On How Much Wilmer Flores Needs to Hit to be a Good Shortstop Option for the Mets

@Mets 

With New York Mets shortstop Ruben Tejada struggling, and Wilmer Flores back to playing shortstop in the minors, I thought it would be a good time to consider whether Wilmer Flores is a reasonable internal option for the Mets. I have discussed this topic before, but it seemed an appropriate time to touch upon the subject again, as Ruben Tejada has looked even worse than last season – five intentional walks in 82 PA’s have boosted his OBP early, but he’s also striking out at a higher rate than ever, with an even lower rate of extra base hits – and Flores has been playing SS regularly for the Las Vegas51s (see these recaps: week 1, 2, 3). Considering it is unlikely that Flores would have improved enough in one off season to go from ‘never going to be a major league SS’ to ‘average major league SS,’ the question I wanted to answer is how much offense Wilmer Flores needs to provide to overcome his poor defense.

For the purpose of answering this question, a value-based statistic like fWAR allowed me to set certain defensive benchmarks for a given fWAR total and investigate the required offensive production. What I really wanted to answer was not whether he could be an upgrade over Ruben Tejada – as of now, almost anyone who is not a butcher in the field would be an upgrade over Tejada – but whether the Mets might have the solution to their SS problem in their own farm system. According to fangraphs, teams have gotten, on average, a little over 2 fWAR from the SS position over the past few seasons. While average production would still be an upgrade over what the Mets have seen since Jose Reyes left, they should be aiming for above-average production from every position, which is why I have chosen to investigate a 3 fWAR benchmark for Flores. Of course, they should be aiming for above-average defense at every position, and Flores is a long shot to ever provide that, but, before a few months ago, it was a long shot that we’d ever talk about Flores as a SS again.

The components of fWAR are determined by UZR for defensive value and wOBA for hitting value. There is also a base running measure that gets added to a players overall offensive value, but I have assumed that to be equal to zero, which is not an unreasonable projection for the average speed of Flores. UZR is not a perfect measure of a players defensive ability, especially over small sample sizes, but it is one of the best defensive metrics publicly available, and it is what fangraphs uses to calculate fWAR, so I am forced to use it. The graph below shows what wOBA Wilmer Flores would need to provide as a SS given certain negative UZR/150 scores. These calculations assume 150 games at SS with 600 PA’s.

Figure 1 – Graph of wOBA v. UZR/150 for a 3 fWAR Mets shortstop.

I did not go beyond -20 UZR/150 because 1) if Wilmer is this poor of a defender, he likely isn’t sticking at shortstop and 2) if he’s hitting this well, they’ll make room for him at 1B or 2B, though they may be discouraged from ever trying him at 2B if he’s that poor of a SS. (For the record, the -20 defender needs a .345 wOBA to reach 2 fWAR). Also, I didn’t go above 0 UZR/150 because that is not relevant to this discussion. If he is an average or better defender at shortstop, then the Mets should give him a long shot to see if he can hit in the majors anywhere near as good as he has showed in the minors. Since 2009, the league average wOBA has dropped from .329 to .314, and from .314 down to .298 for shortstops specifically. One thing to notice is that even an average defensive shortstop would have to hit better than league average to produce 3 fWAR. In between those extremes are three specific benchmarks I’d like to take a closer look at:

Table 1 – Offensive production needed for fWAR total given defensive score.

#Flores4SS	UZR/150	wOBA wRC+ (’13) ’13 Comp
2 fWAR	-5 UZR/150 -10 UZR/150 -15 UZR/150	.313 101 Span .324 108 Scutaro .334 116 Kinsler
3 fWAR	-5 UZR/150 -10 UZR/150 -15 UZR/150	.334 116 Zobrist .345 124 Lowrie .356 131 Pence

 One more comment on the 2 fWAR requirements before I move – they seem very reachable for a hitter of Flores reputation. Even if you don’t think Flores will be better than a Daniel Murphy-type at his best, Murphy at his best produced a .354 wOBA/126 wRC+ in 2011. That was 3 year’s ago, in Murph’s age-26 season, whereas Wilmer Flores won’t be 23 until August. If nothing else, Flores is making a case as a better possible option at 2B for the Mets moving forward, especially considering the relative costs of the two players.

Since two very different offensive profiles can produce the same wOBA, I wanted to tailor these wOBA’s towards the offensive profile of the player Wilmer Flores has been in the upper minors, which is when he ‘finally’ broke out. I also wanted to separate the rates that are more consistently predictable – like K%/BB%/HR%/XBH% – from the biggest variable in a players production – BABIP. With that in mind, I decided to look at a few possible combinations of those rates for Flores, and see what BABIP would be required to generate the necessary wOBA.

Table 2 – Wilmer’s historical rates and the wOBA produced in the majors.

#FloreSS	PA	K%	BB%	HR%	XBH/PA	ML wOBA
AA	275	10.9	7.3	2.9	10.2	.368
AAA	565	14.3	5.8	2.8	10.6	.359
Minors	3098	12.8	5.3	2.1	8	.324
Majors	105	23.8	4.8	1.0	7.4	.241

The 'ML wOBA' column is what those offensive rates would produce at the major league level in the 2013 offensive environment (aka, using fangraphs guts constants from 2013). AAA Las Vegas is known as a hitter’s park, so his numbers may be inflated some there, but his AA numbers were similarly strong. Unfortunately, he wasn’t able to impress in his sparse playing time after getting called up last season – he was limited by injuries – but that number of PA’s is a very small sample size. I have included his 2014 minor league numbers as of May 1 in Table 2. It may be over confident to do so, but I’m only going to consider 14% < K% <  20%, as anything over 20% seems too high considering his minor league career total. I will also consider the ranges: 5 %< BB% < 7%; 8% < XBH% < 10.5%; with HR% = 2.5% = 15 HR/600 PA. Given a few combinations of these rates, here are the BABIP’s that would be required to reach 3 fWAR, and what the production looks like in more familiar terms.

Table 3 – Various offensive outputs that produce 3 fWAR for a -15 UZR/150 SS.

K%	BB%	XBH%	BABIP Req	AVG	OBP	SLG	2B/600 PA
14	6	9	.342	.309	.350	.458	39
14	7	10.5	.328	.297	.346	.464	48
17.5	6	9.25	.357	.308	.349	.459	41
20	5	8	.383	.318	.352	.455	33
20	7	10	.359	.299	.348	.460	45

Table 4 – Various offensive outputs that produce 3 fWAR for a -10 UZR/150 SS.

K%	BB%	XBH%	BABIP Req	AVG	OBP	SLG	2B/600 PA
14	6	9	.326	.296	.338	.445	39
14	7	10.5	.312	.284	.334	.450	48
17.5	6	9.25	.340	.295	.337	.446	41
20	5	8	.366	.305	.340	.442	33
20	7	10	.342	.286	.336	.447	45

Table 5 – Various offensive outputs that produce 3 fWAR for a -5 UZR/150 SS.

K%	BB%	XBH%	BABIP Req	AVG	OBP	SLG	2B/600 PA
14	6	9	.310	.282	.326	.431	39
14	7	10.5	.296	.270	.321	.437	48
17.5	6	9.25	.324	.281	.324	.433	41
20	5	8	.348	.292	.328	.429	33
20	7	10	.324	.273	.323	.434	45

           While the required slash lines might not look too impressive for the -15 UZR/150 shortstop, the BABIP required to get those slash lines should be considered unlikely until Flores shows he will be a high BABIP hitter, and even then some of those BABIP are improbably high for a player to retain. The slash lines and required BABIP’s are much more reasonable for the -10 UZR/150 player, though he’d still either have to have an ~14% K-rate or ~.340 BABIP to reach 3 fWAR. Lastly, the slash lines for the -5 UZR/150 shortstop seem very attainable for Flores – only the combination of the worst K/BB/XBH rates requires what would be considered a high BABIP at .348.

            So, as long as Flores is better than a -10 UZR/150 shortstop, it is reasonable to believe that he could become a 3 fWAR/600 PA shortstop for the Mets, which would be above average. If this is the case, the Mets should give him a multi-season look at shortstop, allowing him time to adjust to the league and another full offseason to work on his speed, agility, footwork and positioning at short. Also, as long as Flores is better than a -20 UZR/150 shortstop, he should be capable of providing 2 fWAR/600 PA, which appears to be way more than what Tejada is capable of providing at this point. For reference, Asdrubal Cabrera was a -17 UZR/150 shortstop for the Indians last season, and they ran him out there for 122+ games worth of innings in 2013. Dating back to 2009, there are 12 instances of a shortstop playing 1,000+ innings in the field with a -10 UZR/150 or worse, with 2009 Yuniesky Betancourt the low man at -19.9 – other offenders include Derek Jeter, Jason Bartlett, Hanley Ramirez, and Rafael Furcal.

            Right now, however, Flores is dealing with a hand injury from Sunday’s game against Tacoma. On top of that, Flores has not gotten off to the offensive start many were looking for. He had an especially rough first week for the 51s, but has been driving the ball better lately (especially the other way), and is playing in the right league and Home Park to get out of a slump. Plus, he has shown improved patience early on this season, and he has had an excellent eye at the plate.

As a fan, it’s hard to watch Tejada keep getting at-bats, especially when the rest of the lineup is struggling to score. I’d ultimately like to see the Mets upgrade at shortstop, and I’m not too picky about what route they go to do so. In the short term, I think the Mets owe it to themselves to give Flores a major league chance at shortstop when his hand heals, as the best possible outcome for the Mets is that he proves capable there – it solves the shortstop problem, solves the problem of where to play Flores bat, and allows the Mets to hold onto all of their valuable assets (draft picks and other players/prospects). If he shows that he cannot handle shortstop, then the Mets know they have to go outside of the organization for a solution, and Flores should be no worse overall than what Tejada has been (-0.3 fWAR on May 1). Plus, with advanced scouting and fielding positioning, the Mets should be able to help Flores mask the problem of his limited range (his biggest knock). Hey, a fan can dream at least, especially when the team is tied for the 3rd best record on May 1!

Could Wilmer Flores be a 3 WAR Shortstop option for the Mets?

            The Mets shortstop issues have been well documented since the departure of star Jose Reyes. Things didn’t actually look so bad in 2012 when Ruben Tejada, Ronnie Cedeno and Omar Quintanilla combined for 2.8 fWAR – that isn’t Jose Reyes production, but it’s still solid production (average team received 2.4 fWAR from their shortstop in 2012), and good value for ~$3 million in combined salaries. Unfortunately, the Mets could not get repeat performances from Tejada or Quintanilla in 2013, who’s combined 586 PA’s produced a -0.7 fWAR. Somehow the Mets are starting 2014 with Ruben Tejada and Omar Quintanilla getting another chance as the majority shortstops, with Quintanilla yet to get an appearance. However, word came out during spring training that the Mets would be using Wilmer Flores at shortstop at AAA again. As you may have read, he was moved off shortstop three years ago due to a lack of lateral quickness – many scouts actually were impressed with his arm and upper body movements but said he would never have the feet to stay at shortstop. After fitness camp this past winter, Wilmer came into the spring camp looking leaner and perhaps a little quicker or more agile than in previous years – this is really yet to be determined, fans should wait to see what professional scouts say after seeing him over an extended period of time, assuming they give him a fair chance. This is relevant to the 2014 Mets because word also came out during spring training that the Mets might be willing to give Wilmer Flores a try at shortstop should Ruben Tejada play his way out of the starting role during April. Now, I’m not rooting against Tejada, I want all the Mets to succeed, I just won’t be confident he can return to the potential he showed even two years ago until I see it. I still think Tejada can be a useful bench piece, but am hopeful the Mets are looking for better long-term (and short-term) solutions.

Before they look outside of the organization for the shortstop solution, and considering the available outside options might cost more than what the front office considers appropriate, it makes sense the Mets would want to explore all internal options. Unfortunately, I haven’t seen any scout or blog come out and say that Wilmer looks good enough to handle shortstop in the majors, so the assumption (at least so far) has to be that he will be a negative defensive shortstop. (I’ve been keeping an eye on him at Vegas and will be tweeting about any plays in his vicinity. I would also like to put together a list of plays in his vicinity that he did or did not get to from week-to-week, so stay tuned for that.) So the question becomes, can he hit enough to be more valuable to the Mets than Tejada (as measured by fWAR)? Of more interest, really, is if he can hit enough to be a good, long-term option for the Mets.

A good option would provide slightly above average or better production – ideally with both the glove and the bat – and average shortstop production has ranged from 2-2.5 fWAR for the past dozen years, hence the 3 fWAR benchmark. Besting Tejada would require average, at best, value from the shortstop position, so I’ll also report on how much Flores would have to hit to reach that mark below. Since I’m not expecting even average defense from Flores just yet, the first objective is to determine how much offensive value he needs to offset a given negative defensive value. fWAR breaks offensive value down into a base running component (BSR) and a weighted runs above average (wRAA) component based on wOBA and a park factor. For this exercise, I will assume a BSR contribution of 0 runs for simplicity, perhaps a fair assumption anyway. The defensive component of fWAR is based on a players UZR score and a positional adjustment, which is +7.5 runs/162 games for shortstops. These two components (relative offensive and defensive value) and a replacements level adjustment of +17.5 runs/162 games are actually all that one needs to calculate fWAR. So, for the first part of this exercise, I needed to determine what wOBA Flores needed to provide given a certain negative UZR rating (really UZR/150). This was a simple calculation that even an Excel spreadsheet could handle, but I needed a little more information for the offensive component – the ‘guts’ of wOBA. Weighted on base average (wOBA) uses the run expectancy of events in the current season to determine the proper coefficients used in its formula. Additionally, wOBA has historically been scaled to be similar to OBP, and that scale varies year-to-year, though year-to-year variations in both the coefficients and the scale factor tend to be minor. Because the changes are generally minor, I am using the 2013 run environment for these calculations. As I discussed in my previous post on predicting key Mets player stats, the fWAR estimations were within +/- 0.3 fWAR in 2011-12 using the 2013 run environment, so, ignoring BSR, I should provide an accurate estimate of the necessary wOBA to make up for poor defense. To make things easier, I will assume 600 PA’s over 150 games (obviously more than he will reach). For the second part of this exercise, I will need to determine what reasonable offensive profiles, if any, Wilmer could provide to produce the necessary wOBA.

Figure 1 – Graph of wOBA v. UZR/150 for a 3 fWAR Mets shortstop.

I did not go beyond -20 UZR/150 because 1) if Wilmer is this poor of a defender, he likely isn’t sticking at shortstop and 2) if he’s hitting this well, they’ll make room for him at 1B or 2B, though they may be discouraged from ever trying him at 2B if he’s that poor at SS. (For the record, the -20 defender needs a .345 wOBA to reach 2 fWAR). Also, I didn’t go above 0 UZR/150 because that is not what this discussion is interested in. If he is an average or better defender at shortstop, then the Mets should give him a long shot to see if he can hit in the majors anywhere near as good as he has showed in the minors. Since 2009, the league average wOBA has dropped from .329 to .314, and from .314 down to .298 for shortstops specifically. One thing to notice is that even an average defensive shortstop would have to hit a little better than league average to produce 3 fWAR. In between those extremes are three specific benchmarks I’d like to take a closer look at:

Table 1 – Offensive production needed for fWAR total given defensive score.

#WilmerFlores4SS	UZR/150	wOBA wRC+ (’13) ’13 Comp
2 fWAR	-5 UZR/150 -10 UZR/150 -15 UZR/150	.313 101 Span .324 108 Scutaro .334 116 Kinsler
3 fWAR	-5 UZR/150 -10 UZR/150 -15 UZR/150	.334 116 Zobrist .345 124 Lowrie .356 131 Pence

 One more comment on the 2 fWAR requirements before I move – they seem very reachable for a hitter of Flores reputation. Even if you don’t think Flores will be better than Murphy at their best, Murphy at his best produced a .354 wOBA/126 wRC+ in 2011. That was 3 year’s ago, in Murph’s age-26 season, whereas Wilmer Flores won’t be 23 until August. If nothing else, Flores is likely going to be the better option at 2B for the Mets moving forward, especially considering the relative costs of the two players.

For the next part of this discussion, I wanted to redefine the wOBA using rates that are generally pretty consistent year-to-year (K/PA, BB/PA, XBH/PA and HR/PA) and a BABIP factor. First, I would like to set the HR rate at 15 HR/600 PA (2.5%) for all calculations – I would like to reduce the number of variables, I saw multiple projection models give him a HR rate of around 2.5% preseason and HR% has the largest wOBA coefficient, which means it will cause the biggest relative variations in wOBA. Second, the XBH% is going to be somewhat BABIP dependent, and I wasn’t sure how I wanted to handle that, so I have to consider them separate. However, the higher a player’s BABIP, the lower the XBH% would need to be to produce 3 fWAR, and vice-versa. Considering HR% is set, variations in BABIP are responsible for the biggest variations in wOBA, followed, in order, by BB%, XBH% and K% (note: XBH% is really 2B% now, as Flores is not a threat for triples without some help). Given his minor league rates in AA/AAA, I will estimate some possible BB%, K% and XBH% rates, and then calculate the minimum BABIP Flores would require given various combinations of said rates. The results will be various offensive profiles that Flores could provide to make up for certain levels of negative defense, finally! :) Before I get to the results, let me recap Wilmer’s minor and major league totals the last year and a half to give an idea of where my estimates are coming from.

Table 2 – Wilmer’s historical rates and the wOBA produced in the majors.

#Wilmer4SS	PA	K%	BB%	HR%	XBH/PA	ML wOBA
AA	275	10.9	7.3	2.9	10.2	.326
AAA	481	13.9	5.4	3.1	11.4	.341
Minors	3014	12.6	5.2	2.1	8	.325
Majors	105	23.8	4.8	1.0	7.4	.241

The ML wOBA column is what those offensive rates would produce at the major league level. AAA Las Vegas is known as a hitter’s park, so his numbers may be inflated some there, but his AA numbers are similarly excellent. Unfortunately, he wasn’t able to impress in his sparse playing time after getting called up last season – he was limited by injuries – but that number of PA’s is a very small sample size. It may be over confident to do so, but I’m only going to consider 14% < K% <  20%, as anything over 20% seems too high considering his minor league career total. I will also consider the ranges: 5 %< BB% < 7%; 8% < XBH% < 10.5%; with HR% = 2.5% = 15 HR/600 PA. Given a few combinations of these rates, here are the BABIP’s that would be required to reach 3 fWAR, and what the production looks like in more familiar terms.

Table 3 – Various offensive outputs that produce 3 fWAR for a -15 UZR/150 SS.

K%	BB%	XBH%	BABIP Req	AVG	OBP	SLG	2B/600 PA
14	6	9	.342	.309	.350	.458	39
14	7	10.5	.328	.297	.346	.464	48
17.5	6	9.25	.357	.308	.349	.459	41
20	5	8	.383	.318	.352	.455	33
20	7	10	.359	.299	.348	.460	45

Table 4 – Various offensive outputs that produce 3 fWAR for a -10 UZR/150 SS.

K%	BB%	XBH%	BABIP Req	AVG	OBP	SLG	2B/600 PA
14	6	9	.326	.296	.338	.445	39
14	7	10.5	.312	.284	.334	.450	48
17.5	6	9.25	.340	.295	.337	.446	41
20	5	8	.366	.305	.340	.442	33
20	7	10	.342	.286	.336	.447	45

Table 5 – Various offensive outputs that produce 3 fWAR for a -5 UZR/150 SS.

K%	BB%	XBH%	BABIP Req	AVG	OBP	SLG	2B/600 PA
14	6	9	.310	.282	.326	.431	39
14	7	10.5	.296	.270	.321	.437	48
17.5	6	9.25	.324	.281	.324	.433	41
20	5	8	.348	.292	.328	.429	33
20	7	10	.324	.273	.323	.434	45

            Finally, if you’ve read this far, I hope I wasn’t too wordy or confusing. If you’re just skipping ahead to see the results (which is fine), there they are, you can decide for yourself what is reasonable and what is not. You’ll notice that these are various ways to arrive at similar OPS totals, but the variations in K/BB/XBH rate describe noticeably different player profiles. While the required slash lines might not look too impressive for the -15 UZR/150 shortstop, the BABIP required to get those slash lines should be considered unlikely until Flores shows he will be a high BABIP hitter, and even then some of those BABIP are improbably high for a player to retain. The slash lines and required BABIP’s are much more reasonable for the -10 UZR/150 player, though he’d still either have to have an ~14% K-rate or ~.340 BABIP to reach 3 fWAR. Lastly, the slash lines for the -5 UZR/150 shortstop seem very attainable for Flores – only the combination of the worst K/BB/XBH rates requires what would be considered a high BABIP at .348.

            So, as long as Flores is better than a -10 UZR/150 shortstop, it is reasonable to believe that he could become a 3 fWAR/600 PA shortstop for the Mets, which would be above average. If this is the case, the Mets should give him a multi-season look at shortstop, allowing him time to adjust to the league and another offseason to work on his speed, agility, footwork and positioning at short. Also, as long as Flores is better than a -20 UZR/150 shortstop, he could be capable of providing 2 fWAR/600 PA, which appears to be more than what Tejada is capable of providing at this point. For reference, Asdrubal Cabrera was a -17 UZR/150 shortstop for the Indians last season, and they ran him out there for 122+ games worth of innings in 2013. Dating back to 2009, there are 12 instances of a shortstop playing 1,000+ innings in the field with a -10 UZR/150 or worse, with 2009 Betancourt the low man at -19.9 – other offenders include Jeter, Bartlett, Hanley Ramirez, and Furcal. Having watched Flores while he’s been playing shortstop the past week at Vegas, I will say that he looks at least as good as Jeter and Ramirez have the past few years, to my eye at least (and over a very small sample size). He has had 2 errors the past 2 games, but those happen, and he's shown some range and a strong arm as well. I have seen nothing to suggest he shouldn’t get a chance to take over for Tejada soon, and personally think he should be better than -10 UZR/150 (though I’m just an optimistic Mets fan with a poor camera angle on the games in Vegas). If that’s the case, the Mets might have just found a long-term shortstop solution in their own farm, and Flores is about to prove dozens of scouts, those who wrote his future as a shortstop off years ago, wrong.