THE BOOK--Playing The Percentages In Baseball

Langerhans went to the A’s from the Braves.  JC at Sabernomics starts with:
http://www.sabernomics.com/sabernomics/index.php/2007/04/thoughts-on-the-langerhans-deal

Langerhans is not a bad player

I responded on his blog, but until it shows up, I’ll post it here:

Not a bad player?  This guy is forecasted by multiple systems as an average to a bit above-average hitter.  He’s also an above-average fielder.  He’s at least an average MLB player barring unknows, like injuries of course.  He’s only got two full seasons under his belt, meaning that he’ll hit free agency when he’s a below-average player.  This is not the kind of player that is traded for a PTBNL.

This trade can only be defended if the PTBNL is a regular player on the A’s roster, or that his poor hitting this year is based on something real that we’re unaware of.

I also see that cash is part of the deal, but this guy will be severly undervalued, so whatever cash is passed over will surely not offset this deal.

He’s 27 and never had a season as an above average hitter.  OPS+ is 99 and 90 the last two years, and 86 for his career.  The projections on Fangraphs all have him around average, but he’s well below average for a corner outfielder.

He is an outstanding defender, so if he’s in center field he has value, and certainly is a good option as a 4th OF, so the Braves should not have traded him for almost nothing.

I dispute the “well below average for a corner OF”, without also considering fielding.

His forecasts are all around average.  If he’s an average fielder for CF, and CF is “neutral” as a position, then, guess what, he’s an average player.

But, let’s say you put him in the corner OF.  All of a sudden, a league average hitter is about -8 runs per 600 PA in the corner.  But, an average fielding CF would be +8 in the corner.  Once again, you are at average.

So, everyone who calls him a great “4th OF”: why is he not considered a fine regular OF?  The dude is average, at least.  And the PBP fielding metrics consider him an above-average fielder in CF, which would make him a great corner outfielder, fielding-wise.

***

In his last year in the minors, he was the best, or one of the best, hitters on a team that finished at the top.  It at least gives us a prior consistent with our belief that he must be at least an average hitter.

***

The interesting thing about Langerhans are his home/road splits:
Turner field: .298 .382 .442
elsewhere: .189 .293 .318

Those are pretty wide.  His K/BB are 2:1 both home and away.  He’s got 34 XBH at home and 26 on the road.  Nothing remarkable.

But his BABIP: .377 at home and .236 on the road.  Only 3 GIDP at home, and 12 on the road.  This is based on around 250 BIP.  A true .305 BABIP will be at .247 to .363, 95% of the time.  He’s pretty much beyond 2 SD in both those cases.  Unless Turner field somehow is really conducive to his game, he’s the outlier of all outliers in this regard.  i.e., can still be pure luck.

I should correct that the PBP metrics have him great in LF (meaning at least average in CF).  Fans also view him as an excellent fielder:
http://www.tangotiger.net/scouting/pos2006_LF.html

When I said well below average for a corner OF, I meant his hitting.

Coming into the year I had him as dead average total, -5 hitting, -5 position, +9 defense, and +1 baserunning.  He’s average in a platoon role, who knows how he’d handle fulltime play?  He’s actually hit better against lefties in what little time he’s had.  Don’t know if he could handle the grind of playing every day if he was given a chance.

How much do you dock him for the 3 for 44 start?  Small sample, but it does represent 6% of his major league plate appearances.  It certainly can’t help his projection.

Still, I wouldn’t gave traded him without getting much in return.  If I had a better option in left I’d at least keep him around as a 4th OF until I was convinced he couldn’t hit.  Maybe the Braves are convinced already due to reasons beyond the stats.

You never know, he might still have a 50 million dollar contract in his future.  Gary Matthews Jr. was one year older than Ryan when he hit .204/.250/.327 in 41 games for the Orioles.  Langerhans has the defense, patience, occasional power and low BA’s of a younger Matthews.

Hey guys, this is by far my favorite blog going right now. I rarely post here because I am literally intimidated by the knowledge you guys (tango, mgl, and all the great comments) have about baseball and sabermetrics.

Anyway, with that being said, can you use a binomial probability test to infer anything about Langerhans’ early numbers? If we consider him a “true” .264 hitter (his marcels)...I plugged in the numbers and got a probability that he would have 3 or less hits of .001041.

Can we make any valuable inferences by doing something like this?

Nonetheless, I think it’s a fine move by Beane assuming Langerhans will turn his offense around and considering his already above average defense. Although, it appears Bradbury doesn’t think the trade has anything to do with the ugly 52 PA’s so far, I’m not so sure Beane didn’t maybe, just maybe take advantage of Langerhans’ slow start this year.

I can agree with your general assessment of having him as an around average OF, though I don’t agree with “platoon” as the term: 15% of his PA against LHP, whereby the league average for a LH v LHP is around 17%.  So, as a LHH, he faces as many LHP as any other LHH.

I also don’t agree with “grind”, as a reason.  He’s had 500 PA in just about all his minor league seasons (if not leading his team, then pretty darn close).

And with Andruw a free agent, this is the guy you’d want there.  Paying Langerhans 1 million$, and wisely spending the other 17 million$ elsewhere is better than getting rid of Langerhans, and giving Andruw 18 million$.

I have him at around -3 in projected hitting, +10 in UZR at the corner, and around even in baserunning.  It also looks like he has a CF arm, maybe even a little weak for a CF, but at worst -1 in CF.  I think that he easily projects as somewhere near an average major league player in the prime of his career, age-wise.  Definitely a bargain at a pre-FA price.  He is worth 5-7 mil on the FA market.

Next year, Andruw will be worth around 3.5 wins above replacement, or 10-14 mil as a FA.  At 18 mil for Andruw, that is too much to pay at 5 mil per marginal win.  Definitely worth playing Langerhan in CF and spending the other 17 mil.  You ought to be able to get almost 5 wins above replacement for 17 mil, which means you get almost 7 wins, say 6.5 wins for 18 mil with Langerhan and only 3.5 wins with Andruw.  That is quite a difference, i.e., a terrible trade for Atlanta.  Of course that is the main reason why ATL traded him - the poor showing in 52 PA this year.  That is what MOST (or maybe many) teams do (bench, trade, or send down a player in a prolonged slump) unless someone is an established good player or a top prospect like Alex Gordon.

Speaking of the A’s, what do you guys think about the Chris Denorfia/McBeth+PTBNL deal?

Personally, I prefer Merchant of Venice to McBeth.

Denorfia
- selected out of college as 555th player
- reached majors at age 25
- has 160 career PA
- on 60-day DL due to surgery

Not someone to think much of.  However, he seems to have hit well in the minors.  I would have presumed that he was a terrible fielder, but Reds fans think otherwise:
http://www.tangotiger.net/scouting/scoutResults2006_CIN.html
(If you click on his name, you will see he compares to Baldelli, Payton, and Finley.)

I’d like to know more about his career path.  Can you help us out Josh?

I didn’t know that much about him before the A’s traded for him. Here’s what I’ve found out:
Hit .340 without power at Rookie ball at 22
posted a dreadful 236/317/304 line in the Carolina League at 23
At age 24, turned into a minor league version of Kirby Puckett (slightly skinnier) for 2.5 years, then plays a half-season for Cincinnati, and gets on-base, and does nothing else.

Most articles on the trade I’ve seen have rated Denorfia as a good CF, excellent corner outfielder.

So, here’s an oufielder who didn’t hit above rookie ball until he was 24, and has 2 HR in 144 ML at bats.

McBeth is a 26 year old converted outfielder who has torn up the minor leagues as a relief pitcher.

That didn’t last long with Langerhans, guys. I’m watching the Padres-Nats and it appears the Nationals have traded Chris Snelling for Langerhans.

Snelling is 25 and Langerhans is 27. My first impression is that it’s a pretty fair deal. It looks like Snelling may be a little better offensively, but when you consider the fact that he’s probably stuck in a corner and doesn’t play defense as well as Langerhans, it probably makes them a lot closer.

Anyway, it looks like Beane comes out the big winner picking up Snelling for a PTBNL. I’ll let you guys go to town on this deal.

When I said Beane got Snelling for a PTBNL (or cash), I meant that essentially that’s all he gave up for him.

Here’s a link to the deal:

http://washington.nationals.mlb.com/news/article.jsp?ymd=20070502&content_id=1943606&vkey=news_was&fext=.jsp&c_id=was

http://shysterball.blogspot.com/2007/05/that-ryan-freel-is-magnificent-bastard.html

But let’s not forget that [Freel]’s not much more than a useful utilityman who somehow conned the Reds into thinking that he was a starting centerfielder.

First Langerhans, now Freel?  It’s like we haven’t made any progress in 20 years.  I wonder what people would say today if they saw Gary Pettis?  Freel’s career OBP is about 30 points higher than average, and his slugging is 30 points lower.  (His 2007 totals exactly match his career totals.) He is, in short, a fairly average hitter.  If you want to argue he’s a bit below average, fine.  Fielding-wise, he’s at least an average-fielding CF, if not an above-average fielding CF.

This is Ryan Langerhans all over again.  Is there a systematic bias against excellent fielding and average-to-below-average hitting OF?  A few years ago, it was Darin Erstad, Mark Kotsay, and Mike Cameron.

If you were to run a regression of outfielders of fielding (using UZR, Fans, or some version of Zone, CF/notCF) and hitting, against salary, I’d bet you’d find a gap.

Roger Clemens: Marcel has him forecast for a 3.13 ERA.  Bill James (BIS) too.  The average of Chone and ZIPS is 3.12.  Clemens career ERA is 3.10.  In short, we see Roger at his average!  Incredible to even consider at this age.

Anyway, I’ve already shown in another thread that Clemens’ career win% is fairly representative of his true talent (.660).  That makes him +.280 wins above the .380 repl level, per 9 IP.  Giving him 20-23 starts of 6 IP per start means around 15 full games.  15 x .28 = +4 wins above replacement.

The free agent value is 4MM per win, or 16 MM.

Assuming he gets a pro-rated two-thirds of 28MM, he will be paid around 18-19MM.

A fair signing if a bit high.  But, that extra 2-3MM is the Picasso value to YES.

Back in April, Westbrook signed a 3/33 extension, starting in 2008.  Take this chart, and multiply all numbers by 1.10:
http://www.tangotiger.net/salary.html
and you see he was valued at between 2.5 and 3.0 WAR.

He’s basically a .500 or a bit above pitcher, meaning he’s +.12 wins above replacement per 9 IP.  Give him about 20-22 full games, and that puts him at 2.5 to 3.0 WAR.  A BINGO signing.

***

Buehrle is on par with Westbrook.  Marcel entering 2007, conveniently enough, has them both at 186 IP.  Their K’s are almost a match.  Westbrook has 10 more walks, but 8 less HR.  Buerhle has a better BABIP.  Overall, their Marcel ERA are almost identical.

BIS had Buehrle with a +.18 ERA advantage, while Chone and ZIPS has an even larger advantage for Westbrook.

This year, Buehrle has had the good fortune to have an even lower BABIP than he’s gotten historically (.268 v career of .289), and he’s sequencing his PAs so that the damage is done with bases empty (77% men left on base v career of 72%).  A “walk year” performance that isn’t.

He is pitching better than expected, so we can give him a little bump.  He’s also 2 1/2 years younger than Westbrook.  Let’s be generous and call him between a .525 and .550 pitcher.  That’s around +.15 to +.17 wins above replacement per 9IP.  Let’s be generous on his durability and give him 22-24 full games.

On that basis, some team will overspend to get him for 5/65 to 5/70, when he’s really worth closer to 4/45.

The problem with signing any pitchers to long-term contracts is that they lose around 10% IP per year due to injury, IIRC, even the good ones.  Basically, if a pitcher pitches 200 IP in any given year, even a good pitcher with very good performance, on the average, he pitches 180 IP the next year, 162 the next, etc.  So 5 years down the road he is only pitching 100 some odd IP.  That 10% loss per seaon includes when a pitcher pitches zero IP because he is on the DL all year or his career is over for whatever reason.

You could argue that the 10% decrease in playing time is around equal to the rate of infaltion in MLB salaries, so that it is a wash whether you sign a pitcher to a short or long term contract. 

However, that does not include a decline in performance from age.  For most FA contracts (which are usually signed near or past a player’s prime), decline in performance cancels out inflation.

So between decline in performance from age and natural decline in playing time that all pitchers experience (on the average), signing a pitcher who is not very young to a long-term contract is probably a bad thing.  Not to mention the fact that you always lose some “flexibility” with long-term contracts, not to mention increased risk.

Buehrle, BTW, has racked up a lot of TBF over the years.  Whether this portends a greater risk of injury, I don’t know.  I assume that you can assess each player’s chance of injury and that it would be different for each player, although I think we have not done a whole lot of research in this area (projecting injury and playing time), at least I have not seen much.  It can’t be that hard to do.  Run some regressions with past injury, past playing time, age, etc.

For whatever it is worth, I too have Westbrook and Buehrle rated around the same, and around 2.5 wins above replacement.  I think my replacement is better than the one Tango is using, as I have both of them as considerably better than .500 starting pitchers.  I have them around .3 to .4 runs per 9 better than the average starter, which is around a .525 pitcher.  I think Pecota has around the same.

I thought that we had a thread recently that showed that no matter how you define replacement pitcher, that a replacement starter is quite good, maybe a .450 pitcher, which is maybe .75 runs per 9 worse than an average starter?

A replacement level pitcher as a starter is .380.  (That same pitcher as a reliever is .470.)

An average pitcher as a starter is .470 (that same pitcher is .560 as a reliever).  If you give him 2/3 of his IP as a starter and 1/3 as a reliever, that makes him .500.

An average MLB starter is .490.  (If he were to relieve, he’d be .580).

An average MLB reliever is .520.  (If he were to start, he’d be .430).

Hafner signed a 4/57 extension, starting in 2009.

Wow, what an interesting case… the exact opposite of ARod.  Hafner has been a fantastic hitter, having his worst season since 2003.  His 2004-06 seasons have been fantastic.  I just did a quick back-of-the-envelope calculation, and I’ve got his estimate of his true talent level as dropping by one win between Apr 1, 2007 and today.  That is about as big a drop as you will find.  Hafner was injured at the end of last season.

He’s also a DH, and I give a severe penalty (2 wins, though perhaps I should only make it 1.5 wins because he already gets a DH penalty for being “cold").

Anyway, it’s always tough to give an extension 2 years down the road, especially for a guy who was hurt and has shown a drastic dropoff in forecast.

On Apr 1, 2007, our estimate for his talent level for Apr 1, 2009 would have been say +3.5 wins above average.  Replacement level for a DH would make him a 4.0 WAR player.  Therefore, a 4/79 deal is what you get. 

Now however we drop him 1 win for his estimate because of his poor play.  It’s not a 4 WAR player, but a 3 WAR player.  That makes him a 4/54 player.

First of all, huge BINGO to the salary calculator, Indians, and Hafner.  You settled on the right price.

But, oh boy, what a-three-months-a-difference-makes.  This was a TWENTY FIVE million dollar injury.

Now, the fun stuff.  The forecast is always based on his performance being somewhat random.  That is, he’s putting up horrible numbers today, and fantastic numbers in 04-06.  We don’t presume injuries, though we do presume his current stats are more indicative of his current health.

Since we know he was hurt, we would really need to weight his current stats more.  But, what if the doctors and Hafner say that he’s pretty much healed, and therefore, we conclude it’s a coincidence that his poor performance (for him) coincides with him coming off an injury?

This is an incredible level of uncertainty here.  And, as you can see, millions of dollars at stake here.

I’d be interested to hear your thoughts on Ichiro’s extension.