Imagine if you (i.e. the average reader) were suddenly dropped onto a major league baseball team. You'd be virtually useless in the batter's box - striking out each time (if you're lucky), and you'd be a defensive liability, forcing you into the DH position.

Assuming that you strike out on three pitches in every at-bat (which is probably being overly conservative but whatever let's roll with it), your wOBA would be .000. That's not a good number. Over 600 plate appearances that translates to something like 180 runs below average.

The DH positional adjustment is -17.5, and the replacement bonus we'll take as 20 runs. Your WAR would therefore be approximately -18 (meaning that you should be paying whichever team you were on almost $100M a year to make up for the fact that you're playing).

While this is an amusing little exercise, consider this: Jose Vidro, over 600 PA last season, would have been worth about -3 wins. That's 15 wins over someone who can't play baseball.

The best player in the game, Albert Pujols, posted a WAR of about 9 in 2008. The difference between him and Vidro? 12 wins, only a little bit less than the difference between ol' Turbo and someone who can't play baseball at all.

Let's present this in another way.

We paid him $6,000,000 to be our starting DH last year. What the hell.

Continue reading this post »

It's 'Let's look at some numbers' time again.

Turn to page 191 of your shiny new 2009 Hardball Times Baseball Annual (don't have one? Matthew is displeased with you). Here Tom Tango discusses the impact of catcher defence by considering stolen bases, caught stealing, pickoffs, balks, wild pitches, and passed balls, controlling each for the pitcher in order to eliminate their effect on the battery's defence.

I'll summaraise the findings for you, in case you're currently invoking Matthew's wrath:

The worst defensive catcher of all time is Mike Piazza. The best was Ivan Rodriguez. Hardly surprising, huh? The really interesting thing about Tango's article is that it shows how many runs a year Piazza cost his team. He rated at roughly -1 win per season with his 'defence' at catcher. Pudge is worth about the same in the other direction.

So the difference between the worst catcher of all time and the best is on the order of 2 wins a season. Which isn't very much.

Now let's take a look at some positional adjustments.

C: +12.5 runs
1B: -12.5 runs
DH: -17.5 runs

Remember that the worst defensive catcher of all time was worth -10 runs. Apply this knowledge to Jeff Clement. With some work he could probably become a passably bad first baseman. He's never played it before and he's not particularly agile, but it's an easy position and he's a hard worker. He'd probably end up as a -5 guy overall. And unless you believe that Clement is the worst defensive catcher in history, that's only a 0.5 win improvement.

For 5 runs of defensive upgrade, Jeff Clement would lose 25 runs of positional value. 25! Potential injuries aside, no matter how bad he looks in the field, there is no way that you shift Clement off his postion right now. As an average major league hitter, a level well within his ability to reach, he'd be worth 2-2.5 WAR as a catcher against 0-0.5 WAR as a DH/1B. Start being less conservative with his defence (after all, he's not going to be Piazza bad), and the difference only grows.

Clement's bat can only take him so far. For him to live up to his potential value, he absolutely has to be kept as a catcher. Moving him from behind the plate should be the last resort.

Lookout Landing/USS Mariner meetup at the downtown Seattle library.

Be there, or be elsewhere. Details will be forthcoming.

Matthew's Addendum: Seriously people, you're going to want to be there.

Graham's Addendum: Unless of course you have a dislike of things which are awesome, in which case fair play to you.

Hi AN. The more cosmopolitan of you may recognise me as one of the people who writes  for Lookout Landing, and the statistically inclined may be aware that I developed an advanced pitching statistic that's gaining traction across SBN (including the occaisional mention here) called tRA. I'm hoping that I can explain the motivation and methodology behind this stat to all comers in this diary*, and answer any questions folks here have about its use.

Please forgive the British English.

Motivation

As measures of pitcher skill wins, ERA, and WHIP don't cut it. The holy grail of pitching analysis is to determine pitcher skill independently of his team, and anything involving hits is not going to do that. A's fans are gifted with one of the top defenders in the game in the form of Mark Ellis, and I imagine that there won't be much argument from anyone that he makes your pitchers better with his glove.

Except there's a problem. A pitcher should posses the same ability whether he's pitching in front of the best defensive unit of all time or on a team of eight Jack Custs. The defence helps after the pitcher has done his thing. In order to get a handle on pitcher skill, we must somehow look at what happens before his teammates get involved.

Method

We can divorce the defence from our pitcher by looking at pure pitcher outcomes and batted ball profiles.

The former consist of strikeouts, walks, hit by pitch, and home runs. Barring some robbed homers and the possibility of Jose Canseco being involved, fielders have nothing to do with these outcomes. Batted balls mean the distribution of ground balls, line drives, fly balls, and infield popups. These are the plays that the defence can turn into outs.

I imagine that the vast majority of AN'ers are at least passably familiar with FIP, which is a statistic that weights the pure pitching outcomes against runs scored, and puts the result on an ERA scale. While an extremely useful stat and an upgrade on ERA, FIP fails to take into account the pitcher's ability to exert at least partial control over batted ball profiles.

tRA does. By determining the average runs scored and outs made on each of the defence-independent outcomes in each season and league, we can credit a pitcher with runs and innings pitched simply by looking at statistics that they have control over. The following statistics are accounted for in the tRA model:

Strikeouts (K)
Walks (BB)
Hit By Pitch (HB)
Ground Balls (GB)
Bunts (BU)
Line Drives (LD)
Outfield Flies (OFB)
Infield Flies (IFB)
Home Runs (HR)

The actual act of deriving the average outs/runs for each of these over the course of a season is fairly involved, so I won't go into detail here.

Once you know expected outs and expected runs, it's easy to calculate tRA:

xRuns/xOuts*27. There. Done. We now have a metric on the R/9 (NB: not ERA) that is completely defence independent. Is that enough? Not quite. Parks need to be taken into account. This is actually a relatively easy adjustment, made possible by a tonne of work from the folks at The Hardball Times. So now we've derived a park neutral, defence neutral pitching statistic. Hooray.

I hope that was easy enough to follow. If you have any questions, I'd love to hear from you.

*It's not a fanpost dammit.

Got questions about statistical analysis? Ask. We might be able to help.

There are people of all sorts on Lookout Landing. Some of you are new to this, some are old hands. Everyone should feel free to chime in with questions, answers, points of contention, etc. Just try to be nice.

For the first time in 5 years, Raul Ibanez is not a member of the Seattle Mariners.

i hope you enjoy Phillie, Raul, and that the fans enjoy your frankly astonishing defence.

You may be gone, but... Mike Cameron is still in our hearts.

We've apparently already made a pick: Reegie Corona, out of the Yankees org. Light hitting shortstop, but I don't know much more about him beyond that.

More as we get it, of course.

Update: Dave thinks he's the new Willie Bloomquist.

Another update: What the hell his name is Reegie?

Another pick: Patrick Ryan, in the AAA portion of the draft, out of the Brewers. Strong groundballer but his strikeouts fell off a cliff in AA last year after being extremely effective in A+ in '07.

Yet Another Pick/Trade: In exchange for cash, we acquired Jose Lugo from the Twins via the Royals. Lugo's never been above High-A but he's a groundballer with good strikeout rates.

So it seems like a lot of people are missing a crucial point in player evaluation:

The stats used to evaluate pitchers are not necessarily the ones we should use to look at hitters.

Analysing run prevention is a different beast to measuring run scoring, and that's because we're trying to solve seperate problems with different constraints. The key concept here is that runs win baseball games, and we should consider the twin goals of calculating present run value and future run value as paramount.

When attempting to measure a player's output in terms of runs scored, we need to figure out the individual contribution to what is a team effort.  Since each player is only 1/9th of a lineup, we can't blindly look at a given player's runs scored/runs batted in. We must determine the components of players that best describe their offensive worth, which is where OBP, SLG, and more advanced metrics come in. These turn out to be reasonably stable year to year.

We run into a different problem with pitching/defence (if everyone will permit me to ignore the difficulties in isolating a pitcher from his fielders for a few minutes). We know how to model runs scored. We take... runs scored against. The correlation here is 1.00, so there is no need to look at components to measure present value of the pitching/defence units. The problem lies in the stability of our run prevention statistics - namely that there barely is any. The components used to evaluate a pitcher should be those that decrease the volatility of our future run value estimate, and therefore they need to be defence independent. This is the province of FIP and tRA.

So when you feel like applying tRA to hitters or OPS against to pitchers, remember the context in which the stats were developed. It's very easy to think that they'll be accurate when applied to the other side of the game, but doing so loses sight of the constraints they were designed to overcome.

WHIP sucks.

What's wrong with it? Let me give you a list.

1) Assumes hits given up are entirely under the pitcher's control.

Did you know that defence isn't entirely static? Funny that. Hits don't occur at the same rate in front of different fielders and in different ballparks, and the variation in defence/park effect far outstrips any variation in hit prevention skill amongst ML calbire pitching.

2) Assumes innings pitched are entirely under the pitcher's control.

Sort of the same as point one but so crucial I decided to mention it twice. Outs are determined by pitcher skill, defence, and park effects. Pretending that the latter two don't exist when trying to measure pitcher effectiveness is a bad idea.

3) Improper mapping of different types of hits to different run values.

WHIP values a single the same as as a home run. At least ERA doesn't have this problem.

The problem with throwing together a set of numbers that look vaguely similar and calling them statistics is that generally the result means, in layman's terms, sweet bugger all. For all intents and purposes, WHIP is ERA except worse. At least ERA tries to measure runs, while all WHIP does is... measure walks and hits per inning pitched.

It's crap like this that gives real analysis a bad name. Fantasy baseball has a lot to answer for.