THE BOOK--Playing The Percentages In Baseball

Thursday, June 25, 2009

Psst… wanna work for Fangraphs.com? (part 2)

The Fangraphs train keeps rolling along, as David offers even more paid positions.  Since the readership here and there overlap, I doubt my name means anything, but, who knows…

Linear Weights, by pitch movement

Dave Allen strikes again, this time from Fangraphs:

Wednesday, June 24, 2009

David Wright

Lots of MSM chatter about the non-MSM BABIP.  Love it!  I made these two posts at Primer:

Although I don’t know why I think this, for some reason I have it stuck in my head that Wright’s high BABIP is likely a statistical fluke, while his high K rate and lack of power are likely something to worry about. Can someone explain to me why I am wrong?

You are not wrong.

Every metric has a certain amount of noise to it. For something like K/PA, it has very little amount of noise. For something like BABIP, it has alot of noise.

The important thing is that this is not an either-or situation.

Thus, if I accept the premise that the metric of K/PA has very little noise—which I do—then the version of it I would accept as the most quiet in David Wright’s case is the one that has 2312 PAs of information to silence the noise.

What I was trying to say is that the metric K/PA allows you to weight recent performance more heavily.

If, for example, the “standard” weights for the last 3 years is 5/4/3, then you would weight BABIP as 4.5/4/3.5, and you would weight K/PA as 7/4/1 (or some such).

You NEVER ignore any past performance.  What you do is weight them based on their persistency to forecast the future.

Since players are human beings, then we want a metric that more closely aligns to his current conditioning, strength, and speed.  K/PA has limited noise, and so, you weight his recent performance more.

If players were NOT humans, then you would have no reason to weight recent performance more, and you’d stick with his career totals.

Batting Order and the pitcher

I reply to a thread at Primer:

Batting the pitcher 8th, or moving everyone down a spot in the order and putting the guy you originally decided to bat 9th (when you moved the pitcher to 8th), into the 1 spot?

If this is the choice, it’s a no-brainer, and it’s the former.  In no way can you put one of the worst hitters on the team at the top of the order.

***

As for the general issue, my research in The Book (see it for free on Amazon’s Look Inside) using Markov chains is that moving the pitcher from 9 to 8 will add roughly 2 runs in a 162 game season.  MGL’s research via his simulator is that it’s a break-even or a slight advantage to keeping the pitcher in the 9th slot.

***

The most egregious thing you can do is move the pitcher to the cleanup slot.  This would cost you 0.1 runs per game (about 16 runs in a season).  Basically, moving the pitcher up the order costs you around 4 runs per slot.  Move pitcher from 8 to 7 to 6 to 5 to 4 and remove 4 runs each rung.  That’s the impact of a batting order.  And remember, this is by far the worst hitter on the team.  That’s the impact here.

I presume most people would think that moving the pitcher to the cleanup slot would cost you half a run a game.

Blogosphere Question of the Day, 06/24; OR Why should OPS die?

Inside The Book blog reader asks his blog readers this question:

I was catching up on the issues of By the Numbers and read the following quote in the the November 2008 issue:
“An OPS of .800 will always generate more runs than an OPS of .700, given the same amount of playing time.”
I know the above statement is not always true, but do you?  I want to give out a prize and decided that I the first person to prove that it is false, using math, will get to choose the first team I will study in depth with my new disabled list database.  I know it is not much, but that is all I can really offer.  Hopefully there will be more of these to come in the future.

One of his readers already gave out the answer, and not the theoretical mumbo-jumbo I am about to give below.  He actually found real-life examples (though I suspect that maybe SB was in there, or park factors, or something).  It’s for this reason that I want OPS to die a quick death among serious analysts (as well as its offshoot, the less obscene OPS+).  It can survive for quick things.  Anway here’s my answer:

Read More

Tuesday, June 23, 2009

UZR in MSM

This time, it’s the Star-Ledger in NJ:

UZR is broken into different components that account for errors and other factors. In Teixeira’s case, his dropoff is tied almost entirely to his decline in the portion of UZR that accounts specifically for range. In 2008, Teixeira saved 9.4 runs just on his range. This season, he’s cost his team runs, posting a -3.6.

That said, even advanced defensive statistics such as UZR have their blind spots. And when it comes to first basemen, and Teixeira specifically, the statistic doesn’t measure one of the most important areas of contribution: scooping balls and tracking wayward throws.

It bothers me to no end that Fangraphs and other sites report numbers to one decimal place.  It implies a level of precision that simply does not exist.  If a UZR of +4.6 really means +4.6 +/- 2.1 (or whatever), then I’d show it as +5 and be done with it.

Canadian banking system

Non-sports post:

Read More

Flip Flop Fly ball

Cool graphics site.

Hat tip: Craig.

NHL draft analysis and spreadsheet 1994-2009

I love the internet.

Barry Code

Hat tip: Patriot as he highlights a site dedicated to Barry Codell, inventor of what has been popularized as Total Average.  This is what baseball analysis would look like if Linear Weights didn’t take hold.  And here is their glossary of what can essentially be described as “let’s add and divide numbers”. 

Codell paved the way.  This site has gone to great lengths to make a religion out of it.  Pluto is also a dwarf planet, unless you also accept Eris as the tenth planet.  You can’t stick with just nine planets.

Monday, June 22, 2009

Donald Fehr

Shysterball is talking about Fehr:

no mention of the fact that a rookie made $60,000 a year when he took over and that the game’s biggest stars are now making ten times what they made back then.

Actually, no mention from Craig if we compare the minimum salary in MLB to the NHL, a team with half the revenue of MLB, and who had to contend with Alan Eagleson, a convicted criminal for his shenanigans as head of the NHLPA:

The minimum NHL player salary in 2005-06 and 2006-07 will be $450,000; $475,000 in 2007-08 and 2008-09; $500,000 in 2009-10 and 2010-11, and $525,000 in 2011-12

That is just frankly embarrassing for MLBPA that it didn’t negotiate at least 1MM a year as the minimum salary for its players.  Not to mention that players in their non-free agency years make 30 cents on the free agent dollar, a rate far far lower than what NHL players get.

What is clear is that the MLBPA has given the rights of its non-free agents away to artificially increase the value of its free agents, all on the likely idea that if one boat rises, everyone else will gain.  Of course, seeing that the percentage of revenue that goes to the players is no different than the other leagues, that too is not a good idea.

Finally, since PED is a workplace issue that Fehr recasted as something to bargain away, Fehr fails there as well.

***

Michael Jackson used to be a great entertainer and his time has passed in a very unkind way.

WAR Lords

Some helpful WAR 101 for you guys in these two posts.

I love the “WAR Lords” blog title as well.  I’m going to have to use that if I ever get to putting out leaderboards.

And Justin chimes in as well.

StatSpeak

Dan is putting a callout:

With that, I am putting out a call to the readers of Statistically Speaking, asking for your help. We would all like to keep this blog going full time, and it will be a difficult task to do once Matt undoubtedly wins BP Idol. So if you are interested in writing for StatSpeak, send me and email at .

He also said that Pizza was leaving, which means it’s as bad a news for us as it is good for whoever gets his time.  And presuming he’s making a rational decision, even better for the recipients.

Contract Buyouts

Sherman (hat tip Primer):

What we are suggesting the Commissioner’s Office do—to save the trade deadline for gossip junkies like me—is to offer a one-time, one-contract reprieve for all 30 teams. Remember the Allan Houston Rule from 2005 when the NBA allowed every team to release one player so that the player’s contract no longer counted toward the luxury tax. The rule, however, did not remove teams from having to pay the contract, just any luxury-tax obligations.

The NHL lets you buy out two-thirds of the remaining contract.  This has been exercised once or twice a year.  The players have usually ended up at around break-even on the deal.  It usually looks something like this:  a player is owed 30MM over 4 years.  The team thinks he stinks.  They pay him off at 20MM.  He then signs a free agent deal for 10MM over 4 years.  The player gets his money.  The team that loses him get to have his contract reduced on the books because they get to spread the buyout over 8 years for cap purposes.  The team getting the player pays fair value for the player.  Everyone wins.

I’d be in favor of a buyout at 50 cents on the dollar.  This will mean that some players may lose out on the deal.  If you make out the buyout too low, say 10 cents on the dollar (a “restocking fee"), then some teams get to get out of really bad contracts and those players will lose out.  Not only that, but giving a team this much leeway means they would take alot more chance at big contracts.

Saturday, June 20, 2009

Video of the week

Loss aversion in golf

Hat tip: King

Using data the tour regularly records on every ball’s green location accurate to the nearest inch, the professors found that birdie putts were made about 3 percent less often than otherwise identical putts for par. (In effect, players tell themselves before birdie attempts, “Let’s just get close,” rather than, “I have to make this.”)

Friday, June 19, 2009

Why things are not always exactly as they seem and an amusing broadcasting comment…

First the broadcasting comment:

I was watching the ATL/PIT 16 inning (or however long it was) game the other day.  Late in the game (extra innings), every time the HP umpire called a marginal strike, one of the announcers (I forget who) would say something like (you often hear the same thing in a similar circumstance):

“Well, they (the umpires) have been out here for 6 hours already.  You’d better be swinging the bat.” They are implying of course, that the umpires want to go home, and understandably so.

What is (glaringly) wrong with that statement and logic?

OK, things that are not always exactly as they seem…

In a couple of pitch f/x articles, it has been shown that the high inside fastball or just the inside fastball is a very good pitch, if you throw hard that is.  In fact, the more speed on the pitch, the better the inside pitch is.  Even an inside breaking pitch is a good pitch I think (according to the pitch f/x data), probably because it surprises the batter, especially the same-side batter who is likely bailing out unless he is definitely expecting that exact pitch.  Anyway, with regard to the inside fastball, you also often hear commentators wax about how pitchers don’t like to pitch inside anymore or that so-and-so does not like to pitch inside (and that it is a bad thing).

My second question (this is like a dual question of the day) is:

Even though the inside or high inside pitch may be a very good pitch when you look at it through the lens of the pitch f/x data, why might it be a bad pitch in reality?  And let’s forget about the fact that most, or at least some, pitchers have to mix up their locations in order to keep the batter from keying in on a certain location or locations.

hSOB (Horizontal Speed off the bat)

Someone asked Dave Allen:

Is that a more accurate measure than just taking the average overall speed, or just a preference?

And he responded:

I guess it is preference and depends on what you are trying to say. In this case I wanted some proxy for how ‘hard’ the balls are hit and I think that horizontal speed is a better measure of that than overall speed.

Additionally horizontal speed does seem to be a better predictor of success of a ball in play. To see that you can plot the run value of a ball in play against the overall speed (like you did at your blog) and horizontal speed and then fit a line to both. In both cases you get a tiny p-value so both overall speed and horizontal speed predict success, but the r-squared for the horizontal speed is a little bit higher than for the overall speed. I get about r^2 = 0.136 versus r^2 = 0.103. So the horizontal speed explains another 3% more of the variation in run value outcome than the overall speed.

Let’s say that our objective is to figure out how hard a ball is hit.  Not just to figure out how hard in its most pure form, but how hard as in a ball that travels far, fast, and is tough to field.  The horizontal speed off the bat (hSOB) seems to give us what we want.  That is, we are measuring the speed of the ball off the bat by only looking at the horizontal component (vertical launch angle of 0 degrees).  After all, we want the ball to move forward.  But, we also want some, not alot, of loft.  And, as has been pointed out by a few PITCHf/x-ers now, the ideal vertical launch angle is around 11 degrees. 

Therefore, shouldn’t we measure the speed off the bat at 11 degrees, rather than 0 degrees?  Furthermore, some power hitters, say Ryan Howard, have the optimum launch angle at say 25 degrees.

In other words, how hard you hit the ball is really another way of saying “how much of that ball did he get” (with the understanding that you don’t want to hit the ball flush on, but a tiny bit below the center point).  And, that amount is entirely determined based on the plane of his swing.

So, while hSOB gives us part of what we want, if the question is “how hard did he hit the ball”, shouldn’t we solve that based on the plane of the hitter’s swing (which we can possibly infer based on the average of all his trajectories, or possibly based on the average of all his hits)?

Am I thinking about this wrong?

Mark Howe on dad Gordie Howe

Cool dad’s day story:

Another rule was to make their own way in the world—don’t trade on Dad’s fame. Murray told NHL.com last year how his mother embarrassed him, at 10 years old, when he tried to get her to confirm to a bunch of strangers that he really was Gordie’s son, as he had been bragging.

“Who are you? Get away from me,” Colleen Howe responded.

“Mom and Dad never had to say a lot,” Mark said. “Dad was about example. He had a million opportunities to say, ‘I’m Gordie Howe,’ and it would have opened doors for him, but that was never the case. He just feels like he’s an ordinary person like everyone else. We learned to be humble and we learned that just because our name was Howe, it didn’t mean diddly-squat.”

Mark Howe remembered one of the sternest lessons about hockey his father ever gave him. It was delivered in an easy-going, friendly manner, but the advice was worth millions.

“We were playing together in Houston and I was young. I had been out with the guys the night before, had a few drinks and I was hurting that morning,” he said. “Dad grabbed me and said, ‘Look how these guys (don’t) take care of themselves. At 31 they’ll be done. If you want to play until you’re 30 or 40, what you do now will determine your future.’

“His dedication to the game was unmatched. I can’t believe the conditions I watched him play through. He was hospitalized before our first Houston game with back spasms, in traction, and played that night. I watched him pass kidney stones at 9 a.m and play in a 4 p.m. game.

“He is, without a doubt, the toughest man I ever met in my life.”

The Poz goes FJM on Harold Reynolds’ a$$ - gather around the kids

Poz is at it again.  For all you aspiring takedown artists, this is how it’s done:

8. It’s not a good thing. Well, maybe it can be a good thing, if the power hitter is on a good team. But it’s not a good thing if you have a bad team. Take Adrian Gonzalez. Please! His OPS is going to be high because he hits a lot of home runs and walks a lot. This is because pitchers don’t pitch to him. Yes, when he hits the home runs, they pitch to him, but the other times they don’t. I appear to be confusing OPS with on-base percentage but I’m rolling now, so please don’t stop me.