Matt Reynolds 2014 Season and the BABIP Effect

Attempting to project how Matt Reynolds 2014 numbers might look at the Major League level as a function of BABIP.

Image from of a Jay Horwitz tweet

            Despite being a second round draft pick in 2012, Matt Reynolds has seemingly come out of nowhere this year to become one of the more interesting prospects in the upper levels of the Mets system. A third baseman for most of his tenure in college, the Mets have used Reynolds almost exclusively at his original position of shortstop – he’s played 256/277 games at short in the minors, only moving over to 2B occasionally in 2014 for the superior defender Wilfredo Tovar at Bingo, and to keep Wilmer Flores fresh at shortstop when he went back to Las Vegas. Defensively, while his range might still be a little light for shortstop, everything else is good enough to stick there, which is something I didn’t believe to be true back in April. What changed my opinion? If you’ve seen my site before, you know that I watch every Mets minor league game available on MiLB.tv, and report on most of them (I will likely report on more of them in the offseason), so watching him for 58 games at shortstop with Las Vegas is what changed my opinion. While he might have the glove to play a middle infield position in the majors soon, it’s his bat that has put the spotlight on Reynolds this season, as he combined for a .343/.405/.454 slash line across the top two minor league levels in 2014.

            Looking at that slash line, the first things that come to mind are: 1) that’s a really high average, and 2) that’s a below average ISO. Now would be a good time to look at how those stats break down across the two levels.

Table 1 – Some Matt Reynolds 2014 stats by level

Team	PA	H	2B	3B	HR	SO	BB	SB	CS	BA	OBP	ISO	BABIP
AA	242	75	5	3	1	41	29	6	3	.355	.430	.067	.433
AAA	301	89	16	4	5	60	21	14	4	.333	.385	.146	.404
Total	543	164	21	7	6	101	50	20	7	.343	.405	.111	.417

Table 2 – Some advanced Matt Reynolds 2014 stats* by level

Team	P/PA	GB%	FB%	LD%	IF%	kL%	kSw%	nB%
AA	3.98	51.2	19.2	24.4	4.7	4.2	12.1	12.0
EL-Avg	3.75	44.9	30.9	15.5	6.7	4.6	14.5	9.1
AAA	3.75	43.3	28.4	22.3	3.7	5.0	15.1	8.3
PCL-Avg	3.77	44.3	28.0	18.7	6.6	4.7	15.3	9.5
Total	3.85	46.9	24.2	23.2	4.2	4.6	13.8	9.9

*Stats available from statcorner, which has 2 less PA (both SO) listed for Bingo.

nB% = (BB-iBB+HBP)/PA

            As a quick intro for those unfamiliar with the website statcorner, it is a great website that compiles unique data from major and minor league baseball’s gameday/play-by-play files. The only problems I have with the website are that the minors stats don’t always agree with the more accepted Fangraphs/Baseball-Reference database, and that the minor league gameday files used to create the database are not always perfect. Specifically, minor league gameday misses pitches and apparently plate appearances, and occasionally mislabels some balls in play – for example (not necessarily Matt Reynolds, forget the players involved), a recent mistake I noticed came on a single lined over the second baseman’s head and into RF that minor league gameday described as a groundball. Also, note that minor league gameday is not consistently available below AA, so while play-by-play files should still be relatively accurate, pitch data will be incomplete. Furthermore, no minor league gameday has pitch identification/velocity available yet, so neither will statcorner. One final thing on statcorner is it’s unique feature of giving you the league average statistics when your mouse hovers over a specific stat, which is where I got the league average stats listed above.

            Back to Reynolds, his combination of high BABIP and low ISO appear to be red flags for his offensive outbreak this season. With respect to his BABIP, he had managed a .290 BABIP with Savannah in 2012 and a .263 BABIP with St. Lucie in 2013, so this was a huge increase. Line drives are expected to fall in for hits at a much higher rate than any other type of batted ball, and stat corner shows he had a 15.7 LD% with Savannah and a 9.9 LD% with St. Lucie. Tim Rohan of the New York Times discussed this with Reynolds recently, and that increase in line drive is not a fluke, but the result of offseason adjustments made before 2014.

"Then, in December, his agent, Jason Wood, had him meet with Rick Strickland, a hitting instructor and part-time scout for the Mets. Strickland concluded that Reynolds was not moving his body in unison as he swung. Adjustments were made and the results were immediate… And the line drives are going to the opposite field again."

As for his ISO, he went from well below Eastern League average ISO (.131 per statcorner) with Binghamton, to nearly Pacific Coast League average ISO (.152) with Las Vegas. A bump in power output was to be expected as he made the jump to the offensively charged PCL, but this was more than just a bump in power output, this was a significant jump to nearly league average power output. Nearly average PCL power will come with skepticism about how it translates to the majors, but his XBH and HR rate jumped significantly in Las Vegas.

Table 3 – Some more advanced Matt Reynolds 2014 stats by level

Team	BIP/PA	(2B+3B)/BIP	HR/FB
AA	70.7%	4.7%	3.0%
AAA	69.1%	9.6%	8.2%
Total	69.8%	7.4%	6.4%

            His rate of extra base hits when he put the ball in play doubled and his homerun/fly ball rate nearly tripled with Las Vegas, which indicates he was driving the ball in the air more often, as shown in Table 2. Due to weather issues and MiLB.tv limitations at AA, I only saw 13 BMets games with Reynolds before his promotion, but I rarely saw him hit the ball beyond shallow outfield depth, and it seemed like he was focusing on his opposite field approach. With Las Vegas, he was driving the ball with consistency and using the entire field. If you don’t believe me, look at the spray charts, courtesy of www.mlbfarm.com.

Figure 1 – Matt Reynolds spray chart with Binghamton in 2014 – 37 RF, 34 CF, 21 LF

Figure 2 – Matt Reynolds spray chart with Las Vegas in 2014 – 37 RF, 53 CF, 33 LF

            That first spray chart is what I’d expect from a Wilfredo Tovar-type slap hitter, as he rarely hits the ball beyond shallow OF depth. The second chart is more promising for future fringe-average power output, which is all a shortstop needs at the highest level, because he’s driving the ball all over the field and deeper to the OF with more consistency. These spray charts show that all of Reynolds homeruns were to left field, although he was able to hit some balls to the warning track and wall in right and center field. Although his HR/FB rate is more acceptable with Las Vegas, double-digit homerun totals appear to be a stretch at the highest level.

            Digging a little deeper into the stats, you can see that Reynolds splits reversed after his promotion, but he showed an ability to hit both lefties and righties well.

Table 4 – Some Matt Reynolds 2014 splits by level

Team	Home	Road	vs. Left	vs. Right
AA	.265/.351/.325	.468/.528/.543	.319/.414/.361	.374/.439/.453
AAA	.369/.423/.531	.299/.347/.431	.393/.430/.560	.306/.364/.443
Total	-*	.368/.422/.476	.359/.422/.468	.335/.396/.447

*Didn’t seem to make sense to combine home stats from two stadiums

            The home/road splits in AAA are a little disconcerting, as Cashman Field is one of the best places in pro baseball to hit, but there is a huge BABIP discrepancy within the splits – .455 BABIP at home vs. .355 on the road. While a .355 BABIP is still higher than average, it’s not unrealistically high (like the .400+ BABIP he has for the season), especially for someone with Reynolds approach and good speed – he’s not a burner, but has better than average speed. So the road splits are perhaps a closer representation of how Reynolds bat would look with a more sustainable BABIP, and a .778 OPS is fine for shortstop.

            So the question becomes, how much should we expect Reynolds to hit at the major league level? At the beginning of August, Toby Hyde attempted to calculate his expected BABIP with a major league BABIP calculator and use that to normalize his stats throughout his minor league career. He concluded that Reynolds improvements in 2014 were mostly BABIP driven, although he concedes that Reynolds has improved as a hitter, and that his future is limited to that of an impact utility player. Hyde had to use the questionable batted ball data with the ‘simple’ xBABIP calculator from Hardball Times, stolen base data for the speed score, and choose major league stadiums that played similarly to the stadiums Reynolds called home in the minors. He describes the process as essentially adding/subtracting singles to Reynolds totals across the three levels.

Table 5 – Some advanced stats from Reynolds first two years in the minors

Team	BIP/PA	(2B+3B)/BIP	HR/FB
2012	73.2%	6.1%	8.1%
2013	72.3%	7.6%	4.3%

             Hyde considered XBH/PA, and said:

“His BABIP is higher because he’s hit more single, not more extra-base hits. His extra-base hit rate between AA and AAA is below that of a-ball."

            Comparing Table’s 3 and 5, you can see that his (2B+3B)/BIP for the year is nearly the same as it was in A-ball, but Reynolds put the ball in play less often in 2014. And while his BABIP was higher with Bingo because he was hitting more singles, his BABIP was higher with Vegas because he was hitting more doubles and triples, as well as more singles.

Considering the NYT piece and the difference in spray charts, it would appear that Reynolds started the year with a commitment to the opposite field, which led to a lot of singles. As the season progressed, it’s likely he learned to recognize those pitches that he could pull and drive with authority, and he started using the whole field more, which led to an increase in extra base hits. While that’s a nice (unconfirmed) narrative, it still doesn’t answer the question of what Matt Reynolds will bring to the majors.

The truth is, no matter how good of a predictive model one uses, unless a player hits a lot of homeruns, his offensive output is going to be dependent on BABIP. Because of this, I’d rather consider a range of potential outcomes given certain set rates and a range of BABIP’s – like I did with Wilmer Flores earlier this year. Specifically, given set K/BB/(2B+3B)/HR-rates, what will his slash line look like as a function of BABIP? Hyde considered what Reynolds would provide with a 20% K-rate in the majors, which is in line with his overall AAA rate. Reynolds did show considerable improvement in that department during August, when he only struck out 17.5% of the time over 137 PA, but he’ll be facing consistently better pitching in the majors, so 20% seems like a fair estimate. Reynolds showed he could take a walk at an above average rate with Bingo over the first half, but his AAA BB-rate of 7% is more in line with his pre-2014 rates, so I will use that rate. Clearly he’s not going to hit many homeruns in the majors, so a 1% homerun rate seems fair (perhaps too generous for Citi Field?). When I was doing this exercise with Flores, I could ignore triples, as Wilmer doesn’t have the speed to hit triples without considerable outfield help. Reynolds is faster and the type of player who hustles out of the box on every play, so he will end up with some triples – he’s had one more triple than homerun in 2/3 seasons, with 0 triples in his debut for Savannah. In the minors, Reynolds has hit one triple for every 3.8 doubles, and averaged a 7.5% (2B+3B)/BIP rate the past two seasons. Given the 72% BIP-rate that’s left, and a 7.5% (2B+3B)/BIP rate, Reynolds would hit 32 doubles plus triples over 600 PA’s (or 5.3% (2B+3B)/PA). A breakdown of 26 doubles and 6 triples is closest to his minor league 2B: 3B ratio. Given these rates, his slash line ranges from .243/.296/.342 with a .300 BABIP to .281/.332/.381 with a .350 BABIP – to save you the time, it’s at .262/.314/.362 with a .325 BABIP, and .301/.350/.401 with a .375 BABIP.

My breakdown credits Reynolds with more extra base potential but less on base potential than Hyde’s, but the result is a similar OPS upside. Given this set offensive profile in the 2014 offense environment (see the guts page) and a 0 UZR rating, Reynolds would need a .340 BABIP to be a ~2 fWAR/600 PA shortstop – this is without considering his ability on the basepaths. While he doesn’t have game changing speed, he is fast enough to steal some bases at a decent clip and could easily have a positive baserunning score. I also manipulated the BB/XBH totals of the Reynolds offensive profile to produce the .251/.339/.318 hitter Hyde’s analysis yields with a .317 BABIP. I then calculated his fWAR given a 0 UZR rating, and that slash line produces a ~1.5 fWAR/600 PA shortstop in the 2014 offense environment. That might sound crazy, but look at Ruben Tejada’s 0.9 fWAR in 377 PA despite his .226/.340/.286 slash line – Tejada has a 4.3 UZR rating at SS so far.

My conclusion may not differ much from Hyde’s, but how I got there was different, and I hope the path was still interesting. Assuming he's a net neutral defender and baserunner, the most likely outcome for Reynolds is still that of a utility infielder, as he’ll likely need a higher than average BABIP to hit enough to start, and no one should expect a higher than average BABIP until a player has established themselves as that type of hitter. Reynolds will only be 24 next year, so he still has room to improve, and he improved a lot over this past offseason, but this is a fair end-of-2014 assessment. As a last foot note, while I don’t normally like to talk about the ‘intangibles’ a player brings to the table, there is something to Reynolds game that makes me believe he can be more than the sum of his parts. Maybe that’s just the Mets fan in me hoping for more than is reasonable from a prospect, or maybe that’s me being blinded by watching his BABIP-driven success with Las Vegas this year, but either way, I’m not betting against Reynolds just yet.

Amazing visualization of first supernova explosions from @sciencemagazine. @TACC #hookem http://t.co/bLhOf8aKgP pic.twitter.com/7FMYO9orKv
— NSF Comp & Info (@NSF_CISE) September 8, 2014