by Mike Rogers

Update: I have included the 2007 numbers for Adam Zornes. Keep in mind that this was over just 67 plate appearances, so his numbers are much more uncertain than his 2008 stats which came across 264 PA’s.

In part one of my look at the Padres 2008 draft, I’ve discussed the stats and scouting reports of the first 5 college bats that the Padres selected – Allan Dykstra of Wake Forest, Logan Forsythe of Arkansas, James Darnell of South Carolina, Blake Tekotte of Miami (FL), and Sawyer Carroll of Kentucky. Here in part two, I will do the same for the next 5 bats that they took in the 2008 draft; Cole Figueroa of Florida, Adam Zornes of Rice, Beamer Weems of Baylor, Matt Clark of Louisiana State, and Derek Shunk of Villanova. So, lets get started…

Read the rest of this entry »

by Mike Rogers

Update: I found an error that occurred in my 2008 numbers across all of the conferences. As a result, the numbers have changed a bit, and I have fixed them, but haven’t updated the blurbs I wrote about each player.

Sorry this is delayed. I got sick last week that ate up my two days off, and, thus, consumed any time I was going to devote to writing this piece. Better late than never, I suppose.

I’m changing this just a little, and dropping the lone college bat that I’ve evaluated that did not sign with the Padres out of the 2008 draft, Jason Kipnis of Arizona State. Also, in my previous post, I outlined my methodology. I forgot to say that I also have an aggregate number which uses park-Adjusted wOBA, park-adjusted Isolated Power BB%, and K%.

Caveats: One, I’m not an expert. Far from it. I’m not presenting this as gospel and I haven’t even gone back far enough to see how predictive (if at all) or accurate these methods are. Two, this, at best, is based on lots of small samples, relatively speaking. I’m dealing with two years of data for each of these kids. However, that’s generally around 350-550 plate appearances total. So, not even quite 80% of a big leaguers season. Add in hundreds of different parks and varying levels of competition, and you’ve got a lot of uncertainty that comes with this sort of thing. Three, regarding that competition, I realized that I didn’t adjust for the level of competition faced. This was an oversight on my part and something I will go back and change at some point in the future. However, I haven’t done that yet, but I’m going to present this anyways. And lastly, there’s big virtual hat tip I want to give to Adam Foster, founder of Project Prospect. I am a regular in the forums over there and through constant email exchange with Adam, he’s been gracious enough to show me his “system” he’s using to help rank prospects. I’ve used parts of his system. Half of my “score” is based on his back-of-the-envelop math and the other half would be too, but I’ve adjusted the weightings in the formula. I’m also using his Speed Score in my study as well. This analysis (for the lack of a better word) isn’t as in-depth as the one Lincoln Hamilton unveiled over at PP a couple weeks ago. He’s been at this for 4-5 months longer than I have been, so I defer to his systems superiority, though from the conversations him and I have had, we tend to agree on most of these kids.

With that long, long intro, lets not stall any longer…

Read the rest of this entry »

by Mike Rogers

Over the past couple of months, I’ve been filling any down time I’ve had (within reason) to importing and quantitatively evaluating college hitters. After about 3-4 weeks of constant tweaking, deciding what works and what doesn’t, I finally started to settle in to a system that analyzed what I felt were the key points to hit on - however, I’m still not 100% satisfied with the results.

Thus far, I’ve got the 2007 and 2008 numbers inputted into an excel spreadsheet (that’s too big to upload to google docs as-is, so I’ll have to do some more copy/pasting and get it up on google docs or edit grid). The things that I’m tracking are:

Avg/OBP/SLG
Isolated Power (IsoP)
Strikeout and Walk percentages (K/PA, BB/PA  — Note: PA’s are estimated since I don’t have Reached On Error results)
A Speed Score that resembles the one Bill James invented many years ago (before I was born).
Stolen Base Runs (SB*.22)-(CS*.33)
Weighted On Base Average

In addition to this, I’m using the Park Factor numbers from Boyd’s World (invaluable tool in my analysis). This allows me to use the Total Park Factor (TPF), which is the park factor of all the stadiums in which a college team played in over a 3 year stretch (much more reliable to use the 3 year stretch than the single year park factors due to their vast fluctuations year-to-year — especially in college baseball). In turn, I’ve used this to park-adjust all of the hitting statistics to give me park adjusted Average, OBP, Slugging %, Isolated Power, and wOBA. I also got the average wOBA for each conference and then averaged the park factors for each conference and got an Average Park Adjusted wOBA (APAwOBA?) for each of the conferences I tracked. Using this and a players Park Adjusted wOBA, I’ve calculated a Runs Above Average total versus their conference peers.

What I plan I doing in my next post (maybe two posts), is looking at the Padres 2008 draft. The Pads took 21 college players, 16 of which are among the 1988 entries into my system thus far. Of the 5 I don’t have, one was a Division 2 college player and the other 4 were from smaller schools who weren’t in conferences I tracked. The ones that I did track, and have 2 seasons worth of data for, are the SEC, ACC, Big East, Big West, Big 10, Big 12, Pac 10, Mountain West, and Conference USA. I would like to add in the other conferences (like the Mid American, for instance), but some of the data is missing and that’s not as straight-forward of a process, but it’s not impossible either.

However, having 2 years of data and 16 Padres 2008 draft picks, I think is a decent starting point for a possible two-part breakdown of the hitters they’ve selected. I’ll hope to have the first 8 college hitters the Padres selected up Thursday night or Friday afternoon.

Rob Neyer’s post (h/t: GLB) about Jake Peavy and Petco got me thinkin’ about park factors. The conventional* way to figure a pitcher’s park factor for a season is to take his home park PF and 100 (for all other parks) and average them together. So Peavy pitches in Petco, which has a PF of, let’s say, 84 for a given year. We do 84+100/2=92. That’s Peavy’s PF for that year, and that’s what we’ll use to adjust his stats.

*When I say conventional, I mean the norm at saber analysis type blogs, message boards, etc., not like hardcore analysis, where I’m sure things often times get much more technical.

But don’t we really want to look at how often he pitched in both his home park and different road parks (and not just assume his home/road starts are split evenly and his road parks average out to 100 PF)?. What if he pitched a bunch of road games (and, thus, not as many as we’d expect at pitcher-friendly Petco)? Or what if he pitched a lot in Colorado and Arizona, two extreme hitter’s parks? Well, I looked at Peavy’s 2003 season, start by start (I chose ‘03 because it’s the year Peavy had the most road starts compared to home starts …), to see if there’s anything here.

Short answer: No, there probably isn’t much to it … stick with the shortcut method.

Long answer: In 2003, Peavy pitched 90.7 innings in Petco (PF: 84 — thanks, BR). His road parks combined PF was about 101 (104.2 innings). As you can probably see, there’s nothing too extreme happening here, as Peavy’s road PF is too close to 100 to make it interesting. He pitched five times in San Fran and LA, two pitcher’s parks, and just four times in Arizona and Colorado (for only 19.7 innings combined). He also threw in some other heavy pitcher parks, such as Seattle, Cincy, and Cleveland. It all pretty much got balanced out to 100.

His old park factor, the quick and easy one, was 92. The new one, going park by park, is 93.3. His old adjusted-ERA was 4.47; his new one is 4.41. Whoohoo! .06 ERA points!

Let’s look quickly at a made up example: Let’s say Peavy pitched 90 innings in Petco again. This time, however, he pitched a combined 36 innings in Colorado and Arizona. The rest of his road parks, let’s say, come out to a PF of 98 (and about 70 innings). If this actually happened, his new PF would be 96.7 — a pretty big change over 92. His adjusted ERA would fall to 4.25, a .20 point drop. If you’re valuing him over replacement level, the more technically correct PF would net him like 5 runs, or approximately a half a win. Noteworthy, at least.

If you care to look closer at what I did, here’s a little spreadsheet: peavypf.xls

The real Peavy season from 2003 is on the left; the made up one is one the right …

So, really, the usual method is fine. You’re probably not going to gain much from getting more detailed, although it could have an impact if you can spot a weird season, where someone plays a disproportionate number of road games in extreme parks.

MGL links to an article in the NC Times about building a winning team in Petco.

Pretty good quote from Sandy Alderson:

“You’ve got to be careful about tailoring a team too closely to your home park because you have to play half your games elsewhere,” Padres CEO Sandy Alderson said. “Secondly, you have to be sure you understand how your home park plays. That’s always debatable. It’s a tough hitter’s park. We brought the fences in a little bit after ‘06, and that hasn’t really changed the nature of the ballpark. You start with good pitching.”

Anyway, I don’t have much to add; just wanted to pass the links along, as this stuff has been discussed quite a bit here and all over the Pads’ blogosphere. Feel free to comment here, or of course take it over to The Book Blog (or the NC Times if their comments policy doesn’t scare you off ; )

Going into Baseball Musings mode for the moment, although I could never do it as good as Pinto.

David Gassko has an article up at the Hardball Times on batted ball park factors (as well as other stuff like k’s and bb’s).

There is very cool stuff. Also, it’s already regressed to the mean, which, well, I’ll let David describe it:

Luckily, there is something we can do to measure and take into account the effect of randomness on any statistic, park factors included. That concept is known as regression to the mean. Essentially, we can add a certain amount of average park effects to estimate a park’s “true” impact on a given statistic. The more luck involved, the more heavily we move the measured park factors toward the average (which is by definition 1.00).

So how do we figure how much to regress? One thing we can do is find the correlation between park factors in a given category in one year and the next. The stronger the correlation, the less luck is involved and the less we regress.

At the bottom he gives you a spreadsheet with all the data for all parks.

Definitely some interesting stuff for Petco:

K’s: 1.08 (that is, Petco inflates strikeouts by 8%)
2b’s on fly balls: .88
3b’s on fly balls: 1.09
hr’s on fly balls: .86
2b’s on line drives: .91
1b’s on grounders: .97
2b’s on grounders: .97

Pretty much things we would expect, I think, although the k’s are interesting. Gassko mentions humidity as a possible reason for this, and it could be other atmospheric conditions. Or it could have something to do with the batter’s eye. How is Petco’s? I’ve never been there, of course, but I think I remember hearing some negative things about it (for the hitters, that is). Perhaps the hitters try to change their approach in a big pitchers park like Petco and end up striking out more …?

I’ve heard a lot of people say that the Padres should look for lin drive hitters (think I’ve said it myself a few times). Quietly, from my parent’s basement, I’ve never understood why. Well, Gassko shows that a “line drive doubles hitter” is going to be hurt quite a bit by Petco (unless, I suppose, his line drives are not the same as the average line drive). However, all other hits on liners (singles, homers, triples) are not really suppressed by San Diego’s park (although I don’t believe many homers are classified as liners and there aren’t many triples in general). Also, another thing about line drive hitters; they are generally good hitters. So, yea, who wouldn’t want one of those?

I’ve mentioned ground ball hitters not being affected by Petco … turns out, they would be, although the effect is not that large.

Anyway, a lot of fun stuff to chew on thanks to Mr. Gassko.

From the bottom of a recent Krasovic article:

“I did come across a theory recently, I can’t remember where, that if you could put together a great pitching staff, you might be better off building the best hitter’s ballpark,” Garagiola said. “The theory is that if you put great pitchers in a great hitter’s park, they still will do what they do because they are great pitchers, and they are then beneficiaries, because their hitters aren’t facing them. It’s kind of counterintuitive, but it makes sense.”

At first I was like, “huh?” But maybe this does make some sense. If you have a hitter’s park, it’s generally going to benefit hitters who put the ball in play. If you have a pitching staff full of high k pitchers, they theoretically won’t be hurt by the park as much as the average pitcher, which is what you’ll be facing throughout the year. Does that make sense?

I’m not sure if that’s right or if the theory holds true at all, but it’s interesting to think about. So what if you build a park that favors fly ball hitters. Load your team with high % fly ball hitters and ground ball pitchers, right? I guess the point is to build a quirky park because it’s easier to take advantage of, beyond the normal advantage that comes with playing at home. So the question comes back to, how do the Padres do that with Petco?

The Padres have a .549 w% at home and .517% on the road (since moving into Petco in 04). If I’m reading this correctly, I believe that’s well below historical trends. So despite having a quirky park, or at least an extreme one, they’ve performed worse at home than you’d expect.

I suppose one way to gain an edge would be to have a bunch of ground ball hitters. The challenge with that is that they have to be good enough hitters to score runs and win games (especially on the road when the park will be more normal). I mean if I played I may only put 10 balls in play all year. I may not be hurt as much by Petco as the next guy, but a team of me would probably score about .2 runs per game.

Have any ideas on how to “beat” Petco or how to use it more to our advantage? Or any thoughts on other parks or park effects in general? Let me hear ‘em.

This is not a good study (if you will even call it that). I debated whether or not to even post it, but I decided that since most of my stuff is crap anyway, I may as well go ahead. Anyway, the theory goes something like this: The large outfield in Petco is a major reason why Greene struggles more at home than expected (or, perhaps, it isn’t). Either way, I figured we should forget about Petco and look at how he does at road parks of various sizes (like Phantom suggests). The question then arises: how do we estimate the size of an outfield? MGL used the scales on this site and a computer tracing program to do just that. I’ll do what he did and classify outfields with 106,000 square feet or under as small, and ones with 116,000 or over as large.

For large parks (that, in his career, Greene’s played in) we get Arizona, Colorado, Detroit, and Washington. That’s a total of a whopping 328 PA’s. In small parks (Chicago — NL, Cincinnati, Florida, Houston, Boston, Philly), Greene has racked up a measly 292 PA’s. How has he hit in each?

Small parks: .223/.277/.318
Large parks: .330/.393/.625

Before you start listing the problems, here they are:

• The sample size is ridiculously small — somewhere around half a season in both cases. The variation could just be all, or mostly, randomness.
• Parks have changed. MGL’s calculations were for parks in 2007. Greene’s career numbers are used here.
• Obviously, a large outfield does not equal a pitcher’s park. Colorado and Arizona are two of the best hitters parks in the NL, and they’re also the largest. There are numerous other important factors like weather and air density.
• It could be other factors like pitchers/defenses faced causing much of the disparity (if it isn’t simply random variation).

I’m sure there are many more … again, please don’t take this one seriously at all (not that you will). Anyway, the point is (I think) that a large outfield does not mean Greene will struggle in that park. It is obviously more than that. When Greene hits the ball in the air, his subsequent success is probably determined by multiple factors, including outfield size, weather, air density, and so on.

Petco, by the way, is the 4th largest outfield that he’s played in. Of course, he hits just .230/.292/.377 there. What’s the main difference between Colorado/Arizona and San Diego’s parks: altitude and weather conditions.

Further (er, better) research is clearly needed.

Yesterday’s post received some interesting comments. I figured I’d try to address some of them here (okay, there were only 3 not counting my own) and ramble on about some similar topics.

Here’s what Kevin (from Padres Nation) said in comment #1:

Petco suppresses all players, or at least the theory goes. That’s always been the theory of park effects as I understood. If you had the road stats of all of the league’s shorstops, that would be a good, although not perfect way, of judging them.

I think the point that Petco suppresses offense for all players is right on (not that this is surprising). There may be a rare player or two who would benefit from Petco, but by and large, it hurts offense of all forms. However, my point was that Greene is hurt above and beyond the average player (who is still hurt) because of how he puts the ball in play (or something else, perhaps). Let’s look at some numbers. I don’t have shortstop data, but I loaded every team’s stats into a speadsheet (from baseball reference) and then took out the Padres.

NL players not in Petco (in 07): .268/.336/.427
NL players in Petco: ………….. .232/.294/.354

So, on the road they get a 15.5% boost in average/14% boost in OBP/21% boost in SLG.

Now let’s look at Khalil’s career (note there are a few PA’s from Qualcomm that I didn’t take out):

Greene on the road (career): .280/.335/.515
Greene in Petco (career): .228/.288/.370

On the road: 23% boost in BA/16% boost in OBP/39% boost in SLG.

Greene is hurt more by Petco (or helped more on the road) than the average player, especially in slugging percentage. I am not really trying to prove anything by these numbers. It’s only a 1 year sample for the league. For Greene, it’s his career, but it’s split in half. There’s still a large margin for error and it’s not like there are any tests for significance or anything here. Consider this just for illustration and bear with me if you can.

From this small exercise we can conclude, at least by these numbers, that (with the note from above in mind):

• Hitters are, on average, a lot worse in Petco than in other parks (Yeah, I think we knew that)
• Greene is worse at home, relative to his road stats, than other players (the margin is wider).

The question then becomes: in what ways is Greene hurt?

Enter Didi in comment #6:

He’s still able to hit HR at Petco, with far fewer doubles than on the road. Including 3B, he had 30/63 hits are XBH at Petco, 44/92 on the road. That difference of about 30 hits is what killing his OBP at home since he drew about the same number of walks.

I wonder if he is approaching his ABs differently at home than on the road. And it’s possible that what would have turned into 2Bs on the road were turning into outs at home due to the size of the OF and the better OF defenses (maybe).

That prompted me to look up Greene’s career numbers on the road and in Petco … but this time we’ll look at singles per ball in play (singles/AB-SO-HR), doubles per BIP, etc.

In Petco:
1b/BIP: .190
2b/BIP: .056
3b/BIP: .010
hr/BIP: .035

On the road
1b/BIP: .184
2b/BIP: .117
3b/BIP: .005
hr/BIP: .060

To make that a little more clear, let’s do it per 550 balls in play, which is about how many Greene had last year.

In Petco
1b/550: 104.5
2b/550: 30.8
3b/550: 5.5
hr/550: 19.25

On the road
1b/550: 101.2
2b/550: 64.35
3b/550: 2.75
hr/550: 33

Plain and simply, he’s been an extra base machine on the road. And that’s the same area where he’s getting hurt in Petco. I don’t think that’s surprising, really. I mean, you don’t hear many people talking about Petco hurting singles. It may increase triples a little, but they just don’t happen enough to be that significant. Again, remember this is just sample data, so don’t make any conclusions based on it. I’m just pointing out what’s happened so far for Greene. Like everyone else, he’s hurt by Petco. It just so happens that he’s hurt a lot more than the average joe.

The next question, I guess, becomes: why is he hurt more than the average player? I’m not sure, but I guess I have a few ideas. He hits a lot of balls that are caught in Petco that would be extra base hits in most other parks. In a more technical sense, Greene could hit balls at a certain angle, bat speed, etc. that work fine on the road, but get caught up in the heavy air in Petco Park. I don’t know how to save it and paste it here, so here’s his hitting chart from mlb.com (you have to customize it … it also won’t let me link it to a customized version). Anyway, I’m really not sure. It’s an issue for another day — or an issue for someone far brighter than I am.

More park factor fun

I’m in park factor mode here lately, as you can see. Geoff Young wrote an article for the Hardball Times on PF’s and, specifically, Petco in 2006. Good read.

The topic of taking advantage of your own park (as a team), which Geoff gets into a little,  is probably yet another topic for another day (among the long list of topics I’ll never get to).

Anyway, here’s my question (that I already asked in the comment section of the last post): If you have unlimited resources (i.e., any kind of data you want), how would you construct park factors? Well, here’s my take. Why can’t we take each ball that a player hits in play and attempt to strip it completely from it’s context.  That is, if we know the angle off the bat, the speed off the bat, and things like that, we can say, to a certain extent, what would happen to this ball in a completely neutral environment (weather, fences, air density, etc.). Then we could take each ball that a hitter put in play and give it some kind of percentage of a hit, based on what usually happens to balls with those parameters. So, for instance, if you hit a ball at an angle, speed, etc. that usually produces a 400 foot fly ball, you could call that a home run, no matter what actually happened to that ball. Of course, I’m just thinking out loud here — this would take tons of work and may not even be practical (or valuable). If there’s anyone who will do it, though, it’s probably Greg Rybarczyk. From the article:

Hit Tracker in its usual form uses observations of hit outcomes (landing point, time of flight) to derive the hit’s initial parameters (Horizontal Launch Angle or HLA, Vertical Launch Angle or VLA, and Speed off Bat or SOB, with spin assumed to be a function of these factors). But, with a few lines of code added, it becomes “Hit Whacker,” using HLA, VLA, SOB and atmospheric inputs to generate a hit’s outcome. With this capability, we can create a procedure for assessing how easy or hard it is to hit homers in any park.

To cover the range of possible batted balls that could become homers, I created a “test set” of trajectories, representing 45 different HLA’s (every two degrees from foul line to foul line), 41 different VLA’s (15 to 55 degrees) and 26 different SOB’s (95 to 120 mph). That’s 47,970 different fly ball paths! I ran this complete test set in each park, in that park’s actual altitude, in the park’s average game time temperature from 2002-06, with no wind (I’ll describe how to account for different winds shortly). The trajectories were evaluated as “home run” or “not home run”, and the results were compiled.

That is essentially along the same lines as what I’m trying to say here. Of course, this is all for component factors, rather than runs based ones.

Anyway, my point of this mini-series, if there is one, is to know what question you’re asking before you go looking for the answer. You may not get there initially (we surely haven’t), but the ride will probably be more valuable.

Or it will just chew up 2,000 words.

Over at Sacrifice Bunt, Melvin wrote a post on Khalil Greene yesterday. That prompted, believe it or not, yet another debate about Greene over at Ducksnorts. As usual, the debate was productive and informative. I’m not going to rehash the whole thing again, but I am going to try to put some of my thoughts into words here, as I’ve been thinking about Greene (again) of late. As always, translating my thoughts into words is always a risky proposition, so correct me if I mess something up or stop reading right now for your own sanity.

Ability vs. Talent

For a great article on this turn to, as usual, Patriot. This is really an issue that I think could possibly clear up a lot of disagreement, or at least allow people to understand what they disagreeing about (and if not … well, I just feel like writing about it).

If you’re trying to measure a player’s ability, you want to remove him from his context completely. That is, you want to remove Greene, at least theoretically, out of Petco and put him in some neutral park. So to actually adjust for this, you want to use component based park factors (like for doubles, hr’s, k’s, etc.), rather than the usual run based factors. What you’re trying to do here, I believe, is estimate the player’s actual ability in a context-neutral situation. So why is this relevant with Khalil Greene?

Well, there’s been a lot of talk lately about Greene and Petco Park. Basically, people are saying things like, “Greene may be hurt by his home park more than any other player.” I’ve said things like that myself, even though I’m not sure if it’s actually true. Here are Greene’s career splits so far (close to 1,100 PA’s at home and on the road):

Home: .228/.288/.370
Road: .280/.335/.515

For a little context, here’s what the Padres did as a team last year:

Home: .235/.310/.378
Road: .265/.333/.440

That is obviously a pretty big split, but it’s no where near Greene’s.

Let’s set randomness aside, and say that Greene really is hurt more by Petco than other players. Let’s say, for example, he has a tendency for hitting long, high fly balls that get hung up in the thick air of San Diego and fall for outs, rather than home runs. A normal hitter, with a more normal distribution of balls in play, will not be hurt as badly by Petco. Note that I’m not really sure this is the case with Greene, I’m just using it to (try to) eventually make my point.

Okay, so about that point … what does this mean? Well, I think it means that if we moved Greene to a completely neutral park, he may very well hit like he’s hit on the road throughout his career. Then, what does that mean? To me, it means that he has more value in a trade than his overall numbers would suggest. Say, if the Braves wanted to deal for him … they don’t care what his line would be for 81 games in Petco and 81 games in road parks. They want to know what it’d be for 81 games in Atlanta and 81 games in road parks.

What does it mean for the Padres? Well, it means that, like I said, they should be able to get more for Greene in, say, a trade because of it. He’s worth more than his numbers, even when park adjusted (with a runs based factor). But, and I think this is the important part, in terms of resigning Greene, they obviously have to consider the context. The fact is, they are going to play half of their games in Petco, and unless they move the fences in where many of Greene’s long flies go, he’s still going to struggle at home.

My point here (and hopefully it’s at least somewhat clear and on the mark) is that although it’s nice to say, for example, that Greene would be a stud in a neutral type park, it really doesn’t matter to the Padres if they decide to keep him. They don’t play in a neutral park. They play in a park that hurts offense a lot, and may just hurt Khalil Greene more than it hurts the average player.

Khalil for MVP

Mr. Greene won the Padres team MVP the other day. I’m not particularly interested in who won the award, but I do like to think a little about the thought process. Anyway,  I’m of the opinion that for things like MVP, something like WPA (that’s Win Probability Added) should be used, although I could certainly be swayed. The question you want to ask is what are you trying to measure? Ability? No, I don’t think so, not for most valuable player. Value? Well, yes, but what kind of value? Stats like VORP and basic linear weights ignore context, at least as far as runners on base and the certain “clutch” nature of an at bat. In many cases this is preferable. Since we know that clutch ability doesn’t really exist, at least to a large degree, then we don’t really want to measure a player’s clutch ability if we’re trying to measure their performance going forward. A double in the first inning with the bases empty is worth just as much as a double in the 9th with the bases load, down by two. A double is a double.

However, if you want to measure value, in a backwards looking way, then I think something like WPA is more appropriate. You could also probably use something like Win Shares, as that also has a clutch component and is a somewhat related stat (and there are other similar stats). But, for simplicity, let’s stick with WPA for now.

Greene was -.51 in WPA last year. He was actually very un-clutch by fan graphs’ clutchiness stat (-1.10). Now, just for completeness, let’s say he was +10 runs defensively and he gets +5 for position. We’ve got him at like 10 runs above average or 30 above replacement (these numbers should be relatively close, but they’re just for illustration only, and definitely not exact).

Now, to make it fun, let’s say pitchers can’t win the award. So Greene’s competition is Adrian Gonzalez. Gonzalez’s WPA was 4.07. His clutchiness was 1.84! That doesn’t matter much going forward (surprise, he was negative in the clutch in 05 and 06), but for something like this, (I think) it should be considered. Just for fun, let’s only give him +3 for fielding and -10 for position. 33 runs above average or about 53 above replacement. By my very crude calculations, it looks like Gonzalez should have ran away with that award.

WPA for fielding

You could of course say that if you’re going to use a WPA approach for offense, then you should use one for defense, as well (for MVP’s). I would say that is correct, but I haven’t seen much discussion or research into this area. I am guessing that’s largely due to the fact that people aren’t that interested in it, and it would probably be a lot of work. But I’m sure there are some players who have clutch fielding years. That is, they make a lot of big plays when the game is on the line (or a lot of runners are on base), and perhaps don’t do as well in less critical spots. None of the fielding stats out there attempt to correct for this because they’re much more interested in a true talent measurement, and things like clutch fielding are probably only going to get into the way. But, yeah, for the MVP I do think that theoretically this would be the way to go.

So, there you have it. A long rambling post about something, I guess. A lot of words. Bad writing. Little substance. That’s the Friar Forecast. Until next time ….

By the way, I’m surprised Derek Jeter fans don’t go to a clutch fielding argument more often. Like … yeah, he sucks in the first inning but he turns it on when it counts … these stupid defensive stats don’t account for that, and therefore they’re intentionally biased toward Jeter. Did you see The Play! WPA fielding, baby!