THE BOOK--Playing The Percentages In Baseball

BTW, if I run 10,0000 seasons in my season simulator, the Dodgers make the post-season 26.3% of the time “with” Pierre and 29.1% of the time “without” Pierre.  I put “with and without” in quotes because all I did was change the Dodgers win percentage for the remainder of the season to reflect the .12 difference per game between Pierre and Ethier and then assume he plays either 40 games (with) or 10 games (without).

That is a fairly substantial difference I think, and more than I thought, so I stand corrected.

The other problem is that in Sheehan’s set of data, most of Pierre’s games leading off come without Furcal in the lineup and instead a replacement level player at SS. I think that’s a larger problem. The data that he’s using for comparison isn’t really equal in conditions. Talk about cherry picking to make a point.

Pierre is .12 runs per game worse than the average of Ethier/Kemp?  That is .12*162 = 19 runs per season, which is enormous.  Or, I should say ENORMOUS!  I’m glad MGL corrected himself.

A player like that can only exist as a platoon player, PR, or late-inning fielding replacement.

I will also disagree with MGL on the batting order.  While it won’t matter much in most cases, when you are talking about the leadoff hitter, it will definitely make a big difference.  Moving Pierre from leadoff to 8th will save, as a pure guess, at least 5 runs, if not even more.  If you say 10 runs, I won’t even be surprised.

When we say it doesn’t matter much, it’s when we talk about moving a guy 3rd to 4th or 2nd to 5th or something like that.  Something reasonable.  Should Ichiro be leadoff or 2nd?  Sh!t like that. But leadoff to 8th?  No.  It matters then.

Yes, 19 runs is a lot, but we are talking about the impact of him playing 40 or 10 games during the rest of the season. That is the argument I am addressing, not whether Pierre or Ethier should be in the lineup, period.  19 runs per 162 is 3.5 runs in 30 games.  While nothing to scoff at, that is not going to make that much of a difference to the Dodgers from here on in. Of course that depends on your definition of “much of a difference” I guess.

Tango, the reason I say it does not matter much (Pierre’s slot in the BO) is that, one, Pierre is not THAT bad of a hitter, and two, there is some benefit to having that kind of speed at the top of the order regardless of how often he does or does not get on base.

I’ll leave that paragraph above, no matter what happens, and I am off to run some sims, with him batting leadoff and then 8th.  It might make a big difference whether Furcal is in the lineup or not, as with Furcal, Pierre is the worst hitter on the team and without Furcal, the SS is.  Not that my sim is the end all of course.

...sound of computer running…

With Pierre batting leadoff, they score 4.235 rpg in 100,000 games.

BO is:

Pierre
Ethier
Martin
Manny
Loney
Kent
Blake
Berroa
Pitcher

If I switch Blake and Pierre, they score 4.208 rpg, which is worse by 4 runs per 162!

If I put Furcal back in the lineup and Pierre bats 1 and Furcal 8, they score 4.378.  If I switch them, I get 4.424, a difference of 7.5 runs.  But that is switching Furcal, a good hitter, from 8th to 1st.

One of the “problems” in isolating the effect of moving one player in the BO, is that we have to move at least 2 of course, and we are not sure who the effect is coming from.  I suspect that most if not all of those 7.5 runs is moving Furcal up.

How about if we bat Pierre leadoff, Ethier 2, Furcal 3, and move Martin to 7 and Blake to 8?

4.428.

I don’t see any evidence that batting Pierre leadoff is problematic, at least according to my sim.

Switching a bad hitter (but OBP-heavy) for an average hitter (who likely isn’t strong in OBP in your forecast) is not what I’m really talking about.

The Furcal thing is much more interesting.  I presume you have Furcal as an even, or worse, hitter than Blake.  But, Furcal is heavy on OBP.

So, switching Pierre for Furcal gives you a 7.5 run difference which is no big surprise.

Can you post your forecasts for your Dodger players, in this format:
http://www.insidethebook.com/ee/index.php/site/comments/optimizing_the_batting_order/

Then, I’ll run it through my non-sim process, and see what I come up with as an optimal order.

By the way, I don’t go for the argument of only 40 games or 10 games or even one game.

If someone is 20 runs per 162G worse than someone else (say like Willie Bloomquist and the average player), are you going to say that it’s no big deal to make Willie a regular because there’s only 6 weeks left to the season?

Or how about sitting a 3.50 ERA pitcher in favor of a 4.50 ERA pitcher, because it’s only 7 starts?

Sorry, but I don’t buy it, if for no other reason than it shows a pattern of bad behaviour.

Sorry, but I don’t buy it, if for no other reason than it shows a pattern of bad behaviour.

Tango, that is a different story.  I am only responding to the ubiquitous, “Playing Pierre could mean the difference between making the post-season or not” (of course it COULD) argument.  That’s all.

Of course it is a very bad move, especially when you consider how much worse it is than 10 bad sac bunts or IBB’s or something like that.

I am merely trying to quantify the difference between Pierre playing 10 games or 40 games.  That’s all.  People can make anything they want out of it.  I just don’t want people to be making judgments based on the crap that Sheehan puts out there (without Pierre, they score 50% more runs, etc.).

Tango, I just don’t think a Markov lineup simulator does very well, especially when baserunning comes into the picture.  As I said in another thread, I am even having a hard time finding an advantage to the pitcher hitting 8th.  And it seems to depend a lot on exactly who the other players are, as it should. 

Here are the projections for Dodger players for a league where the averages are:

Per 500 PA, not including IBB and SH:

s 80, d 24.5 t 2.5 hr 13.5 bb+hp 45 so 80

OBA is scaled to a league average of .338 and SA is scaled to a league average of .433.

The last number is lwts per 630 PA (around 150 games)

Ethier 84.1, 28.2, 3.3, 12.8, 43.2, 71.2, 364, 460, +7.1

Jones 59, 20.6, 1.4, 22.6, 54, 98.4, 336, 445, -.1

Berroa 86.7, 20.8, 2.2, 8.8, 30.6, 81.6, 320, 378, -20.9

Loney 89.7, 27.9, 4.8, 11.5, 36, 63.2, 361, 462, 6.4

Kent 76.9, 27, 1.3, 13.5, 45.9, 64.8, 352, 433, 1

Pierre 113.4, 18.4, 5.25, .4, 33.3, 28.8, 363, 374, -10.9

Kemp 92.5, 27.7, 5.0, 15.1, 32.4, 100.8, 368, 498, 14.4

Furcal 93.4, 23.3, 4.9, 8.37, 46.8, 56, 375, 432, 4.1

Martin 87.2, 25.5, 1.5, 11.5, 59, 57.6, 389, 446, 11.3

If this were another team, the issue would have a different flavor. Since it’s the Dodgers, though, the Kemp-Ethier/Jones-Pierre discussion is fraught with implications beyond “who’s better?” evaluations. By playing Pierre and Jones, the Dodgers continue their recent practice of telling young players that experience for experience’s sake will be valued as much as their talent. Even after they were ready: Loney sat. Kemp sat. Ethier sat. Aybar got traded. LaRoche got traded. DeWitt sits. Nomar played. Luis Gonzalez played. Pierre was signed and played. Jones was signed and played. Casey Blake was traded for and played. Danny Ardoin played catcher while Martin moved to third!

It seems to me the Dodgers don’t want to play anyone who even has a whiff of being “unproven”, and will always prefer to play known qualities over uncertainties. So telling Kemp/Ethier that one of them will sit because Juan Pierre has played for seven years prior, even though it’s consensus that both Kemp/Ethier are better players, tells everyone in the organization that it’s not a meritocracy. Is that what you want for your org?

I think that the issue is overblown to say the least, considering that Pierre has started one game since last Monday (in only one game in the last 6 was there a lefty pitcher).

Isn’t part of the issue the fact the Dodgers have signed Juan Pierre to a long contract at like 10 milion a year or thereabouts.

By using Juan Pierre as a “scrub” the manager implicitly makes the general manager look like an idiot. Perhaps not the smartest move from a standpoint of job security.

I suppose that is an issue for managers.  I don’t know.  I don’t know how much discussion goes on between managers and GMs as far as playing time goes.  It probably varies a lot among teams.

But, as I said, he is not playing much anyway, so this is much ado about nothing.  My guess is that Torre knows that Ethier and Kemp are better players and he just had to say something good publicly about Pierre for various reasons (maybe to save face for the GM).  Managers always have to say something good about players unless they are criticizing them to the media for some reason.

I also think that most GMs leave playing time issues to the manager as long as they seem somewhat justified, regardless of a player’s contract.  Not to say that these issues don’t get discussed or that GMs don’t ask questions and make suggestions.  It also depends a lot on the manager (and the GM of course).  If you hire a Joe Torre to manager your team, I assume that you are not going to interfere much in his decision-making.

I understand that Pierre is a really good base runner and all (and I’m not 100% sure if you’re including that here), but I can’t figure out how he isn’t an absolutely pathetic hitter. If a .247 EQA is “not much worse than average” for a CF this year, then the CF crop is truly awful. And I know OPS+ isn’t a great measure by any means, but that’s 68 and that isn’t good either.

So my question: How does Pierre rank as not that bad? What am I missing?

If you want to discuss how bad or good a player “is” you better cite some numbers other than how they have done in some 341 PA in a career which includes over 5,500 PA.  Until you at least do that, there is nothing to discuss.

You see, using your framework, Ryan Ludwick IS one of the greatest hitters that ever lived, and Khalil Greene and Jeff Francouer ARE two of the worst hitters that ever put on a uniform.