Crooked Numbers

Playing With the Lineup

Nice article on batting order. It seems to me that when we think of ‘ideal’ batting orders, we are thinking of the best way to order
offensive parts When The Parts Are Working. When we run a simulated season, it is typically done Strat-o-matic style, as you did it. IIRC, the other studies I have seen operated similarly. So, two questions:

1) Is it true that in reality, offenses bunch up their successes more often than randomly – less successful than the norm when Randy Johnson pitches, better than normal when a Yankee middle reliever is on the mound, and if so, wouldn’t we expect to see a larger effect if we ran simulations that way?

2) Is the data on success-bunching available to do so?

–Mike Wendt


Interesting idea. I could see adding a “pitcher difficulty” factor — perhaps harder in innings 1-5, easier in 6-8, and harder in the ninth, but I’d like to back that up against real data first. For that, let’s go to the videotape. Here’s how batters did by inning in 2004, broken down by league:

2004    AL      1      .270   .340   .433
2004    AL      2      .263   .327   .407
2004    AL      3      .278   .341   .446
2004    AL      4      .284   .344   .471
2004    AL      5      .279   .343   .451
2004    AL      6      .280   .347   .463
2004    AL      7      .267   .339   .424
2004    AL      8      .254   .328   .388
2004    AL      9      .249   .323   .401

2004    NL      1      .272   .344   .441
2004    NL      2      .256   .320   .416
2004    NL      3      .261   .329   .425
2004    NL      4      .271   .339   .436
2004    NL      5      .264   .332   .429
2004    NL      6      .271   .340   .448
2004    NL      7      .260   .333   .421
2004    NL      8      .259   .333   .403
2004    NL      9      .250   .316   .400

Interestingly, both leagues see marked declines in the last three innings (as well as the second inning), so there may be something to this. But I wonder how much of that is just the cyclical nature of the lineup. I would attribute the dip in the second inning to the
likelihood that the bottom of the order is up in that inning more often than not. But the last three innings I can’t really explain other than to wonder about defensive replacements, pinch hitters, and relief specialists.

I’m not sure this answers your question about bunching success against particular pitchers, but it does raise some interesting questions about teams batting better earlier in the game and thus “bunching” their success in those innings.

–James Click

More on the Lineup

James –

Thanks for the articles. One question that should be fairly straightforward to address and might generate an interesting column:
with a lineup of 8 average players and one Barry Bonds, where is the optimal place to bat him? My gut tells me that it’s clearly no lower than 4th, but it’s the 2-3-4 (and even 1) question that’s interesting.



Running Barry through the lineup of eight average players yields the following results:

Pos  Min   Mean  Max
1    805   930  1106
2    797   925  1092
3    774   920  1076
4    784   925  1073
5    786   917  1076
6    780   910  1094
7    769   906  1050
8    765   901  1045
9    772   900  1070

Keep in mind, though, that this doesn’t take into account the fact that he would more frequently be batting with no one on base if he was leading off. If instead Mike Matheny and Jason Schmidt are substituted in the last two spots in the lineup, things look more like this:

Pos  Min   Mean  Max
1    637   796   924
2    677   795   923
3    664   796   948
4    672   798   954
5    662   792   925
6    645   780   913
7    653   771   889
8    641   761   885
9    653   764   886

In this case, he peaks in the fourth spot, but the difference between
there and the first three is small enough that it cannot be stated with
much confidence that that’s the “best” spot for him to bat. The extra
PAs by Bonds by batting him higher are muted by the decreased batters on
base in front of him.


LABR of Love

You seem to have run a pretty smart draft. I get that pitchers are more
of a gamble than hitters, but you can’t tank those categories. In a way I think that means one should go even more pitcher heavy to protect you from injuries. That said, you did get some nice value picks. I like Capuano in particular and you can’t complain about the prices on Willis
and Rusch.

My main concern with your team is catching. Yuck. You better plan to
turn some big trades there. This year I’m in the National Fantasy
Baseball Championship
and I’m finding that with 2 catchers on each of 15
teams, you have to get two decent ones. Even taking one ‘back-up catcher’–the fantasy term for having a Mike Matheny play for you every day–creates huge problems for your team.

Look at it this way, I’ve created a sort of VORP for fantasy players in which the number 0 represents the ideal players (rather than >replacement–I guess you could call it FVOIP) I’d need to win the
overall NFBC. I’m working with last year’s NFBC data set when they had 195 competitors so each number here represents a gain or loss of one point as it applies to the 2004 NFBC standings. Some are well over 0; most are under it.

Catchers kind of suck in general so even Pudge goes at -6.867. J.D. Closser, my kind of guy, has a -12.277. John Buck is similar and you can get those guys late in a draft or cheap at auction (a fantasy mag I have here says $9 and $7 respectively). By way of comparison, Koyie Hill rates at -21.532 and Ross is at a -24.347. To look at it another way the only everyday outfielder in all of baseball who is a worse fantasy player than even one of your catchers is Orlando Palmeiro who comes in at -23.742 (though I added a bit to his pessimistic PECOTA batting

In other words: Dude, get some catching!


Hi Brian,

Very interesting system. I like the concept of “Roto VORP.” I agree with you 100% that rostering a Matheny for the whole year can really hurt your squad, in that you’re dragging your batting average down, and also suffering the effects of opportunity cost, where you’re losing precious runs, RBI and HR that might have been racked up by a better option.

I think where you and I disagree, basically, is on the evaluation of the two catchers in question, Hill and Ross. Hill’s shown enough BA ability and doubles power in the minors that should he win the job, I like him to go something like .250-10-50-50 (basically similar to what the Player Forecast

With Dave Ross, I’m actually quite bullish (again much like PFM). As surely as I’d call a major regression if Ross lucked his way into a .330 BA last year, there’s no doubt in my mind that last year’s .170 was a huge fluke. I like Ross to hit about .240 this year, which won’t kill me. More importantly, even a pessimist wouldn’t question his ability to hit for power. Given the starting job in LA, I like him to hit 15+ HR. Call it .242-17-58-52.

I imagine your projection system relies too much on what Hill and Ross did last year, where I see both having only sporadic playing time in ’04. Give both regular gigs and I think you’ll see by year’s end that I got a couple bargains.

Of course as the wise Jeff Erickson told me: “If you don’t like your team after the draft, that’s when you know you messed up.” 🙂

–Jonah Keri

The Veterans Committee Ballot

Ted Simmons would compare favorably to the 13 enshrined catchers across
the board with the exception of Fielding Runs (which I’m guessing is already factored in when calculating his WARP).

Did Simmons fail to get 5% of the vote, and if so, how did that happen? How can an 8-time All Star with three top-10 MVP finishes at a position
with light Hall representation be essentially ignored? I was checking for his 5-year WARP3 peak, and he was so consistent that his best 8-year run is actually better than any 5-year run….

–David Salvia


First off, thanks for reading. Simmons did indeed fall off after garnering a mere 3.74 percent of the vote in his only appearance on the ballot (1994). I’m not sure whose puppy he ran over, or whether it was common knowledge that he was a lousy tipper, or if there were rumors that he was dabbling in the black arts making the rounds… for him to fall off the ballot so quickly was unwarranted, even if there were a number of good candidates up for vote (the top five from that year are all HOFers now).

According to current VC rules, Simmons won’t be eligible for consideration until 2009, 21 years after he’s been retired. He’s on what I’ll call The List of the Damned, players who are:

a) not active
b) not recently retired, awaiting their first shot at a Cooperstown ballot
c) not on the VC ballot
d) not on the ineligible list
d) above the JAWS positional averages

Bobby Grich: 117.1/46.8/82.0
Bill Dahlen: 121.8/40.7/81.3
Lou Whitaker: 118.0/39.6/78.8
Dwight Evans: 113.0/41.8/77.4
Darrell Evans: 113.7/39.0/76.4
Buddy Bell: 106.7/43.6/75.2
Brett Butler: 103.5/44.7/74.1
Willie Randolph: 110.7/36.9/73.8
Ted Simmons 101.0/40.8/70.9
Rick Reuschel: 101.8/39.6/70.7

Note that these numbers may differ slightly from what’s on the web, as the new DT cards went up a couple of weeks ago and I haven’t updated my spreadsheet yet.

Anyway, I hope that a new VC system is in place by the time Simmons is
eligible. I think he’s a very legitimate candidate who deserves a shot.

–Jay Jaffe

Bottom of the Ninth

Ranking Prospect Lists

I recently did a little number-crunching, and thought you’d be interested in the results. For years, I’ve been collecting prospect lists from various media outlets, and my goal was to find out who has done the best job of predicting prospect greatness.

My methodology was pretty simple. I assigned 100 points to the top prospect on each list, 99 points to the #2 prospect, and so on, down to #40. (Note: I chose 40, because for years, the BBP list included only 40 prospects.)

Next, I multiplied that factor by the number of career win shares each player has tallied to date. Then, I simply took a sum of those 40

The results for 2000:

1. Baseball Prospectus: 97,128
2. Baseball America:    93,502
3. John Sickels:        90,776

This was the first year I began collecting lists, and unfortunately, I didn’t collect very many of them. But these are generally considered to be the three most prestigious prospect lists.

1. Baseball Prospectus: 100,134
2. Baseball America:     81,264
3. John Sickels:         78,872
4. TeamOne Baseball:     78,530
5. Top Prospect Alert:   76,973

The problem in 2001 is that some outlets considered Ichiro! a prospect, while others did not. If we remove Ichiro!, here are the results for the top 39:

1. Baseball Prospectus: 89,811
2. Sickels:             78,872
3. TeamOne:             78,530
4. Top Prospect Alert:  76,119
5. Baseball America:    71,052

I figure it doesn’t make much sense to continue this study beyond 2002 (as the sample size for the players involved is too small), so here is the final year of the study:

1. The Sporting News:   56,926
2. Baseball Prospectus: 51,831
3.         48,788
4. Baseball America:    47,968
5. Sickels:             44,598
6. TeamOne:             44,038
7. Top Prospect Alert:  40,654

The problem here is that TSN included several players (such as Vernon
and Corey Patterson) who weren’t considered prospects by anyone
else. If you remove those two, Baseball Prospectus easily ranks #1.

So, in other words, Baseball Prospectus’ list ranks #1 from 2000-2002!
Congratulations on a job well done!


Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe