Tuesday, July 31, 2012

The Quest For .500!

As of me writing this, the Mets are 50-53 on the curtails of a miserable 3-13 performance since the All-Star Break. They've scored 467 runs (putting them on pace for 734, or 16 runs more than in 2011)  and have allowed 476 runs (putting them on pace for 748, or 6 more than in 2011). They were 52-51 at this point last year, in the midst of a 5-game winning streak that started in game 102 (This year's Mets are in the midst of a current 2-game win streak that started in game 102).

Enough comparisons to last year. What I'm looking to do right now, is try and project where the Amazin's end up. Using Math!

The Mets have 59 games left. The winning percentage of their opponents in those remaining games is .491, which is slightly above the Mets' current percentage of .485. This is a little misleading, however.  28 of their remaining games come against teams over .500, who share a winning percentage of .568. The other 31 come against teams below .500, who share a winning percentage of .414.

Those are magic numbers: 31 and 28. A 31-28 record down the stretch yields an 81-81 record, which the Mets haven't achieved since 2008, when they went 89-73.

It can't be that simple though. Certainly, the Mets aren't going to win every game against the inferior competition, and lose every game against a winning team. So how likely *is* 31-28 from here?

Thanks to Baseball-Reference, I was able to get a breakdown of not only the Mets remaining schedule, but the number of remaining games against each, and how they've fared so far against each team. I broke all this down into this table:



Looking at this, we see that this (simple) model projects the Mets to go .... 30.638-28.362 from here! In a sense, it's fair and reasonable to say the Mets will finish very close to .500 this year. But that's also a rather simple way to look at it.

In my opinion, 31-28 represents neither the best nor worst case scenario, but a fair middle ground. Arguments can be made in either direction.


THREE REASONS WHY THE METS WILL FINISH ABOVE .500

(1)
The Phillies have traded away Shane Victorino and Hunter Pence.
The Marlins have traded away Gaby Sanchez, Omar Infante, Hanley Ramirez, and Anibal Sanchez (among others).

These two teams have arguably gotten worse since they've last played the Mets, and they represent 18 of the Mets remaining games. Based solely on their current head-to-head records, they project to go 12-6 in those games. Given the above changes, outpacing that projection isn't unreasonable.

(2)
The model above suggests that, because the Mets were swept in Houston, they won't win another game. I'm not a betting man, but if I were I would think it fair to say the Mets will take at least one of three.

Of the sub-.500 teams the Mets play going forward, only one (Miami) is within eight games of .500. After that, it's Milwaukee (-10), Philadelphia (-12), San Diego (-16), Colorado (-26), and Houston (-34). All of these teams have significant issues, and represent a prime opportunity for the Mets to fatten their record.

They also have left room for improvement against some opponents. The Mets have gone 4-17 against Cincinnati and Washington this year, and with 9 games remaining can look to improve upon that. If they even go a paltry 3-6 in those games, it represents a net gain against the projection.

(3)
The opportunity for internal improvement. It's no secret that the bullpen has been, well, terrible. But there are some bright spots - Josh Edgin has pitched great, giving up only a single earned run since his ML debut, and sporting a gaudy 16.2 K/9. Jon Rauch has quietly been solid, with a 1.32 ERA in his last 18 appearances (2ER/13.2IP). He struggled between mid-May and early June, but he has a 2.17 ERA otherwise to go along with his career best 1.062 WHIP.

On offense, Ike Davis' consistent resurgence has turned first base into a position of strength. His 20 home runs lead all players from the position, and his ability to hit somewhat competently against lefties has allowed additional flexibility for Terry Collins. Similarly, the improvement of players like Andres Torres and the out-of-nowhere-mashing of Scott Hairston both bode well for this team.

THREE REASONS WHY THE METS WON'T FINISH ABOVE .500


(1)
Their pitching. The bullpen has struggled. And struggled. And struggled. They could improve as the season wears on, but it's unlikely that they will make significant strides. The starting pitching has also been shaky lately, and plenty of uncertainty surrounds the vacancies left by Dillon Gee and Johan Santana. Can the Mets win with Hefner filling in? Is Matt Harvey as good as he looked? The jury is out right now, but these are questions that need answers soon, before it's too late.

(2)
The rest of the offense. The Mets are getting almost nothing from their outfield, outside Scott Hairston. Jason Bay is still lost, Andres Torres' OBP is high, but his .320 SLG undoes the bulk of that. Mike Baxter's return offers a glimmer of hope and solidifies a potential platoon in left field, but that leaves the Mets with, at most, 1.5 effective outfielders at any moment. The infield (Davis/Murphy/Wright/Tejada) has been collectively solid of late, but it's going to continue to hurt the Mets when 5-9 in the batting order are dragging.

(3)
Emotions. The Mets' last homestand saw the wheels fall off the wagon, and then the wagon combusted, and then the ashes threw a hanging curveball to Matt Kemp. The Mets have been, by most admissions, playing better than their talent would suggest. When their long-touted locker room chemistry began to lull, things just got worse. Pedro Beato's demotion is a step in the right direction, but after three terrible weeks it leaves one to wonder if they can regain the feel-good vibe from the spring.




The Mets had a .544 winning percentage in April, May, and June. They're .292 going into their last game in July. The level of competition going forward suggests that they could go either way. If they play the way they have so far this season, .500 is attainable. If they catch a few breaks, they could end up better than that. 

Thursday, July 12, 2012

A look at Wins Above Replacement: Ruben Tejada

When Ruben Tejada has been healthy, he's been a very good shortstop. So good, in fact, that Mets fans have begun to forget about the guy who used to play shortstop for them. See? Forgotten.

The old guy is currently hitting .264/.336/.378, with 22 RBI and 27 XBH in 384 PA's. 
The new guy is currently hitting .325/.381/.405 with 12 RBI and 13 XBH in 182 PA's. 

Though a smaller sample size, it's not very tough to see who's doing better at the plate. But let's make two further comparisons:

FIRST: How does Tejada at 22 compare to the old guy at 22?

Well, he won't play more games, because Jose (Who-se?) Reyes appeared in a career high 161 at age 22. In those games (733 PA's), he hit .273/.300/.386, with 58 RBI and 60(!) stolen bases. It is important to note here that many SABRmetricians discount the value of steals (an opinion I share. Argument: The Mets stole 130 bases last year, and scored 718 runs. They're on pace to steal 85 this year, and score 742, their highest total in five years). Anyway...

The people who spend their time collecting all these stats so people like me can whine about them have come up with a stat called Wins Above Replacement (WAR), which basically sums up the overall value of a player. The nice thing about it is that you can use it to compare seasons and careers of multiple players; its formulated to be consistent and comparable from year to year, and, to a limited extent, from position to position (If someone is a 5 WAR second baseman, it doesn't really hold that they would be a 5 WAR center fielder. Also, it isn't necessarily fair to say that a 4.0 WAR first baseman is a better or more valuable player than a 3.7 WAR shortstop, as first base is the far-easier defensive position). Also, it's important to note that WAR is a "counting" statistic; it doesn't fluctuate like batting average. It accumulates or subtracts based on a player's performance relative to "replacement-level". In other words, for a player's WAR to decrease, they would have to play worse than replacement-level.

There are two primary sources for WAR: Fangraphs and Baseball-Reference. Each site calculates it somewhat differently, so someone's WAR on Fangraphs (denoted fWAR) wouldn't compare directly to someone's WAR on Baseball-Reference (denoted rWAR). You can find more information on how they're calculated on their respective sites.

Jose Reyes' age 22 season earned him a 1.4 rWAR and a 2.3 fWAR. Ruben Tejada has a 1.6 rWAR and a 1.3 fWAR this year. According to Baseball-Reference, Tejada's play in 41 games alone contributes more value to the Mets' success this year than Jose Reyes' play in 161 games did in 2005! On Fangraphs it isn't yet, but Tejada is certainly on the pace to top Reyes' season. To provide a closer comparison, however, I'm going to modify these a little bit. I'm going to call this "Marginal WAR", or MfWAR and MrWAR. I'm going to do this by dividing their WAR by plate appearances, to calculate how valuable each player is on a direct proportion to playing time.

Jose Reyes:
MfWAR: 0.00314
MrWAR: 0.00191

Ruben Tejada:
MfWAR: 0.00714
MrWAR: 0.00879

Depending on which site you use, Ruben Tejada has been between 2.27 and 4.6 times more valuable than Jose Reyes on a per-plate-appearance level (NOTE: This does not mean that each Ruben Tejada plate appearance is so much more valuable than one from Reyes. I chose to use PA's as I feel they best represent a player's portion of offensive and defensive playing time.) Were Tejada to play at this level for an entire season, it would translate to roughly 5.5-6.0 rWAR (Baseball-Reference defines 6.0+ as "all-star"). Is this sufficient to say that Tejada is an all-star caliber short stop? Certainly not in itself, though it does provide a good context about his level of play. 

To consider his "all-star" eligibility, we should look at how he stacks up to other major league shortstops. Which brings us to...

SECOND: How does Tejada compare to other MLB Shortstops in 2012?

Ruben Tejada is currently the second-youngest everyday shortstop in the major leagues (He is older than Starlin Castro of the Cubs. He is also older than Atlanta's opening-day shortstop, Tyler Pastornicky, who last appeared as a pinch runner on May 30th and has been replaced by Andrelton Simmons.) Tejada presently ranks 14th in fWAR (tied with Jose Reyes and Alcides Escobar) and 9th in fWAR (tied with Zack Cosart) in the MLB. (Fun Fact: both sites rank him above Derek Jeter!). However, in both cases, the bulk of players above him have played significantly more. So, for fun's sake, let's break down the top 15 players on each scale by Marginal WAR:

(Click to enlarge)

[NOTE: National League shortstops are in bold, American League shortstops are italicized]

We can draw two quick and dirty conclusions from this chart:

1. The two sites do show inconsistency in their grading of shortstops. Troy Tulowitzki, for example, fails to even show up in the Baseball-Reference list.

2. Though his playing time has been limited, Ruben Tejada has the second-highest MbWAR and fifth-highest MfWAR among NL shortstops.

Certainly due to the uncertainties of baseball, and to his level of inexperience, it's unfair to use this data to say that Tejada is an all-star caliber shortstop. However, it does seem fair to say that Tejada, when on the field, has certainly played like one. Let's hope that it's a trend that continues, because he looks to be headed for a bright and prosperous future.