the general conversation about the prospects.
Consider, for example, the write up on Dejong last year.
Scouting Report
DeJong spent time at third base and, less frequently, shortstop in 2016. Based on what I saw in the Fall League (40 range at short, below-average arm), I have him projected to second base, though I suppose he could be passable at third, as well. He has plus bat speed and makes hard, all-fields contact, though his average raw power really only plays to his pull side in games. I don’t think his 22-homer season at Double-A was an aberration as much as I think it was an exaggeration. He puts the ball in the air pretty regularly, and the bat speed and bat path for some game power are certainly present. He doesn’t track pitches well and can get swing-happy, so the overall hit/power combo is right around fringe average because there’s a good bit of swing and miss going on here, as well. That’s probably a little short of profiling as a regular at second or third and, unless you’re buying heavily into the “we can hide bad defenders because of batted-ball data” theory and think he fits at short, he’s either a low-end everyday guy or very solid bench bat. I’m inclined to put a firm 40 on bench players who can’t play a premium position, but I buy into DeJong’s bat enough to bump the grade up a bit.
What is in there of substance. His HR total was an exaggeration not an aberation? Well no- he hit 38 HRs between AAA and MLB last year. Eye testing the glove at Fall League was worth no insight at all- he graded out as a plus defender at SS. All the assessments are lazy data: he puts the ball in the air regularly (more useful to know exactly how often), his power plays better to pull side (how much better?) etc.
Conclusion- overall hit/power combo is right around fringe average…inclined to put a firm 40 …but bump the grade up {to 45].
Like many of these eye tests, all that really means is ‘player doesn’t have obvious plus-plus physical tools, so scouts aren’t enamored.’
Does anyone have a good suggestion of a player scouted by Kiley and/or Eric who: (1) Kiley/Eric rated much differently than BA; (2) because of something they saw with their eye test; or (3) because of some other insightful reason; that (4) proved insightful and accurate? I’ve not yet seen one.
I’m a White Sox fan, so I’ll read the detailed descriptions of the ChiSox prospects when Eric & Kiley write one. I don’t really read the descriptions of any other team’s prospects, unless it’s a guy with a lot of hype that’s shooting up the Top-100. I’m really just looking for the FV number.
Maybe Eric & Kiley are writing a description to fit their number rather than fitting an FV to their qualitative report. They did write a “Making of the Top-100” piece that might answer this question.
I don’t see it as a big problem either way. If they have a sense that Player A projects to be a 45 and Player B a 40 and their qualitative reports on each player aren’t fine enough, it doesn’t matter to me because I’m not looking for that granularity. As I wrote above, their track record is much better than KATOH’s for evaluating prospects in trades. Are they better than BA or KLaw or BP or MLB? Again, I try to take an average of them all, so it doesn’t matter to me.
I don’t think Kiley or Eric are poor at the eye test or traditional scouting.
I just agree with the executive quoted in Kiley’s article:
““There is a big population [of scouts] working in this game who are lucky to have jobs. They don’t know anything. They’re professional BS artists,” one executive said.”
IMO there really isn’t anything articulable that the eye test can catch that is not better automated and assessed using data.
That’s really the heart of the Hosmer matter- the ‘eye test’ sees leadership, grit, heart, and effort, throw in some power, RBIS and a few visually impressive plays at 1B- and the eye test is sold.
The eye test doesn’t even attempt comprehensive coverage because it is not possible. Groups like profootball focus, opta, etc., attempt to get a set of eyes on all matches, but a modern computer vision data scraper is far more effective.
The eye test is left with a bull**** claim, ‘i know a player when I see one,’ and other than that it’s just a take-your-best-guess on tools grades. Beyond that, there is a lot of narrative with scouts arguing ‘you really have to see the player in warmups to know what he’s all about,’ or ‘I knew he had it in him when I saw how determined he looked after losing in the final,’ and other such arm chair psychology. “Make up” assessments are typically trojan horses for prejudice or worse..
Kiley and Eric are great, the scope of the coverage they provide is excellent- but give me KATOH 100 times out of 100 if actually pressed to assess prospects in a measured way where accuracy matters more than story.
It’s just a lousy remit, they have the ‘be the republican on MSNBC’ role- don’t get me wrong, I understand that this series is popular, but I don’t think the assessments are high quality (but rather a pivot towards broader appeal).
KATOH wasn’t actually more accurate though, and every major league team, no matter how progressive, still employs a massive scouting department.
KATOH rankings were far more accurate (40%+) one my brief look at the top 10 ROY vote getters- especially when considering how vague the ‘FV’ is with respect to when this value materializes.
Are you saying the traditional scouting was more accurate, if so by what measure. Or are people taking the easy cop-out “they are equally important, and best practice is to take equal portions of both?”
Sure the easy cop out is popular in and outside the industry, but the reality is that data science laps the eye test as medical doctors do witch doctors.
There is simply a body of workers throughout sport, mostly former players at some level, that won’t admit that they are essentially bull**** artists who aren’t able to innovate fact-driven and accurate assessments, so they conform to what is conventional for as long as they can. It happened in medicine, finance, and every other industry- it is happening too to sport.
So sounds like you think scouting generally has no value even those employed at baseball teams not just the ex-scouts like Kiley, KLaw, etc at publications? Perhaps scouts become obsolete in the future when all TrackMan data (xWOBA, etc.) is available to KATOH & Zips and they figure out its predictive ability in their algorithms.
I think what’s more likely is the emphasis of scouting will change to providing context for why KATOH & Zips still make so many errors in their player forecasts since they’re only providing the mean or most likely forecast in that player’s probability distribution. I mean KATOH forecasts Wes Rogers as #28 on its 2018 Top-100 (he’s one of many laughable head-scratchers on its list) and this was a guy who was passed over by everyone in the Rule 5 draft. KATOH doesn’t currently map to how front offices regard prospects. I certainly don’t want my White Sox trading KATOH’s #79 Eloy Jimenez for KATOH’s #6 Zack Granite.
Jeff Sullivan writes articles semi-regularly about a batter changing his stance or a pitcher changing his release point and provides photos so it’s obvious to someone like me. Do you think these articles of Jeff’s are bogus? If not, then why would scouting, which is the work of Jeff’s occasional article every day, be bogus? Are you demanding that level of technical detail be provided for every prospect? Can’t we, as readers, assume that Kiley & Eric have done that work (more so on a team’s top 10 prospects) and are simply providing us the high-level summary?
A changed stance or changed release point is best measured via data- which is easy. Release points are today mostly assessed via data.
There are no other high grossing industries that put any serious amount of trust in eye testing, not for financial markets, architecture, infrastructure, manufacturing, medicine, travel, logistics, defense, energy, IT- you name it. Sport is the one high grossing industry besides entertainment where most of the people with professional capital are uneducated- so when they get old, they coach, present on TV and scout. Therefore these three jobs have some of the lowest competency standards in the world among all high grossing industries.
There is no particular reason to think that the eye test adds more than it detracts.
I asked commentators- which players have been flagged up by the eye test more accurately that data modeling? No one gave an actual answer, and the answers offered (oddly names like Lindor, Sano) were way off the mark.
Similarly, I’ve asked ‘what articulable thing can the eye test measure that data cannot,’ and you’ve answered ‘changed release point,’ or ‘changed bat swing:’ again both wrong like the Sano & Lindor examples, as both of those are best measured by data (which is the current MLB industry standard for assessing both of those factors). The low-information eye testers can sometimes notice those things too, but just with far less accuracy, less information, less precision and way more bias and conformity.
There are clubs that only employ scouts because the front office doesn’t want a war of culture on their hands- scouts are considered ‘baseball men’ whereas an ivy league data scientist is not. Coaches and players tend to treat ‘baseball men’ differently than non baseball men. There are clubs that would go away from eye testing entirely and pivot to computer vision – just as their are league executives that want to go away from having umpires eye-test balls and strikes and instead go completely to computer vision. The backlash in either case would be substantial, both inside clubs and with the less educated or more traditional fans.
But to answer your question, no Kieth Law’s eye test doesn’t see anything that isn’t better measured with computer vision and data processing. Eye testing is obsolete now save only for it’s folksy appeal to traditionalist. Sure- scouts can just learn to use data, and since games happen, baseball professionals will be watching them, and so making observations with one’s eyes will always be part of the sport- just like it is currently for MLB’s data analysts, who also have eyes and watch games. But the sophistication level of pro sport eye testers is extremely low, the eyes and mind simply cannot come close to processing most of the relevant information during any given play, let alone game.
Thom, after reading through as many of your posts as I could at this time, it seems to me that you are missing 3 things that I think scouting provides more than a stats-based tool.
1) You’re citing extremely small samples (for example, your Acuna stuff below is including an 8 PA for a 170 wRC+ showing in Rookie-Ball or a 49 PA sample for Tatis for his .182 ISO). Even for guys like Hoskins, his 212 PA sample is not overly large. Similarly for other players you or others have brought up in your examples of trying to prove yourself or disprove others.
2) Improvement: Players improve. They improve as they move up levels and get better coaching, better facilities, etc. I think this can be seen in the case of DeJong. He did not show the “tools” necessary to play SS, so Eric projected him to 3B, the position he had played his whole minor league career outside of 11 games at SS in 2016. You say DeJong grades out as a plus defender at SS? 1. We know that defensive samples are either not large enough or not reliable to definitively say much of anything after just one season, let alone less than that (which DeJong has in the majors). I guess you’re using UZR/150 or or DRS (which show him more as average at SS at +3 and 0, respectively), because Clay Davenport has DeJong as a below average SS both in the minors (-1) and the majors (-5). So, plus shortstop defense?
3) Projection: I think the eye test is useful in projection while stats based is more reflective. Scouting eye tests talk about things like putting on weight, getting stronger, growing, making adjustments, etc. Can these be quantified in stats? Yes. But only after the fact. Scouting and using your eyes can show you that a player with large feet, but who is short is likely to grow (Pedro Gonzalez or Oneil Cruz – can’t think of older guys since I’m not that old myself). Yeah, Eric and Kiley are going to miss on guys. But that is because they’re to project and predict, and yeah things happen to humans all the time that throws them off that track.
So yeah, these lists aren’t going to reveal some hidden gem. They’re not supposed to, as you have already stated in other comments. I’m not sure what you’re looking for in lists like these, maybe you just need KATOH and the Fringe Five back in your life?
None of those examples favor eye testing over data.
KATOH tests physiological data (including weight and projected weight) that correlates with success.
Putting on ‘weight’ is best measured in data: pounds. Getting functionally stronger is also again best measured in data- maybe by proxy such as bat speed or EV. Data far exceeds the eye test for processing such information, particularly the different between gaining strength and gaining functional strength.
There are zero professional sports teams that train based on the eye test- all use data. High intensity work load is not measured with the eye test. Fitness tracking is not measured by the eye test. All is done with sensors and data analysis, all across pro sport.
Similarly growing is measured in data. Height X at Time 0 – Height Y at Time 1= growth. The eye test is pretty poor at assessing growth, whether height or bulk or posture or elbow strength or whatever else. No serious physical assessments in sport are done based on the eye test- Dr. Andrews relies on modern data-driven medical surgery and rehab procedures, he doesn’t have a former SP who ‘knows pitching elbows’ conduct an eye test.
“Making adjustments” are again best measured and assessed with data. Consider how many nonsense eye test ‘adjustments’ have been falsely reported over the years: how many eye testers believe in things like clutch hitting, ‘confidence players,’ and lineup protection? The eye test sees a player bunt, and concludes ‘this player has made great adjustments to be a real team player, a true Yankee (or whatever).’
So to answer your points-
(1) Yes its a small sample size. One thing I appreciate about KATOH is that it is anchored on actual performance- even retroactively. So KATOH uses, let’s say, 2010-2016 to project 2017. Then 2017 happens, and KATOH goes back and ‘learns’ how to adjust co-efficients to incorporate the new information from 2017. If a variable stops being predictive, it gets left out of the model. This is how nearly all serious assessments are done in all high grossing industries- save only when an industry has a bunch of uneducated big shots that want jobs after their prime earning years (i.e. entertainment and sport, maybe some others like politics, education, consulting and law).
It would be a big improvement if Eric and Kiley anchored their scores to actual machine-learning regressions- but they won’t likely do that, because eye test traditional scouting is supposed to remain ‘pure’ of modern data computation. It’s literally an anti-science approach (though I’d expect Eric and Kiley are as data friendly as most any scouts). There is simply a ‘we have to do this because alot of coaches and scouts are uneducated’ tradition to rely on ‘baseball people’ to make ‘baseball assessments’ with their eyes- it’s why we have umpires: but this dynamic has no purpose apart from traditional appeal. Umpires and scouts would be more efficient automated, sad and disturbing but true.
(2) Sure players improve. Dejong may have improved after Eric scored him. But eye testing and data projections are both supposed to estimate improvement as accurately as possible. Eric’s writeup didn’t see much FV for Dejong even when accounting for forseeable improvement- whereas data analysis did. It’s easy to understand this- nearly all differences between scouting and data-driven assessments come down to simply one question: does the player have super obvious physical attributes? If so, then scouts will likely overrate the player, as groups in the aggregate have difficulty avoiding over emphasis on something obvious that everyone sees.
(3) Your argument is literally backwards. A player with a large frame is known to be likely to fill out and add strength. Every club employs sports scientists who measure player dimensions -using data rather than saying “oy, he looks a big’un” – and then use development models that predict added weight. Every club manages weight gain programs using sports scientist platforms like prozone or catapult. Scouts and eye testers, even fitness coaches, have a massive disadvantage trying to eye test physical development. That is why, at lower sporting levels, the scouts typically love the over-developed kids and underlook the smaller, weaker kids with high production. Data analysis is far more precise in predicting, within a given confidence interval, whether someone is likely to grow or otherwise change in any articulable way.
There is a reason all these answers (Sano, Lindor; weight gain, etc) are far off the mark. It’s because I asked a question that has a null set as an answer. What high-information events are better measured and processed by the eye test measure as opposed to modern information technology? None- but Arod and Matheny want jobs, so the industry has to kinda pretend that their ‘baseball eye’ adds value, and the cottage industry of scouting has popped up by catering to the sensitivities of the ex players.
You’re dumping on this series – no pay wall, by the way – because it didn’t predict 90th percentile outcomes for DeJong and Hoskins? Was anyone in the industry bullish on DeJong headed into 2017? If it can be done much better, sounds like you need to start a daily blog for me to complain on!
I’m critiquing a process that operates like this.
“watched a pitcher once in the fall league, heard from a few guys that his command is 40, looked worse to me so I gave him a 30.’
That’s a 100 year old approach that only exists because of regressive approaches.
Consider Miguel Aparicio above- with the 20 current hit grade. There is no actual sound basis for such a low grade- he his about .300 at rookie ball and low A as an 18 year old, keeping his K rate under 20. This is just an arbitrary grade. The write up doesn’t even match with the grading- the writeup says ‘if he maxes out his hit tool he can get to a soft 50,’ whereas the grade shows max 55. While Aparicio did run up a terrible 94 PA as an 18 in A ball, he doesn’t have anywhere near a ’20 grade’ hit tool profile- which generally should be 3 standard deviations worse than average (99.7%). There is nothing close to a serious argument that Aparicio’s hit tool is currently worse than 99.7% of minor leaguers, prospects, or anything like that – he has a .270 avg for his career with a K rate around 16%: with 40 XBH.
It’s simply not a serious approach. I didn’t critique the Dejong writeup for failing to predict a breakout, I critiqued it for making patently wrong observations based on (1) quick look at tools; (2) talking to a couple people and conforming to industry standard; and (3) abstaining from a more in-depth analytical assessment. To nit pick about Aparicio is maybe a bit much, the 27th ranked prospect- but the write up didn’t even match the grades, the insights are pretty much non-existent and based off of one watching plus asking a couple guys what they think: KATOH and other data driven models are far better, or merely looking at his fangraphs data page provides world’s more relevant and accurate information.
“Who is the last under the radar guy on these fangraphs eye-test lists to do far better than projections would expect?”
Off the top of my head, and over several years.
Acuna last year (was noted as a top 100 guy with a chance to be THE GUY based on nothing but scouting as he was hurt),
Honeywell (Kiley noted him as a Top 150 45+ FV straight after being drafted when he had little track record).
Fernando Tatis was another guy scouts were on last spring before the stats jumped to back them up.
If you’re looking for guys in the majors, there’s Lindor (who didn’t really start hitting until he made the Show), or Sano, or Mazara. I remember Realmuto getting tagged as a prospect sleeper by Kiley before he came up and became a first division regular as well.
All of these examples are wrong and poor.
Acuna absolutely mashed at all lower levels. You label his data profile as ‘nothing’ but over look that he RC+ of 145, 138, 170 & 139 before he was given a 55 FV last year. He was already very highly rated by the data- the scouts did not like Acuna first, rather the opposite.
Lindor wasn’t scouts only by any means- he was a consensus top talent, all american, #8 pick in the draft. Again, Lindor has always projected very well, the scouts didn’t have any head start.
Your Honeywell info is off the mark- Honeywell had posted a 1.07 ERA, 2.20 FIP, and a 10.7/0.7/0.3 slash in his rookie ball debut as the 4th pick from the 2nd round. It was by no means surprising or ahead of the curve to give him the modest grade of 45 FV. How is that considered an insight?
https://www.fangraphs.com/blogs/evaluating-the-prospects-tampa-bay-rays/
Similarly, Sano was a consensus scouting and data darling since his big money signing in intl FA: he posted RC+s of 191, 131, 153, 146, 203, 145, 156 in the minors. When was he overlooked by the data analysts? Did we really need the eye test to tell us- big power, doesn’t move great, prone to Ks. Of course not- the players you are listing are some extreme examples of consensus talents.
Tatis is different that the others, he has no success yet above A ball. Even so, he was highly valued by data analysts immediately, despite low walks and limited power last season during his debut. His power increased this year- but the data already showed that an 18 year old who posted ISOs of .153 and .183 during his debut would gain power. Power-relative-to-age is a major factor in projection systems, so there is nothing about Tatis’ power spike last year that was unforeseen by the data approach.
IMO no one is going to be able to come up with a fact-based answer because eye test scouting has zero competitive advantage on data analysis and hasn’t actually provided any insight. It’s just narrative.
Surely Sano and Lindor are very poor examples.
Kiley also identified Albies before his breakout.
No he didn’t. Notice that no one saying ‘scouts ID’s so and so’ are using any links. Kiley merely praised Albies before you noticed he was already a consensus great talent with top production.
https://www.fangraphs.com/statss.aspx?playerid=16556&position=2B
Albies rocked a 156 and 168 RC+ at two rookie ball stops as a 17 year old playing SS – at that point he was on the radar of every data-driven projection system. He showed great bat control immediately (10% K rate in rookie ball) as well as very good speed.
The weakness of these answers demonstrates that – no one actually has a good example on the top of their head as far as a FG eye test find that proved insightful. There isn’t a Mookie Betts or Hangier for the eye test lists, while we can all name plenty of big time players that the FGs data team found that were being overlooked by the eye testers and general industry.
Albies was a data star immediately & never slowed down in his development, and every prospect projection system I know of expects 17 year olds to gain power in a nearly linear fashion that is similar to how Albies increased power.
What in the world do we even imagine the eye testers can see that gives insight on a player like Albies? What articulable observation point? Loose wrists, confidence, leadership, heart, poise, desire to win, gamesmanship, being a real gamer, being a team player? Most all of these are either dubious or best measured with data.
Bat speed and swing plane are best measured with data (which is the current industry standard for assessing such things). So too with speed. Defensive ability is still mixed simply because of cost: not all lower level clubs have installed modern information technology, but as soon as they pay up they will switch away from eye test and towards stat-cast type assessments of fielding ability. Controlling the strike zone is best measured by data, consistency is best measured by data, fitness, power, EV, hard contact, pull/opposite field, plate discipline, bat-to-ball ability, etc.
What thing does something think that an eye test could notice about Albies that data cannot more accurately measure?
Couldnt disagree more. KATOH was almost entirely useless IMO. KATOH’s hits were no more numerous than its, at times, ludicrous misses. Scouting the stat line is a terrible idea in identifying MLB talent. A thoroughly researched report of industry consensus and direct scouting is incredibly valuable. It not only allows insight into how organization view their own talent, but also into trade valuation. Traditional scouting does a much better job of identifying talent than mathematical approaches to date.
Thom, these lists are probably the most detailed scouting reports available to us. I was a huge fan of KATOH, but I think it’s pretty well established that KATOH is a complement to existing scouting and shouldn’t replace it (and Chris said as much as well). Moreover, Eric and Kiley have written up a lot about what they look for and how they evaluate, and the detail they provide is much more transparent than what you see elsewhere.
And yes, the miss rates are huge, but that’s true with all prospect analysis (including statistical approaches). The reason for that is generally speaking, your favorite prospect (whoever it is) is going to bust. If you wanted to be on the safe side, you would just project every prospect to bust and you’d have a really good hit rate. But that sort of defeats the whole purpose of it.
The only quibble I have is that we forget how high the bust rate is when looking at FV rankings, but even there Eric is much more conservative with FV rankings than other outlets.
BTW, it’s really unfair to go after anyone for missing on Aaron Judge. First of all, it is the right call to be down on players with contact issues (which is exactly what KATOH says you should be worried about). Second, absolutely no one called it…no projection system, no scouts, nobody…people were debating whether he would stick in the league as a starter, not whether he was going to be an MVP candidate. And third, IIRC Eric was actually higher on Aaron Judge than a lot of other outlets.
I agree that these lists are the most comprehensive lists we have- but that is partly an editorial choice at fangraphs.
Obviously, the majority of the sporting audience/ESPN audience is regressive, anti-analytics and laughs at WAR. To the contrary, the majority of fangraphs readership is generally pro analytics. As for Chris saying that data analysis complements traditional scouting- that is merely what analysts are expected to say, because to say otherwise is unpopular.
The write ups by Kiley and Eric do not articulate a single thing that is best measured by the eye test as opposed to comprehensively and accurately assessed by computer vision. Alot of what they’ve written by assessments is accurately reflected in what I stated – they look at tools, look at production, factor in industry opinion and recent narratives, ask themselves about trade value- and then pick a number. There isn’t an articulable, high-intelligence, unique-to-pro-scouts “thing” to see- which is why there really isn’t anything to say as far as what scouts can see that data cannot capture. The most typical answer is ‘game environment, leadership, how a player warms up, how a player handles pressure, how a player is as a teammate, how much effort the player uses, etc.’ – all things that either can be measured by data or are merely armchair psychology/narrative.
As for the idea mentioned above that the ‘traditional scouting lists’ are more in line with trade value than KATOH- well of course that would generally be true, if the ‘traditional lists’ are meant to represent industry consensus whereas KATOH is intended to measure performance. For some smart FOs like the Pirates, however, KATOH tends to explain the transactions better.
Lastly, I am certainly not going after any talent assessors for missing this or that player. I am more generally asking- when has the eye test on fangraphs ever provided useful insight, moreso than the data analysis? I looked at last years top 10 ROY vote getters to see if any of them had been singled out with useful insight by the fangraphs eye test: none were, but rather, as is typical, KATOH and data assessments were more accurate. This makes sense, KATOH is anchored on actual performance and facts, it shows which players are doing the things that factually correlate most with success. Traditional scouting lists are often aggregators of lowest common denominator eye tests.
What player’s performance was better predicted by the traditional FG scouting lists than the metrics? There are of course answers, but I don’t think anyone can come up with a good answer because eye test scouting is conformist, low information and regressive by nature- the obligatory “both approaches are exactly equal” is just bending over backwards to not embarrass the traditionalists.
Thom, you’re not going to like the answer to this. Chris actually did include Top 100 rankings in his model at one point…he called it KATOH+. The model fit was way better, and the scouting lists were generally much faster to catch onto changes than the stats were when he updated them every few months.
Why do the Baseball America-type analyses do so well? Partly it is because it is become much more data-savvy…the advantages of statistical modeling is bleeding over into the scouting world. This is how it should be. You should always be funneling insights from scouting into your quantitative models, and you should always be incorporating quantitative results into your scouting. This type of mixed-methods analysis is the best way to make rapid progress and stay ahead of the curve.
Additionally, while I think realistically some things (like speed and arm tools) will be completely taken over by computers, it’s really hard to capture developmental changes (in fielding, hitting and power) with quantitative modeling. Growth curve models are a great invention but they are better for capturing between-player differences than within-player change. Idiographic quantitative methods are still in their infancy, and the signal is utterly swamped by noise. And the error in quantitative measurement for a lot of the things that are relevant for development (such as strike zone awareness, bat to ball skills, loose wrists, field agility) is so big that even if you could model it correctly we couldn’t get accurate enough measurements to make it work.
Finally, I think you’re getting too wrapped up in the idea that scouting is unscientific. I think at one point it was unscientific, but scouts have had to up their game lately. There are dinosaurs who think Eric Hosmer is the bomb and they’ll be out of a job, but rigorous qualitative data collection and analysis can be very scientific. Given all that Kiley and Eric have written up about what they look for, it seems systematic. It is good to be skeptical of certain scouting reports when they don’t line up with the statistics we’d expect but the whole enterprise is probably effective than scouting the stat line because it is so comprehensive, if not anywhere near as efficient.
I know about KATOH+, it was less predictive than KATOH.
The only reason data analysts have to pay lip service to traditional scouting is that coaches, traditionalists, and players are highly offended by data analysis.
Your argument is purely prose. To say modeling should “always incorporate eye test information” is incorrect. Let’s talk about sports scientists- which are used at every club. They run regression models to estimate things like ‘load,’ and they have KPIs that suggest not to go over certain amounts of training load because of correlation with injury. Now let’s say a fitness coach comes in and says, ‘nah, dont’ worry about Ole’ Hoss, he is a machine, throw away that stupid spread sheet.’
What should a professional do? Should they tweak their best practice, medical-industry-standard assessment to “funnel insights’ from the eye tester? Of course not- they often do because of club politics, but those who are caught can lose their license.
Consider Brett Hull. The data says he shouldn’t play, as he has a concussion. The eye testers say, ‘dag gum data, in my day we didn’t even wear helmets, he’s a man and he’s going to play.’ Seriously, that is the level of sophistication we are dealing with in the eye tester world. We rightly disapprove of all medical staffs that disregard modern, information technology driven best-practice assessments on the basis of ‘well, he looks alright to me’ eye testing.
Consider the difference between eye testing player health in the NFL compared to data -testing. There is literally nothing that the eye test will pick up that a data assessment will not- computers can voice record a patient’s spoken systems, process the information, and analyze it- automated to a tableau output plugged into a machine learning program at a local hospital/university.
It sounds balanced to say “oh of course, scouting should be filtered into a multifaceted analysis,’ but it is just empty rhetoric. Can anyway say how the inputs should be weighted? So let’s say our KATOHlbs model predicts a 17 player to add 16 pounds of weight over the next 3 years, considering every articulable and predictive input variable, physiologically, performance-wise, nutritionally, work ethic, and demographically, taking into account their fitness program. And let’s say a strength coach really likes the kid, and says “no, I believe in him, he’s a two sport athlete, he’ll add 35 lbs.” What weight should be given to each output? What if the data analyst knows that being a two sport athlete actually has a strongly negative correlation with weight gain?
The easy and popular answer is “something like 50-50,” simply because this answer comes off as friendly, inclusive, open-minded, and relaxed. But it is not a serious answer. If someone were to say, ‘recent studies suggest a 84-16% split in favor of the data,’ I’d concede that’s possible, but I’d expect that it’s simply a function of some clubs not yet being caught up and purchasing modern systems (or firing the Dave Stewarts, etc).
I get it that it’s popular to simply assume that eye test inputs have predictive value, but the data indicates otherwise. The unserious approach to weighting eye test input data (i.e. “ah, just split the difference”) indicates that there isn’t a known value of eye test observations as a predictive indicator. If there was, we’d cite the papers and say, ‘well the data shows that the eye test input variables improves the data-only model by an adjusted R-squared of 0.05,” but instead no one has any serious minded proof- IMO that is because there isn’t one.
Thom, I hate to say it, but you’re wrong on everything here. KATOH+ had better model fit than KATOH. And moreover, you need qualitative data to figure out what to put into the model, and qualitative data to interpret what the model means. And you’re talking about measurements taken in a clinical setting as though they can be done just as easily in a non-clinical setting, and the examples you’re citing are way less complicated than measuring something crazy like “hit tool 3 years from now”. And, on top of all that, you’re trying to essentially do a partition of variance of how “useful” qualitative evidence is vs. quantitative measurements when you should be thinking of it as an interaction between the two.
Just like life in general, one never knows what’s going to happen. One can use both eyeballs, crunch any number of stats, and win all sorts of arguments with pundits and friends, and still be wrong about what a player may end up doing. Teams miss all the time on minor league talent (draft, trade, or otherwise), so why would two guys at FG be any more accurate? They are using the best of what they have got to use to make predictions. No one has a crystal ball, so no one should be treating these lists as destiny.
I imagine that if there are a majority of successful predictions in the long run on a list, then the list is a success. Check back in five years to see if these Rangers make anything of themselves, as one year of output (Hoskins, Judge, etc) may be too early to say regarding long term success. In the end, it’s easy to harp on the misses, recent or otherwise, but misses are inevitable in any line of work or judgement.
Sure, it’s tautological that no one knows the future.
But this statement
“They are using the best of what they have got to use to make predictions.”
is likely wrong.
You could make the same arguments about witch doctors, i.e. “who can fathom the mysteries of the universe, etc etc, everyone is just trying to make the best assessments they can, no one is perfect,” but data-driven medical science is no-question more accurate that medical eye tests. Same applies here.
I think you may have written more about how useless these lists are (it definately SEEMS like it), then the actual list.
You’re like that one beat writer in every city, that can’t let it go. I’m in Chicago, so say a Sullivan or Wittenmeyer or going back, Bernie Lincicum. May have made a good point, but can’t let it go for a season or 10 and the way it’s said is to provoke (or, using a more current term, clickbaity).
So you think these guys are being lazy somewhere and not using all available resources for their write ups? No way. Given their experience in these matters, I think they do and deserve the benefit of the doubt.
As for analogies, I’ll go with one about ‘perfect’ predictions. If one hopes to drive from point A to point B, does one not take into account the quickest route, typical traffic patterns, and weather conditions before ones goes? It seems logical that, with all this information pre-considered, that everything will work out great on the trip. However, as you know, sometimes it doesn’t work out that way because one can’t predict things like accidents, slow drivers or detours. One does the best one can, however, based on one’s experience and expertise. If it doesn’t work out perfectly as predicted, one shouldn’t be surprised. It’s a prediction, not a fact.
They are not being lazy, but they are tipping the hat to the old school method of heavily weighting what they see in one game, mixed with a look at the data and a talk with scouts.
So Tyler Phillips (#23) above. Eric saw him once- command looked shaky. Very likely, Eric simply watched and did not exhaustively hand-score the placement of each pitch. Very likely, Eric watched as a normal scouting viewer- from the stands very likely, where it is extremely difficult to tell in which of the strike-zones 9 zones a pitch located. Eric didn’t have use of replays or slow motion. Given these difficulties with eye testing, Eric could only be expected to accurate assess each pitch type and location with moderate efficiency- the eyes and brain simply can’t observe, record and process a substantial amount of the relevant information.
So Eric saw, probably once, what looked like shaky command. Logistically, we are talking about a guy who isn’t a top 20 prospect. So how many scouts did Eric discuss Phillips with? Well, with 30 clubs having 20 prospects in their top 20s, we can assume that if Eric talks to 5 scouts for each top 20 prospect, that’s 3,000 conversations. How timely can such conversations take place? It would be unlikely that Eric spoke to more than 5 scouts about Phillips, probably one or two- and of course those scouts have networked and talked with others. It is very easy to be caught in an echo-chamber when catching up with a few scouts about any given player.
So Eric’s calculus for command was likely roughly as follows:
(1) Saw him and his command looked poor;
(2) spoke to some scouts who had him graded at 40
so….30 now, 45 FV.
It’s really that simple. I’m not saying Eric is lazy at all, turn him loose on a data system more accurately measuring all the relevant information and I’m sure his work would be high quality as usual. But send him to sit behind home and try to assess what each pitch was and where it located- is an iffy proposition, because eye test observation is no where near accurate enough compared to information tech.
As a Rangers fan, this is always my favorite day of the offseason. You all do amazing work on this site and just wanted to say thanks for everything you do.
Also, at some point last season, the Rangers changed how they allowed Minor League pitchers to pitch (fastballs only first time through the order, having to hit command check marks before promotions, etc). Have you heard any rumblings from other teams on how they feel this is working/will work for the club down the road? Texas has always struggled to develop pitching, so wondering if I should get my hopes up about this.
Drew Robinson? He was on your “Best of the 45/40 guys” list so I assume he was omitted here by mistake.
Never mind, he graduated through service time. Thanks for all the amazing work, this year’s prospect content has been some next level shit.
Look up yohander mendez career stats and then look at how many HR he gave up last year!
Would love to hear something on Starling Joseph – OF. Maybe next year