Batter's Box Interactive Magazine Batter's Box Interactive Magazine Batter's Box Interactive Magazine
What could that title possibly mean? The first part refers to a Diamond Mind 2004 projection disk that Baseball Primer's Dan Szymborski has created, using his own projection system called ZiPS. Dan has produced 5 builds of the disk so far. I have added transactions up to and including the Javy Lopez signing (with Baltimore). The "Dudek" part refers to my own efforts to create Manager Profiles for all the AL teams.



Diamond Mind is a sophisticated and flexible baseball simulation. I asked the program to play 15 computer-generated seasons using current rosters with the players in roles that I've decided are likely given current rosters. There's still a long way to go before spring-training rosters are set, so I expect to produce a 2nd edition in mid-January (by which time almost all of the big free agents will have signed) and a 3rd edition at the beginning of March.

And so they played - little computer-generated players hitting, pitching, fielding and getting injured over 15 years. Here are the median performances of all the AL teams, as well as the number of times in 15 years they made the playoffs (playoffs were not simulated).

 Team....... W-L .. R-RA playoffs
Boston...... 99-63 911-681 15
Toronto..... 91-71 861-767 3
New York.... 89-73 882-769 6
Baltimore... 81-81 756-758 2
Tampa Bay... 63-99 696-872 0


Cleveland.... 82-80 741-742 6
Kansas City.. 80-82 783-767 8
Chicago...... 74-88 717-820 1
Minnesota.... 73-88 721-782 2
Detroit...... 66-96 641-800 0


Oakland..... 97-65 763-624 12
Seattle..... 85-76 781-708 3
Texas....... 79-83 802-809 1
Anaheim..... 77-85 706-723 1



Before anyone gets excited about the relative positions of the Yankees and Blue Jays, we should remember that the Jays are nearly set, while the Yankees signing of Kenny Lofton is on the horizon (their defence needs all the help they can get). Many people think the Yankees have improved themselves substantially. Vazquez, Brown and Contreras will replace Clemens, Pettitte and Wells in the rotation, and will soon find out what it means to have the Yankees defence behind them. Offensively, the major change so far is Sheffield in and Nick Johnson out. But is that enough to overcome the aging lineup?

The Orioles are set to close the gap even further, with the Vlad signing apparently imminent; nevertheless they still don't have a third baseman and possess a starting rotation of largely unproven pitchers.

Minnesota needs to do a lot of work, especially if they insist on handicapping themselves with Rivas and Guzman as their middle infielders. With a lack of high-obp players in the lineup, the sim is very pessimistic about the Twins chances of scoring a lot of runs this year.

Anaheim has added some good pitching, but are in need of offence (they are projected to finish 13th in run scoring). Seattle has lost some talent, and drifted back into the pack, which makes the Athletics the prohibitive favourites to take the West again.



The Szymborski-Dudek Diamond Mind Simulation, 1st edition | 27 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.
_Jay - Monday, December 22 2003 @ 07:13 PM EST (#22408) #
I played about a quarter of a season with the 2nd edition of his projection disk before abandoning that endeavour when I received the 2003 Season Disk. The one point that stood out for me was the production projected for Josh Phelps when slotted in the five slot every day. I think had had 10 homers in the first 25 games and was definitely slugging at or around .700 in my sim that early.

I enjoyed playing with the projection disk but readers should also note a couple of important points about those projections. Again I used the 2nd edition but I don't imagine much has changed.

1) Lefty-Righty splits were not included, meaning O-dawg hit the same, regardless of who was on the mound.

2) I felt he was a tad generous with the projections of our young players but that may just be me. Robert may have accounted for this though by using realistic lineups instead of computer selected based on projected results. For example, I believe the projections had Lexi at 440 at bats, and hitting over .320 Avg.

I don't want to seem negative because creating the projections and then exporting those into DMB players is an extremely large amount of work. I was so grateful to see and to use the disk. But in reading the results, as Robert says, remember to include one or two grains of salt.
robertdudek - Monday, December 22 2003 @ 07:28 PM EST (#22409) #
Jay,

I agree that the projections of young players are probably overoptimistic. But almost none of the rookies in the AL were given a regular job (I gave Rios about 150 AB and Gross about 200). I also had Hudson sit against half the lefthanders and I made many choices that I know the big league club will actually make, even though they make the team worse (e.g. Rivas as the evryday 2B).

There are still some problems at this early stage: as mentioned Baltimore doesn't have a 3B, Anaheim has no credible 1B and the Yankees DH is Ruben Sierra.

I like to view this as an estimate of the current strength on paper, with the proviso that some teams are just beginning to fill out their everyday lineups and key pitching roles.
_Jonny German - Monday, December 22 2003 @ 07:42 PM EST (#22410) #
Great stuff Robert. I'd like to hear more about the details of the Jays performance, i.e. how many at-bats for Phelps, how did Hinske do, what was your bullpen usage like?
robertdudek - Monday, December 22 2003 @ 08:04 PM EST (#22411) #
I didn't record stats from all the individual seasons. The players are supposed to perform based on their ZiPS projection (available at Transaction Oracle - see link above).

The bullpen usage I can tell you about, since it's part of the Manager Profile. The closer role was mostly Aquilino, with Kershner coming in if there were a bunch of lefties, or Ligtenberg when Lopez was tired. Ligtenberg was the main setup man, with Kershner and Speier backing him up. Haines was the main mop-up man; Chen was the second lefty. Chen and Walker were spot starters; the rotation was Halladay, Batista, Lilly, Hentgen with Josh Towers skipped when an off-day allowed.
Coach - Monday, December 22 2003 @ 08:17 PM EST (#22412) #
I'm sure that all the Manager Profiles are as realistic as the Jays' -- Robert is the reigning champion of the Three Run Homer League against some tough opposition -- but ZiPS, like any other prediction system, isn't a flawless crystal ball. It's quite impressive, and I'm very grateful to Dan for creating the DMB disk, but there will always be players who ZiPS and my hunches don't agree on, like Greg Myers and Vernon Wells a year ago.

Some surprising results, for sure. Cleveland over .500? The AL Central is almost as bad as I thought, but I've been assuming that the Royals had a slight edge on the Twins, with the Tribe another year or two away. Out West, I'm surprised that the Angels didn't give the Mariners more of a battle, but not shocked.

In the East, it will be interesting to see how much Guerrero affects the "standings" if he becomes an Oriole, but since he won't help Baltimore's pitching, I suspect it won't make an enormous difference. I also can't imagine Lofton will be much of an upgrade in New York. Any system that has the Jays winning the wild card over the Yankees in that final three-game series at the Dome has my enthusiastic endorsement.
robertdudek - Tuesday, December 23 2003 @ 12:02 AM EST (#22414) #
Baltimore is a serious pitcher's park.
robertdudek - Tuesday, December 23 2003 @ 12:05 AM EST (#22415) #
Also, it's probably a quirk in the ZiPS, but the run scoring environment that results from the sim is a lot lower than the one we had in 2003.
_pete - Tuesday, December 23 2003 @ 12:25 AM EST (#22416) #
very cool stuff. by the way, the orioles are planning to slot in melvin mora as their third sacker, with he being such an all-purpose guy and all.
robertdudek - Tuesday, December 23 2003 @ 12:51 AM EST (#22417) #
Thanks pete - I'll make a note.
_Jurgen - Tuesday, December 23 2003 @ 12:57 AM EST (#22418) #
Robert:

I'm not surprised Boston did well, but New York's drop surprised me. On paper they seem to me a better team than the 2003 edition. I'd be curious to know what your simulation would predict for the 2003 Yankees. Better, same, or worse than how they actually did?

I'm wondering how the new faces in the East performed in your simulations: How many homers did Javy average? How many innings did Brown pitch? Did Schilling stay healthy? Did Tejada keep up with the Nomars and Jeters? Did Sheffield outslug Giambi? Did Batista outpitch Escobar?
robertdudek - Tuesday, December 23 2003 @ 01:11 AM EST (#22419) #
One thing to do would be to take a look at the ZiPS projections. Once Lofton signs, Bernie pushes Sierra out of the DH slot and the defence improves. Right now, there's virtually no depth on the offensive side. My guess is that when injuries hit the sim Yankees no one was available to step up and produce. Kevin Brown is, I think, a particular injury risk. The sim Yankees won 100+ games 4 times, but were in the 80-89 win territory more often than the sim Jays.

In reality, the Yankees will find some sort of replacements if their guys go down to injury. One key thing to remember is age - virtually all the sim Yankees are expected to be worse than they were in 2003 (by the sim).
robertdudek - Tuesday, December 23 2003 @ 01:25 AM EST (#22420) #
Let me say that the Yankees as currently constituted are NOT better than the 2003 version. While the new players, as a whole, are better than the group that left (but not hugely so - e.g. Brown isn't much better than Pettitte: league, park and defence have created the illusion he is), the large majority of the rest of the guys are well past 27.

The Yankees "problem" is that while they are treading water, their 3 rivals have taken big steps forward. Toronto has quite a bit of young talent, which would naturally push them forward even if they hadn't solidified their pitching. So it shouldn't be a surprise that they have gained on the arthritic Yankees.
_Jurgen - Tuesday, December 23 2003 @ 02:08 AM EST (#22421) #
Brown is a huge question mark. He's obviously about a thousand times more useful to the Yankees than Weaver even if he spends half the year on the DL, but the question is how good can he be when he's not on it.

Brown ('01-'03):

2.86 ERA, 390.1 IP, 8.0 K/9, 3.0 K/BB, 0.6 HR/9

Pettitte ('01-03):

3.82 ERA, 543.2 IP, 7.3 K/9, 3.6 K/BB, 0.7 HR/9

Are league, park, and defence worth an entire run? Well, the peripherals suggest it's a possibility. Brown had a disastrous 2002 which kinda throws off his numbers, but he will be 39 next year. And with his history of injuries, a disastrous 2004 shouldn't be out of the question.

I just have a hard time believing that Sheffield, Quantrill, Gordon, Giambi, Williams, Posada, Jeter, Matsui, Rivera will also all be that much worse in 2004. Sheffield and Quantrill show no signs of slowing down, and Giambi still managed to hit .325 EqA while half-blind and suffering from a major injury.
robertdudek - Tuesday, December 23 2003 @ 02:14 AM EST (#22423) #
I'm particularly worried for Quantrill - his low strikeout rate makes him a risk with that defence. Remember Mr. Jeter and Mr Giambi will be playing in the infield every day (until Giambi gets hurt).

At some point you have to ask yourself how long you can keep rolling the dice. Take 10 players in their 30s - the odds are very good that, collectively, they won't be as good in 2004 as they were in 2003.

Soriano and De Paula are going to be the only young players on the team. This is an old team. It's a tough division. It's going to be fun to watch.
robertdudek - Tuesday, December 23 2003 @ 02:19 AM EST (#22424) #
Tejada is an improvement on defence. I made the sim go with Roberts at 3B and Mora istead of Gibbons in the outfield. Riley and Ainsworth figure to be effective pitchers. The Yankees are old and have little depth; their defence sucks. I think the sim is wrong about how many runs the Orioles will allow, mostly because (IMO) it favours young players.

Anyways, the main point is that it says the Orioles are a .500 team right now, which I think is correct.
Mike Green - Tuesday, December 23 2003 @ 09:12 AM EST (#22425) #
Robert, your prognostication of Carlos Tosca's use of the pitching staff is in my opinion a bit optimistic. The roles you have prescribed for the pitchers are pretty much optimal, and there is little evidence that Carlos will do so well.

Q: why would a Yankee team with pretty much unlimited cash and in need of a good defensive centerfielder be interested in Kenny Lofton and not in Mike Cameron?

A: they're not as smart as some would have them be.

Great work, Robert; your explanation of the limitations of the sim are much appreciated.
_Jonny German - Tuesday, December 23 2003 @ 09:38 AM EST (#22426) #
Also, it's probably a quirk in the ZiPS, but the run scoring environment that results from the sim is a lot lower than the one we had in 2003

Is it possible that ZiPS is exaggerating the park factors? Because besides the curious situation with Baltimore, the numbers Dan has posted for the Jays seem to me to be optimistic across the board for the hitters and pessimistic, for the most part, for the pitchers.

Remember Mr. Jeter and Mr Giambi will be playing in the infield every day (until Giambi gets hurt).

If Lofton does sign, I think the Yankees would be wise to hit a lot of ground balls at Bernie in spring training, to see if he can be a serviceable first baseman.

It's possible that when we look back on this offseason we'll say the Mets outdid the Yankees. They've only added two important pieces, Matsui and Cameron, but they haven't lost anybody that was worth keeping.

The Yankee equation so far:

Plus
Sheffield, Vazquez, Brown, Quantrill, Gordon

Minus
Johnson, Clemens, Pettite, Hammond.

What's the status of Nelson? Busted, I know, but is he returning or retired or what?

The roles you have prescribed for the pitchers are pretty much optimal, and there is little evidence that Carlos will do so well.

I think Carlos will make better use of his pen with better relievers available to him. Many of his bizarre moves last year seemed to be out of lack of confidence in most of the bullpen. I also think the front office should and will put him on a shorter leash if he continues to favour stunts like 5 reliever innings.
_Mick - Tuesday, December 23 2003 @ 10:09 AM EST (#22427) #
Q: why would a Yankee team with pretty much unlimited cash and in need of a good defensive centerfielder be interested in Kenny Lofton and not in Mike Cameron?

I won't completely disagree with your conclusion, but I think it's less a matter of the Cashman Yankees being dumb and much more a problem of the King George Yankees liking "names." Who is Mike Cameron, to the casual fan? Oh, he's the guy with the fluke four-homer game who played out west for, uh, Oakland? Or Colorado? But Kenny Lofton! All those World Series! All those stolen bases! Sexxxxxy!!!
Pepper Moffatt - Tuesday, December 23 2003 @ 10:15 AM EST (#22428) #
http://economics.about.com
I won't completely disagree with your conclusion, but I think it's less a matter of the Cashman Yankees being dumb and much more a problem of the King George Yankees liking "names." Who is Mike Cameron, to the casual fan? Oh, he's the guy with the fluke four-homer game who played out west for, uh, Oakland? Or Colorado? But Kenny Lofton! All those World Series! All those stolen bases! Sexxxxxy!!!

I thought there was a far more rational explanation: The Yankees couldn't sign any more Type A Free Agents, as per MLB rules.

Mike
_Andy - Tuesday, December 23 2003 @ 10:28 AM EST (#22429) #
Are you saying the O's will sit Gibbons, or DH him? Everything that I have read points to Mora at 3rd, Hairston at 2nd and Roberts the odd man out. I actually would guess that Gibbons will play at 1st next year.

The loss of Johnson really hurts the Yankees defensively becuase Giambi has to play 1st every day now, and I don't know if his body is up to that challenge. And if he goes down, they have to rely on Sheffield for the power on that team. I'm not sure if this year's Yankees will be better or worse than last year's team; all I know is that they need a lot of things to go right/stay right just to meet expectations.

It's really difficult for any simulation to factor in a manager's usage of his bullpen.
Mike Green - Tuesday, December 23 2003 @ 10:35 AM EST (#22430) #
Mike M,

My recollection from another thread was that the Yankees could sign the same number of type A free agents as they had lost, and that they had lost eight, and had so far signed five. I'll check.

Mike G
_Joe Dimino - Wednesday, December 24 2003 @ 09:45 AM EST (#22431) #
http://www.baseballprimer.com/hom
Hey Robert, just for fun, could you run a couple of seasons with the proposed Manny/ARod chain of events and set what the differences are?

Manny to Texas
ARod to Boston
Nomar to LA
Magglio to Boston
Odalis Perez to White Sox

Am I missing anyone? Who else was LA giving the White Sox?

I'd love to see how this would change things . . .
_Ed - Wednesday, December 24 2003 @ 07:20 PM EST (#22432) #
Did you happen to take note of the NL standings in your 15 simmed seasons?
robertdudek - Wednesday, December 24 2003 @ 10:47 PM EST (#22433) #
I haven't set all the NL Manager Profiles yet. I plan to provide full standings in the next edition.
Mike Green - Friday, March 26 2004 @ 11:29 AM EST (#22434) #
http://www.baseballprimer.com/articles/rnjohnson_2004-03-24_0.shtml
Robert, it would be interesting to see what the simulation results would be with the personnel changes since December.

Also, I am curious how the ZiPs projections for minor leaguers are derived. The projections for Rios and Gross are excellent, and much better than for Quiroz. COMN for Primer article with ZiPs projections at bottom. Can you point me to a description? Thanks.
The Szymborski-Dudek Diamond Mind Simulation, 1st edition | 27 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.