Lineup by OBP

DMXRoid — Sat, 31 May 2008 23:46:23 -0000

We can all agree that the Indians have had trouble coming up with a working lineup, right? I'm hardly the first person to think about it, or write it, and I mean, come on, when Jamey Blake Carroll has the plurality of AB's out of the 2 hole (heh), there has to be a problem.

I've seen a couple other fanposts with their own modest proposals for how to arrange the lineup, and I didn't want to be left out, so here's mine:

Organize the lineup by OBP, from top to bottom. That is, the hitter with the highest OBP on the team hits leadoff, the second highest hits in the two hole, etc....

The Indians don't have a collection of players that fill out a traditionally arranged lineup card. Other than Grady, who's a fine traditional leadoff hitter, nobody else's placement in the batting order makes any particular sense. Is Jhonny a 5 hitter? Why? Because he'll occasionally go through a hot streak and pop off some home runs? Can anyone provide a good reason to hit any Tribe player in any particular slot?

Probably not. And since the Tribe is struggling mightily to score runs, there's even less reason to be constrained by traditional thinking about lineups. The concern is to get runners to cross the plate.

Obviously, there's no way to guarantee that a run scores in any scenario, so what you want to do is maximize the chance that you'll score a run. What detracts from scoring runs? Making outs. Who's good at not making outs? Players with a high OBP. A player's OBP is also indicative of the probability that he'll reach base. If a player has a .400 OBP, you know that there's a pretty good chance that he'll get on base about 40% of the time. Of course, that's not a rule, obviously, and there are streaks and slumps, but I think you can adjust for that by doing the lineup based on OBP over the past 10 games or so.

It comes down to this: you have a better chance of scoring a run by sending up a guy with a .400 OPB followed by a .380 OBP than you do by follownig with a guy with a .350 OBP.

Now, one obvious problem with my idea is that it only maximizes the possibility of scoring ONE run, as opposed to maximizing the total runs scored in a given scenario. A better way to organize your lineup is probably by OPS, since that combines your ability to not make an out with the number of bases you're likely to pick up in the course of not making that out. However, I'm not sure where the sweet spot is, where it becomes worth it to increase the risk of an out for the reward of an increase in the output. If anyone else has an idea, that'd be excellent.

Please poke holes in this idea, I'm curious as to how well it holds up to scrutiny.

It really is Wedge

DMXRoid — Fri, 30 May 2008 06:42:26 -0000

Evaluating a manager's performance is really really hard to do in quantitative terms. Sure, we can all get pissed at calls that were or weren't made after the fact, but it's difficult to translate that into what a manager brings or takes away from his team in terms of wins or losses. Plus, we all remember when the manager screws up, but nobody says "Oh man, he totally called for the lefty specialist at JUST the right moment last night, it was AWESOME".

I think that one way to solve this problem is to look at the Pythagorean differential of a manager over a period of time, and compare it to the historical average. To be sure, this isn't perfect, as there are a lot of factors that go into the differential, many of them simply involving luck, but it's better than nothing, and it makes my point for me.

The average Pythagorean differential since the modern baseball era began in 1901, using baseball-reference.com's formula with an exponent of 1.83, not the traditional exponent of 2, is -0.404. That means that over the past 100 years, teams won about a half game less per year than their hitting and pitching would make you think.

What I'm interested in is the average Pythagorean differential that follows a manager around. For the 179 managers who have at least as much time in the job as Wedge (5 years. Man, that's a loooong 5 years), the average differential is -0.553, so still about a half game.

Here's a quick rundown of the top and bottom 12 managers by this stat:

Top 12

Seasons Py. Diff Manager

6                    3.1667 Ron Gardenhire
8                   2.8750     Lum     Harris
5                   2.8000     Ossie     Bluege
8                     2.6250     Dick     Howser
6                     2.1667     Dave     Garcia
7                     2.1429     Pete     Rose
5                     2.0000     Jim     Frey
7                     2.0000     Herman     Franks
7                     1.8571     George     Gibson
7                     1.7143     Frank     Lucchesi
15                   1.6667     Jack     McKeon
5                    1.6000     Bob     Melvin

Bottom 12

Seasons Py. Diff Manager

5               -3.0000     Eric     Wedge
10                -3.1000     Fielder Jones
7                  -3.1429     Preston Gomez
6                  -3.1667     Tony     Muser
5                  -3.2000     Russ     Nixon
7                  -3.4286     Hugh     Duffy
9                  -3.4444     Buddy     Bell
6                  -3.5000     Johnny     Keane
7                  -3.8571     Bobby     Bragan
6                  -4.1667     Ned     Hanlon
9                  -5.1111     Jimmie     Wilson
5                  -5.6000     Joe     Kelley

That's right, Eric Wedge is the 12th worst manager _ever_ by this very very very limited, subjected, and probably ultimately useless metric. Dusty Baker and Bobby V beat him out. Wedge's teams are under-performers of historic proportion, a full 2.5 runs behind where they should be every year.

Now, that 2.5 games would have only made _any_ difference last year, when it would have pushed us up over the Red Sox, but the Tribe was 5 games over expectation last year anyway. Here are the differentials for the Indians since 2003:

2003: -5

2004: -1

2005: -3

2006: -11 (!!!!)

2007: +5

It's immediately pretty obvious that 2006's -11 performance is probably the main cause of Wedge's place so low on the list. Remove that year, and his average gets bumped up to -1, barely a half run worse than the average for managers with as much time in as he had. It's also worth noting that none of the negative years hurt the Tribe's playoff chances in any way. In '03 and '04, the Indians weren't anywhere close to the playoffs, in '05, 3 games would have put them at 96 wins, still 3 less than Chicago's 99, and in '06, even adding 11 wins to the 78 the Tribe won wouldn't have been good enough to crack 3rd place, with 3 other teams in the division picking up 90+ wins on the season.

OK, that's the part where I'm fair and a little objective about it. This is the part where I use shoddy numbers to back up my frustration with this team right now.

As I said before, managers don't have a lot of impact on the outcome of a game. It's decisions at the edges, with pretty small differences in probabilities, things like when to make a hitter bunt (unless his name is Andy Marte, then there's a stone rule that he must bunt on all counts in all situations), when to steal, when to go to the bullpen, how to use guys in a platoon, etc...

And yet, as small of an impact that those factors have, they're where Wedge is _really_ failing to perform. The series against the Reds is a perfect example. Wedge got completely wrecked that entire series by one of the worst managers in baseball, Dusty Baker. He was completely unable to adapt to not having the DH, over thought situations like his absolutely unnecessary double switch, added to his weird obsession with making Marte bunt to get to Droob, his increasing tendency to steal when the run expectation chart opposes it, etc...

There's also the impact a manager has by winding his team up so tight they can't think straight, but that's ground that lots of other people have covered much better than I could at 4:30 AM. My point was mostly to kick around the idea that Pythagorean differential can tell us something about how good a manager is.

Thoughts are appreciated.

SB Nation User Blog: DMXRoid

Lineup by OBP

It really is Wedge