SB Nation User Blog: Aaron Fix

The Twins and Bill James' Pythagorean Theorem

Aaron Fix — Mon, 02 Jun 2008 05:21:02 -0000

Joe Posnanski thinks that Ron Gardenhire is the best manager in baseball. He gives a few reasons for thinking this, and all of those reasons are pretty easily refuted, but one in particular got me to thinking. He mentions that the Twins under Ron Gardenhire have consistently outperformed their Pythagorean record, as given by Bill James' Pythagorean formula for expected winning percentage. He is absolutely right. Over 6 full seasons, Ron Gardenhire's teams have won, on average, three more games than James' formula would suggest. Before we read too much into this trend, let's think a bit more about this formula.

The following observations about this formula are important:

1. There is no theoretical reason why expected winning percentage should be equal to RS^2 / (RS^2 + RA^2).

2. The formula will have downward biased estimates for teams which score very few runs (e.g. the Twins).

3. The formula is known to perform poorly in the presence of very good bullpens, and is also known to be biased towards the mean (so that really good teams will have lower predicted winning percentages, while really bad teams will have higher winning percentages as predicted by the formula).

By (1) I mean that this is simply a statistic that James stumbled upon which fit the data reasonably well. This is not a description of the data generating process. As an example of (2), consider a team which scores 800 runs and gives up zero. This team's expected winning percentage, according to the Bill James' formula, will be .500. That's a pretty poor prediction. I do not know the particular reasons for (3), but they are known issues, and are especially relevant to Ron Gardenhire's tenure with the Twins since the ball club actually won fewer games than predicted in 2007 and 2005, the two losing seasons.

So is beating Bill James an indicator of managerial excellence? It doesn't look that way. If you examine the performance of the Twins versus their "expected" performance over the past 20 years, you will notice some autocorrelation (meaning that beating the formula one year makes the team more likely to beat it the next year). This probably has a lot to do with the fact that a team with a strong bullpen one year tends to have a strong on the next year. You'll also notice a lot of negative differentials during losing seasons, and positive differentials (meaning that the team won more than expected) during winning seasons (an example of the formula being biased towards the mean). And you'll notice that the differential is largest in 2002 when the Twins won 94 games and the division despite scoring only 768 runs, the fewest they have scored while winning the division during Ron Gardenhire's tenure (an example of the formula being biased against teams that score very few runs while giving up even fewer).

This formula is, in my opinion, quite useless anyways since there is no theoretical justification for its use. If a team does not live up to their Pythagorean expectation, then this is probably because of a problem with the formula and does not necessarily mean the team's luck is about to turn. In the case of the Twins, the formula may be worse than useless. It is actually quite misleading. Unless the Twins start to score and give up more runs (which I don't see happening given all of the quality young pitching and anemic young hitting), or see their bullpen become mediocre (despite losing Neshek, I think we will still have an above average bullpen this year due to Nathan, depth, Rick Anderson and the fact that we have a long line of potential major league starters in the minors right now) then don't expect the Twins to live up to their Pythagorean expectation.

As of tonight the Twins are 29-27, with 261 runs scored and 269 runs against. The next time someone says "but their Pythagorean record is actually 27-29", tell them to go to hell.

PS: As a side note, Posnanski also mentioned that under Gardy the Twins have won 56% of their one-run ball games. This is not very impressive considering that during all 6 years the bullpen has been pretty lights out, and the Twins have won 55% of their games overall. I need to think about this issue a lot more, but I think that the strong bullpen might mean that they actually should have won more than 56% of those games.

Could the Twins add The Big Hurt?

Aaron Fix — Sun, 20 Apr 2008 23:03:08 -0000

No, and that is really depressing. It's not because we couldn't afford him, we certainly could, and his big right-handed bat would fit really nice right there between Mauer and Morneau. Do you think the Twins might hit a few more right-handed homers with him batting in the middle of the order?

Signing Thomas would mean releasing Monroe, which is already a good idea even without Thomas on the team. But Thomas wants to play more than once every 3 days against lefties, and the Twins already have a good young left-handed DH. And unfortunately, the Twins have given themselves a glut of half decent to pretty good major league outfielders so it would be hard to get Kubel into the field on a regular basis to let Thomas DH. Unfortunately, the holes in the Twins lineup are at the left side of the infield. You could argue that all the outfield spots are holes, but the Twins have committed themselves to a few young guys with a lot of upside, and there is no reason to mess with their playing time.

But when I go to bed tonight you can be I'll be dreaming about the Big Hurt crushing long balls off of Mark Buerhle in Chicago.

Can anyone think of a way the Twins could make this work? I sure can't.

Walks and the Minnesota Twins

Aaron Fix — Sun, 13 Apr 2008 15:45:50 -0000

It is no secret that the Minnesota Twins pitchers don't walk batters. If there is a defining characteristic about the team besides a glut of weak hitting middle infielders, it is that Twins pitchers throw strikes, challenge hitters and don't give many free passes. It is also true that Twins pitchers, up and down the system, are quite good. Obviously there is a lot of correlation (and causation) between not giving up walks and pitching well, and the organization understands this. From top to bottom, the Minnesota Twins as an organization seem to understand that giving away bases for free is one of the worst things a pitcher can do for himself and his team. And the Twins are right. The Team is able to compete year after year, despite generally poor (or at best average) offense. As a dedicated fan, this is a point of pride for me. I like watching 22 year olds come up through the system and challenge guys like Vladimir Guerrero on the mound.

What I find very interesting is that the Twins place no such premium on taking walks. During spring training Ron Gardenhire held privately and publicly encouraged Delmon Young and Carlos Gomez (two of the freest swingers I have ever laid eyes on) to keep hacking. I read in one of the Minnesota dailies that Delmon Young swung at more pitches than any other hitter in baseball last year. Carlos Gomez, for his part, swung through three straight pitches that were nowhere near the plate with the bases loaded and either 0 or 1 out last night. No one in the organization, at least publicly, seems to acknowledge the logical corollary to what they understand at a very fundamental level for their pitchers: walks help hitters.

In support of my thesis, I took a look at the team data for the last few seasons and the results are striking. In 2006, when the offense was actually decent, the Twins were 22nd out of 30 teams in BB. The next year they were 19th, and so far in 2008 (through 10 games, the stats were not updated since last night) the Twins are 26th in the majors with just 29 walks, roughly half of Cincinnati's 57 and Oakland's 56. As for pitching, the Twins have issued just 14 walks in 11 games in 2008, less than half as many as any other team in the league and exactly one quarter as many as tthe league-leading Boston Red Sox! In 2007, the Twins issued the 2nd fewest walks in the league (the Indians issued the fewest and went to the ALCS).

I am left to conclude that there is an enormous logical inconsistency at the organizational level. Granted the Twins order is generally not too threatening and thus pitchers may feel more comfortable challenging the hitters, but it is not at all clear where the causation lies here. In a lineup full of free swingers, hitters get to see fewer middle relievers and get to hit with men on base more rarely. Moreover, becoming more patient at the plate can make a guy a better hitter by forcing pitchers to give him something he can hit. I am more tempted to conclude that the ability to take a pitch makes a hitter better, rather than that the Twins don't walk often because they are not threatening hitters.

Has anyone else been bothered by this? How can an organization that recognizes that issuing walks helps the other team completely fail to recognize that taking walks helps our team?

Why start Monroe?

Aaron Fix — Tue, 01 Apr 2008 04:21:59 -0000

There were a lot of positive about last night's game (the first few were all Carlos Gomez) and I'll let other people post about those and then participate in the comments. I want to submit an open question and then give my best answer (and hopefully others will comment).

Why did Ron Gardenhire start Craig Monroe, clearly the worse of the two options for DH, also clearly the most right-handed of the two options for DH, instead of Jason Kubel?

The easy answer is that he wanted to give a nod to the veteran and show he believes in him after Monroe suffered through a terrible 2007. If so, then fine. The club still won and Monroe probably appreciated the start.

My belief is that there is more to it than that. I believe that what is at work is the same perverse risk aversion that has led Gardy to carry 3 catchers on the major league roster.

There is no question that starting Jason Kubel, who is left-handed and a better hitter, gives the team a better chance of winning against a right-handed starter. But it removes an option for Ron Gardenhire late in the game. Having a left-handed power bat on the bench in the late innings is just good baseball wisdom, and Ron Gardenhire wants that option. If Jason Kubel starts, he gets more at bats and thus more chances to help the team win, but if Ron Gardenhire wants a left-handed pinch hitter, he doesn't have a good option. Since he can't stand the idea of not getting to exercise his conventional baseball wisdom, he sacrifices what is obviously better for the team to give himself the ability to make the type of decision late in the game that he likes to make .

This is a similar sort of situation to his catcher fetish. The chances of a Mauer pinch hit followed by an injury to Redmond are so incredibly low that they will probably never cost the Twins a game. But Gardy's risk aversion gets the better of him, and he uses a bench spot on a guy who will never play a major league inning instead of a deserving hitter. The outcome is the same; it is worse for the team, but allows Gardy to make the types of decisions that he is comfortable making.

Different Managing Styles

Aaron Fix — Wed, 26 Mar 2008 19:07:21 -0000

Near the end of June 2007, the Boston Red Sox called up CF Jacoby Ellsbury from their AAA affiliate. He shuffled back and forth between AAA and MLB before making the playoff roster and eventually helping the team win a world series. In 127 regular season plate appearances, he hit .353/.394/.509. In the playoffs, he managed .360/.429/.520 in 28 plate appearances. He is young talented and full of potential, and has shown that he can deliver on the biggest stage, but has struggled a little in spring training and is still only 24 years old.

Here's Red Sox Manager Terry Francona on his decision to bat Jacoby Ellsbury low in the order:

"You put a young player a little bit at risk when you put him right at the top of the lineup early on," Boston manager Terry Francona said. "I know what he hit last year. It was impressive, but you're still trying to develop guys."

In the middle of May 2007, the New York Mets called up CF Carlos Gomez to the major league roster. He more or less stuck, but suffered a serious injury to his wrist in early July that kept him out of the majors for two months. He stuck despite hitting .232/.288/.304. He is young, talented and full of potential but has yet to perform on the biggest stage, and had his first and only season in the majors interrupted by an injury that has been known to decrease a player's ability to hit for power. Baseball Prospectus says that due to the Mets' carelessness in bringing him up too early and his injury, 2007 was a complete waste in terms of his development. This spring, his hitting coach expressed concern that he was swinging so hard that he could easily hurt his back (his manager encouraged him to swing harder). He started out abysmally this spring before getting hot late (though he still has a lower OBP than either of the guys he beat out for the starting job in CF). He is only 22 years old.

Here's Twins Manager Ron Gardenhire on his decision to bat Carlos Gomez leadoff:

"Gomez real good... (grunting)... me like Carlos Gomez... (beating chest)... he run hard like Nick Punto. Also play bad like Nick Punto. If him not hit leadoff, Nick Punto. What on base percentage?"

Gardenhire and Francona each have their own method of nurturing their young talent. Who is right? I guess all we can do is sit back and see which player has the better year.

http://noblingblings.blogspot.com/

What is Ron Gardenhire Talking about?

Aaron Fix — Thu, 06 Mar 2008 22:18:13 -0000

"Kubel proved he can hit the second half [of 2007]," Gardenhire said. "I've decided that with him what you see is what you're going to get. He's going to be that speed. And when he turns it up a notch, we don't notice.

"It's not that much of a notch."

That is from a story by Patrick Reusse, inexplicably titled "Young has look of a force for Twins". I have been racking my brains for answers as to what Gardy could possibly mean, and I have come up with the following possible interpretations (translations?) of this quote:

...

http://noblingblings.blogspot.com/

"It's not that much of a notch."

Kubel is a good hitter. He is not very fast. When he runs a little faster, I can't tell the difference. He can't really run that much faster.
Kubel hit really well in the 2nd half of the year, which is when I saw him hitting well. His bat is going to travel at that speed (the speed it traveled in the second half of 2007). When his bat goes faster, it looks the same to me, perhaps because it is not actually going that much faster.
Kubel hit well in the second half last season, and since I saw that I expect him to do the same this season. However, I have also seen him struggle in Spring Training. What you see is what you get. Also, he is going to drive his car about as fast as he did last year. When he turns the radio or heat or AC up a notch I don't notice, perhaps because I am getting a bit senile and this also has the side effect of making me not notice when my starters aren't actually playing that well.
With Jason, I have decided that what I see him do is actually what he does, unlike many other players who do completely different things than what I see them do. Take Nick Punto, for instance. I see him hitting .310 with power and driving in a lot of runs, when in fact he actually sucks at at hitting. Maybe I should look at the box scores and if I did I could see that Punto is 0 for 4 for the third straight night and is in a 3 for 47 slump, but since what I see is an 8 game hitting streak, I keep putting his name in the lineup, sometimes near the top of the order.
When Jason Kubel stands in front of a mirror, he can see his reflection. His image travels at the speed of light. I cannot tell the difference between light speed and light speed plus epsilon.
Kubel hits well, and when he hits even better, we (the coaching staff) don't notice.
I'm old. My knees ache and Jason Kubel is a Russian submarine. Sometimes he speeds up and sometimes he slows down. Do you think he will take me to the dentist? I have a banana in my arm.

Whatever he is thinking, I am glad Patrick Reusse decided to close his article with such words of wisdom (The article, once again, is titled "Young has look of a force for Twins" and it refers to Delmon Young, not Jason Young Kubel or anything like that).

ZiPS: Twins Win Division!

Aaron Fix — Wed, 01 Feb 2006 03:57:35 -0000

http://yankeefan.blogspot.com/2006/01/its-early-projection-time.html

I don't have this Diamond Mind Baseball software, and I don't plan on playing any computer games to simulate an entire baseball season, but thankfully someone else has and we get to look at it. Apparently the ZiPS projection favors younger players and handicaps older players, and the guy running the projections has to choose the starting 9 and the 5 man rotation for all 30 teams, so there is some uncertainty, but I found this interesting nonetheless. If anyone is familiar with the software perhaps they would like to shed some light on the subject.

What's noticeable is that according to this projection the Red Sox are noticeably better than the Yankees, the Twins somehow win the central with 86 wins and the White Sox tie with Detroit for third in the division. There are other surprises all over the place (I don't think the Cubs or Cardinals are going to be that good, I don't think Houston is going to be that bad). Has anyone ever ran these projections, and what sort of accuracy did you find? What can affect the results? Any insight would be appreciated.