Wednesday, May 03, 2006

Catching up with the Local Baseball Collective

This morning, the ATL is waking up to a 12-14 baseball team, which isn't great, but it sure beats the 9-14 record that was staring back at us on Sunday morning. When the Braves were losers of five in a row, I was trying not to let their slide bother me, mainly because the Braves were losing a bunch of close games (three one-run games, a two-run game, and a three-run game) and as the Nats learned last year, those close games tend to even out over the course of a year. Sure enough, the Braves have won three games in a row by a total of six runs and are back within shouting distance at five games back. The Braves are +3 in run differential, which indicates that they are what they always seem to be in April and May: an average baseball team.

26 games into the season, the villains on the team (other than Jorge Sosa, whose collapse from a misleading 13-3 2005 has been entirely predictable) have been Adam LaRoche and Jeff Francoeur. The Braves are getting nothing from two power positions and from two players who rank 4th and 1st respectively on the team in at-bats. Francoeur probably has a month to sort out his hacktastic hitting approach or else he's headed to Richmond and we're going to have a Kelly Johnson-Matt Diaz platoon in right. (Diaz has been disappointing so far, but we can't judge a guy on 31 at-bats.) LaRoche, unfortunately, is getting to the fish or cut bait stage. This is his third year as a regular and he's gotten off to a sterling .200/.304/.450 start, despite the fact that he's showed enough plate discipline to walk 12 times in 92 plate appearances. (How many of those walks are the result of Francoeur hitting behind him? The Rockies walked LaRoche last night in the 7th with two outs and a runner on second in a 4-4 game and the strategy paid off, as Francoeur struck out on five pitches and looked awful doing so. The good news for the Braves is that it's easier to find players who can hit and play right or first; it isn't as if they're trying to find quality catchers or shortstops on the market. The bad news is that they pretty much exhausted their farm system last year and their top prospects remaining (read: Jarrod Saltalamacchia) are off to slow starts this year in the minors. Scott Thorman might be an intriguing option at first, since he's hitting .318/.389/.506 for Richmond.

As for the pitching, the starters are hot right now, with the exception of Jorge Sosa and his buddy Kuato hanging out above his belt. (Speaking of which, it's pleasant to know that no one with any sense can look at the increasingly portly Andruw Jones and allege that his homer binge in 2005 and 2006 is the result of the cream or the clear.) Davies has the best peripherals on the staff right now, with a solid 27/9 K/BB ratio in 30 innings...if you ignore those seven homers conceded. (Insert stock "sample size" caveat here.) My only concerns are that Davies started off well last year and then tailed off badly and Thomson started off well and then ripped his ring finger open for no apparent reason. We also haven't reached Tim Hudson's annual Oblique Vacation. The bullpen is pitching reasonably well and we don't have to worry about Oscar Villareal's arm spontaneously combusting from overuse now that Hudson, Smoltz, Davies, and Thomson are all pitching well. I still don't feel good entrusting one-run leads to Chris Reitsma, but what are you going to do when you're not willing to play ball in the ridiculously overpriced closer market?

One other thought: I'm grumbly about the fact that the Bravo Club is apparently a thing of the past, which means no free schwag for people like me who go to 15 games per year. I'm also grumbly that last night, the Braves drew a smallish crowd of 22,000 and still, the ice cream stand near our seats ran out of vanilla ice cream, which seems like a pretty basic staple, especially on a warm night.

The good news is that I accidentally kicked a middle-aged Mets fan with a "Spring Break Panama City 2006" t-shirt on Friday night while scaling the railing between the concourse and my seat. I've never felt less guilty about kicking someone in my life.


Ed said...

Regarding Sosa, how (barring injury) is a pitcher's collapse from a 2.55 ERA to a 6.65 ever "entirely predictable?"

I realize that Sosa had a pretty high WHIP for a guy with such a low ERA last year, primarily because he walked a ton of people. He probably should have had an ERA closer to 4.00 based on the rest of his stats. But still...we're talking about a current BAA of .330 and 6 HRs given up in 21 innings (last year, the numbers were .241 and 12 HRs in 134 innings, respectively). I don't see how this degree of horrendous pitching was written in the tea leaves.

Can't argue with the rest of your post. I have to do a double-take everytime I see Francouer's stats and find a big fat "0" under Bases on Balls. With the exception of three games, he's been essentially another pitcher at the plate.

Anonymous said...


Here's where we stat nerds with the calculators and the computers outshine the tobacco-spitting scout types:

Sosa's collapse, by any objective measurement, was inevitable. Two metrics that measure luck, Strand Rate (from Ron Shandler) and BABIP (Batting Average on Balls in Play -- Baseball Prospectus), were heavily slanted in Sosa's favor last year. His strand rate was 85% and his BABIP was .269; major league averages hover around 72% and .300 respectively.

Throw in Sosa's terrible K/9 and BB/K rates last year and you have a pitcher who was very lucky. Indeed, every analyst worth his salt predicted backsliding:

Shandler: "Potentially facing the biggest drop-off of any player."

Baseball Prospectus: "In short, bet heavily on a regression in 2006, but hope that working with a real pitching coach taught him a thing or two."

Oh, and Michael, sorry to crash your blog here. Suffice to say I'm a regular reader, fellow Atlanta resident, and Member of the Tribe (MoT™). Keep up the good work.


Michael said...

What Anonymous said. Sosa's peripherals last year (home runs allowed, walks allowed, and strikeouts) were consistent with a pitcher with an ERA of 5. Thus, an ERA of 6-something is more plausible than his ERA last year. Unfortunately, his replacement in the rotation - Horacio Ramirez - also has terrible peripherals and Anthony Lerew isn't pitching well for Richmond right now, although he did have a good start on Monday. Chuck James would be the ideal fifth starter for this team (with Sosa in the pen and Horacio in the California Penal League), but I doubt that Bobby is going to go in that direction.

peacedog said...

Ed, to hammer home what anon said, it's been discovered that Strand Rate, and IIRC BABIP, normalize over time. So that over the long haul, you can expect to get something close to those numbers. It can vary. . . strand rate is jsut the percentage of base runners you allow that the bullpen prevents from scoring. A good pen helps, *yes*, but even a good pen has bad spots through a season. Ben Sheets had a couple of really good years but mediocre win/ERA totals thanks to a hapless bullpen.

The stat nerds out there are heavy believes in focusing on what some call "base performance indicators" - K/9 and BB/K that anon aluded to. Hr/9 is a big one too (Sosa was very good in that last category last year, though, a saving grace then). Those are for pitchers. Batting Eye, Contact %, Walk %, Power Index, and Speed Index are important for hitters. And groundball/flyball ratio is getting big on both sides of the equation.

There's lots of good reading out there if you find the sujbect interesting. The prospectus is a fantastic site and IMO worth the money, but I love baseball.

Francouer worries me right now. So young and talented, but he's going to have to find a switch up top and flip it, I think.

Salty is off to a slow start - he was actually hitting .200 and pulled it up to .260 before slumping again. But the extra base hits have started appearing. Thorman is having an intersting year. I was kind of hoping KJ would get a few months at Richmond, as they might aid in predicting what we can realistically expect from him in the future.

The rest of the prospects are too young to discuss (Jones, Andruws, Montgomery, Ramirez); and only one of those guys is an OF anyway (Jones). Escobar is a guy who could make the leap in the right situation, but he's an MI and that's no help with the OF. Burrus isn't really someone I ever expect to hear from, and he just got healthy after starting the season injured (but he does play OF; however I think he walks less than Francouer).

peacedog said...

Also, fabulous Total Recall references. 10

Ed said...

Goodness. What I have run into here? The United Front of Sabermetricians?

Joking aside, thanks all for the input. I wonder if I can play devil's advocate here for a moment in my completely unscientific way...

I did remember that Sosa spent a great deal last year pitching with runners on base (i.e. because of BBs), so it makes sense to me that his ERA was substantially lower than it otherwise might have been.

But it seems to me that you all put too much stock in these metrics if you think that a pitcher's collapse by FOUR earned runs from one year to the next is "inevitable" based on some abnormal metrics.

A pitcher, especially a young pitcher, can improve his peripherals. Brandon Webb comes to mind here. He had a great year ERA-wise and terrible peripherals in 2004, only to improve upon those peripherals in 2005.

Moreover, there are pitchers who defy these metrics. Take, for instance, Tom Glavine. In 2001, he pitched 219 innings and walked 97 and struck out 116. In other words, his K/9 and BB/K were absolutely awful. Moreover, his BAA was .261, so that his BABIP must have been much, much below average.

His ERA for that year? a solid 3.57.

Now, judging from the reasoning in the comments here that Sosa was absolutely destined to fall on his face, Glavine was due for a rough year in 2002. After all Glavine's one saving grace that year (like Sosa in 2005) was that he didn't give up the long ball.

Except that he didn't have a rough year at all - going 18-11 with a 2.96 instead. Now, frankly I don't remember if Glavine made any radical adjustments in 2002. I do know that his BBs and Ks didn't radically change, nor did his BAA. I also know that Glavine has had a number of these kind of seasons in his career, and the sky has yet to fall down on him.

Now maybe Glavine is simply the exception to the rule, but there are a number of other pitchers who have been successful with some unhealthy K/9 and BB/K. Russ Ortiz is a guy who comes to mind. Certainly, he has been terribly recently, but he had about a 5-year period where he lived on the edge and did well (most pitchers aren't usually successful for very much longer).

Now, maybe his struggles in 2005 are a testament to the prophetic functions of the sabermetricians, but why didn't he have his inevitable collapse in 2003 after a disturbing metric year of 2002?

In other words, am I allowed to raise an eyebrow at Sosa's high ERA this year? Is it really that predictable?

peacedog said...

Seasons are nothing. The sample sizes presented in 1 season are *not* generally accurate of what you can expect over a longer haul. A career .240 guy can string together .300 for 500 Abs one year. The fact that we have this snap shots and then they are so neatly broken up is certainly interesting. It allows for people to correct things over the longer haul, I think.

That said - what was predicted is that Sosa wouldn't duplicate his succdess. Not that his ERA would be 6.whatever. *that's* important to note. I'll try to look it up later if I can, but Sosa's xEra last year was notably higher than his output. People generally don't pitch far removed their xEra over the long haul. I'll see what numbers I can produce for he and glavine, it should be interesting. It was also predicted that the kind of pitcher Sosa's BPIs (Base Performance Indicators) pointed to was noticeably worse than what he did last year.

Sosa's collapse was inevitable, unless there was a change in his fundamentals (there has not been). The era is just details. His current era may well be *above* his xEra. It,s very possible given that I believe he's notably above his career Hr/9.

Glavine's .269 in 2001 is *way* below the ML average. And it helps explain the year. But it's just one thing.

Here, by the way, is what you aren't seeing in Glavines 01 to 02 shift, because you just glanced at the stats (and there's nothing wrong with just glancing). See, 01 was *way* below Glavine's career norms. K's down a touch. BB's up noticeably. This is a guy who rarely walked as many as 90 in 240 inning seasons. He walked 97 in like 215 in 91. His BB/K was 1.23 in 01. 1.63 in 02. K's were 4.7/5.1 respectively I think. The latter is much more in line with his career numbers. That's a noteable difference. The BB/K number isn't great 0 but glavine was below 2.0 a number of times in his career. He had a down year in 01. Rebounded in 02, but probably pitched slightly above his head. It happens. The outputs look familiar to you, to all of us. But there are differences in those seasons - significant ones. They aren't there at a glance, except in the ERA and Wins. But those are outputs based on inputs.

Era is a nasty stat - for example - because it's filled with a lot of noise. The defense that plays behind a pitcher, the job the bullpen does cleaning up his messes, and the park he plays in can have significant effects on ERA. In fact, those things generall matter as much as the pitcher's ability. It's worse with Wins - since it's almost entirely offense-centric. Ask Clemmons what last year was like (I can get you his number ;) ).

Ortiz benefitted from good bullpens and offenses, if memory serves. And the Voodoo that is the new Candlestick (ain't calling it anything else). And of course he had good backing in 03 with the Braves. He had one era below 3.68 int he span you are mentioning. I think if we are able to dig into the numbers we'll find a mediocre pitcher in great environments who got a little lucky.

But yes, the bottom line is - people performn outside of their expectations all the time. Sabermetricians know this. They expect it. People just can't keep up doing that. If we could mythically have 1000, 1500, 2000 AB seasons (without fatigue setting in other than how it relatively sets in over a 500-600 AB season), and we took major adjustments out of the equation (*major* delivery/stance changes, a philosophy change), yes we'd probably be in the middle of a correction for Sosa right now. The severity of the correction will depend on a number of things, many that aren't controllable by Sosa (or any other pitcher).

Sabermatricians understand all of these things. They spend a serious amount of time scrutinizing their own methods. That's why recently things like groundball/flyball ratio is coming into play for pitchers and hitters, for example. Because a line drive hitter and a ground ball hitter and a fly ball hitter are not the same thing. And power will mean different things to each one of them.

Ed said...

Thanks peacedog,

I'll readily and cheerfully concede the field here on the legitimate grounds that I don't know enough about the subject (though I'm learning from the volume of output on this blog).

I think more than anything I bristled at the notion that Sosa's utterly wretched pitching so far this year was something that sabermetricians were observing
and collectively shaking their heads knowingly. I think his terrible pitching is a surprise (to its degree), but I also understand that the metrics from last year predicted a significant dropoff from the lofty heights of 13-3 2.55 at some point in the future: a sort of market correction for a pitcher, if you will.

peacedog said...

His xEra last year was something like 4.6. I know his actual era was right at 2 runs better, which is outlying (contrast that with Glavine, who during the years we discussed was only a over a run better once). I don't have his current xEra right now, but like I said I bet he's *above* it. And I bet tehre are sabermatricians shaking their head saying "we expected a drop off, but he's worse than we thought".

Shandler said this about him after last season: "Outplayed his skills based xERA by two full runs, thanks to an elevated strand rate and maybe a little Mazzone Magic. . . he won't be able to conjure up in '06. Potentially facing the biggest drop off of any player".

Interestingly enough, I'm pretty sure some guys somewhere - Baseball Prospectus maybe? - did a study and showed that Mazzone does appear to affect the bottom lines of his players. Generall, he makes them better. I've never had a chance to read it, though. I'd like to.