AL Central - Winter Rankings

by Xeifrank on Feb 4, 2008 11:36 PM CST

I am using my baseball simulator that I wrote last year to rank all 6 baseball divisions based on 2008 ZIPS Projections. The simulator uses advanced sabermetric algorithms to pit two teams against each other taking into account nearly every facet of a baseball game. It is still light on modelling defense, as there doesn't seem to be any one defensive metric that the great minds can agree to. In this study I took roster depth charts from both Rototimes and Yahoo and played each team against each of the other divisional teams, with #1 pitchers going up against #1 pitchers and #2 pitchers going up against #2 pitchers etc..., each game was simulated 2000 times, played at both home and away (park factors involved). All the games are summed up and an average winning percentage is calculated. Below are the AL Central Results. The White Sox, obviously did not fair too well in this study. ZIPS is not particularly kind to the ChiSox. ZIPS creator Dan Szymboski made this comment, "they're no longer even a heavy favorite to beat out the Royals for 4th place in the division". I am waiting for the Bedard trade to be official before posting the AL West/East rankings.

Detroit .5459
Cleveland .5230
Minnesota .5112
Kansas City .4639
Chicago .4586

vr, Xeifrank

SouthSideSox is a community driven site. As such, users are able to express their thoughts and opinions in a FanPost, such as this one, which represents the views of this particular fan, but not necessarily the entire community or SouthSideSox editors.

0 recs | Comment 52 comments

Story-email Email Printer Print

Comments

Well....

I was thinking the Sox would be around 80 wins, but I could just easily see 70 and a finish behind KC. I think the Royals are going to have a pretty decent year (you know...relative to their recent years).

So anyway, I think your simulation is pretty believable.

by hitlesswonder on Feb 5, 2008 9:57 AM CST 0 recs

XeiFrank? As in Z (Zolly) Frank?

As in Z Frank Chevrolet? If so, good one.

Question re your projections:

What lineup did you use? For example, was Owens or Swisher in CF? Crede or Fields at 3B? Uribe at 2B, Richar? Did Thome DH against LHP?

Not saying you're offbase (yet), but I suspect there's still changes to come...

Next time, Bhoov, do some research.

by Chiburb on Feb 5, 2008 11:12 AM CST 0 recs

i also wouldn't use ZiPS

PECOTA has shown itself to be better than ZiPS in the past. but i'd imagine this is about right based upon ZiPS. one other problem with ZiPS that could be lodged is that dan did no projection for alexei ramirez. i realize that one isn't for the faint of heart but he could, arguably, be a starter and PECOTA likes him. and one problem with this guy's calculation in general, methinks, is that he did it before dan released his updated projections on monday. not sure if he used the updated numbers for swisher or dotel (or even included them).

by larry on Feb 5, 2008 11:24 AM CST up 0 recs

what does

dotels ZiPS look like?

12/12/07- We'll miss you Andy Gonzalez

by The Deacon on Feb 5, 2008 11:30 AM CST up 0 recs

Is any projection system...

very good at projecting relievers? The small sample size they pitch and general volatility of relievers makes projecting them (successfully) a rather difficult task.

by CWSKeith on Feb 5, 2008 11:48 AM CST up 0 recs

of course not

and i imagine we all know to take this stuff with the appropriate hunk of salt. but if you're going to compare teams, you may as well do it with the correct lineups and players as possible. dotel wouldn't make much difference, anyway.

and i don't know what dotel's numbers looked like. i'm sure it can easily be found on dan's site; he posted his projections in excel files yesterday.

by larry on Feb 5, 2008 11:53 AM CST up 0 recs

ZiPS...

only has 25 IP for Dotel.

4 homers, 11 BB, 29 K - 4.32 ERA.

The big thing will be the starting pitching, though - the bullpen won't matter if the starting pitching is as bad as the projection systems are making it out to be.

by The Jerry Royster Experience on Feb 5, 2008 12:17 PM CST up 0 recs

ZIPS etc...

First of all, thanks for all the thoughtful comments/suggestions from everyone. I will try to answer a few of the questions. I did not include Dotel in the simulation. I didn't get his ZIPS in time, but seeing how his ZIPS projection isn't all that good, it won't make a difference. As far as using PECOTA over ZIPS, yeah that could've been done, but I don't have access to any PECOTA projections, I believe you have to pay for them. I could've used MARCEL or CHONE, but studies have shown they are all about the same when it comes to accuracy. The lineup I used for the White Sox was the following...

Cabrera
Swisher
Thome
Konerko
Dye
Pierzynski
Fields
Quentin
Uribe

This is the exact line up that RotoTimes has listed for the White Sox, which is what I tended to go with for all teams, unless I saw something obviously flawed.

vr, Xei

Author: Dodger Sims

by Xeifrank on Feb 5, 2008 12:33 PM CST 0 recs

Sounds like this is pre-Santana trade also

Just curious, what do you come up with for Sox Runs Scored/Runs Allowed? I'm assuming the starting pitching is what's killing us.

"Swisher's versatility allows us to give Owens a chance to come back and show he's ready to be that sparkplug like Podsednik was a couple of years ago." - KW

by ChicagoPete on Feb 5, 2008 12:36 PM CST up 0 recs

sounds like...

Actually, it was post Santana trade. I held the rankings and then re-ran all of the Twins games with no Santana, moving all the starters up one spot in the rotation. The Twins were actually very close to the Indians pre-Santana trade. The sim didn't weight starts due to injury concerns, so the Twins could be docked more if Liriano doesn't make too many starts.

As far as the runs per game goes, I tracked runs per game, but that can be misleading. A better method would be to track runs scored per 27 outs, as losing teams tend to bat more times. Perhaps a later enhancement.

vr, Xei

Author: Dodger Sims

by Xeifrank on Feb 5, 2008 12:52 PM CST up 0 recs

wow

They're still playing .511 even without Santana? That's impressive.

"Swisher's versatility allows us to give Owens a chance to come back and show he's ready to be that sparkplug like Podsednik was a couple of years ago." - KW

by ChicagoPete on Feb 5, 2008 1:05 PM CST up 0 recs

ZiPS...

likes Liriano and Slowey quite a bit, and projects an average offense.

ZiPS really loves the Twins bullpen.

by The Jerry Royster Experience on Feb 5, 2008 1:14 PM CST up 0 recs

How much difference does it make?

The White Sox won 72 games last year, but finished with a 39-33 record vs the Central.

By my count they havent gotten worse and every team doesnt have this massive improvement.

Besides they havent finished below .500 vs the Central since 1996. Me thinks other things go into "it". I dont think I will lose any sleep over this one.

"There is not a liberal America and a conservative America. There is the United States of America."

by Tdogg on Feb 6, 2008 12:59 PM CST 0 recs

hmmm

i'm not getting your point about the record v. the central.

2/6/1958: flowers of manchester

by larry on Feb 6, 2008 1:09 PM CST up 0 recs

Am I missing something?

"In this study I took roster depth charts from both Rototimes and Yahoo and played each team against each of the other divisional teams, with #1 pitchers going up against #1 pitchers and #2 pitchers going up against #2 pitchers etc..., each game was simulated 2000 times, played at both home and away (park factors involved)."

Isnt this White Sox vs central teams only? If I am mistaken I apologize.

Of course mine wasnt scientific or anything. :-)

"There is not a liberal America and a conservative America. There is the United States of America."

by Tdogg on Feb 6, 2008 1:30 PM CST up 0 recs

i guess i'm wondering

what you think this "other things that go into it" are and why whatever these things are supposedly will continue to prop up the sox record. last year looks sort of like an aberration to me. prior to last season, the sox since 1996 either had not been a below .500 team or, failing that, didn't finish below 2nd in the division. so it seems pretty intuitive that they shouldn't have been below .500 against the division in those years.

2/6/1958: flowers of manchester

by larry on Feb 6, 2008 1:41 PM CST up 0 recs

Come on Larry.

I dont know. Maybe it was the second full moon rising over the horizon during leap years. Maybe the Sox steal signs from "only" central opponents.

There were some some sub .500 ball clubs in those 11 years. I didnt make any claims of profound wisdom. I just said I wonder how relevant it is.

Even by your own logic if the team is "only" close (above or below) .500 (as I think most would agree), chances are IMO they don't finish in the cellar of divisional games.

"There is not a liberal America and a conservative America. There is the United States of America."

by Tdogg on Feb 6, 2008 1:48 PM CST up 0 recs

sheesh

sorry for asking. was hoping for a bit more insight into 'why' than a full moon.

2/6/1958: flowers of manchester

by larry on Feb 6, 2008 1:53 PM CST up 0 recs

Sorry didnt mean it like that

If you are truly asking for my opinion why, I truly believe these knuckleheads honestly do care more about beating division rivals more then anyone else, especially lately.

Thats why I am curious about the other years. Sometimes its just goofy like that. Bears/Packers example.

Hell I believe Detroit is significantly better than the White Sox, but METHINKS they arent thrilled to play them.

"There is not a liberal America and a conservative America. There is the United States of America."

by Tdogg on Feb 6, 2008 2:01 PM CST up 0 recs

Correct me if I am wrong...

but wouldn't it be a fairly easy exercise (for those of you who dig through historical numbers) to just compare the Sox record for each year to their record against the central that year? Wouldn't that provide evidence of whether or not they are overperforming against the central?

by CatBrains on Feb 7, 2008 10:26 AM CST up 0 recs

Did that already Cat

The other years was in reference to the actual results vs the simulation with the same parameters. Frank liked the idea but thought it was too much work (I totally understand).

"There is not a liberal America and a conservative America. There is the United States of America."

by Tdogg on Feb 7, 2008 10:35 AM CST up 0 recs

OK, WAY too many folks

hopping on the "methinks" bandwagon for my taste. Since when did SSS turn into effin' Shakespeare?

Fer Chrissakes, already.

Jeff Cox: I'm going to call you "FN".

by winningugly on Feb 6, 2008 1:34 PM CST up 0 recs

lol

Screw you WU.

"There is not a liberal America and a conservative America. There is the United States of America."

by Tdogg on Feb 6, 2008 1:40 PM CST up 0 recs

i'm a trendsetter

2/6/1958: flowers of manchester

by larry on Feb 6, 2008 1:42 PM CST up 0 recs

Im curious

XiFrank not to make a lot more work for you but could you run some previous years, with zip data under the same parameters and compare to actual results?

"There is not a liberal America and a conservative America. There is the United States of America."

by Tdogg on Feb 6, 2008 1:38 PM CST 0 recs

re curious

Yeah, that's too much work. But it's a good idea.
vr, Xei

Author: Dodger Sims

by Xeifrank on Feb 6, 2008 4:04 PM CST up 0 recs

Good Diary

You get a Pizza Trophy, but I'd be more interested in seeing multiple simulations of the regular season, not just central.

AIM: SouthSideCheat

by The Cheat on Feb 6, 2008 2:23 PM CST 0 recs

USS Mariner

just did the same thing, using assumed rosters after the Bedard trade is completed. The results weren't pretty for the M's either.

http://ussmariner.com/2008/02/05/the-2008-mariners-through-zips/

AIM: SouthSideCheat

by The Cheat on Feb 6, 2008 2:26 PM CST up 0 recs

multiple sims

of the regular season? Perhaps at a later time. I was just doing a divisional ranking. One other thing I can do with my sim is rank bullpens by starting a 5000 game simulation in the top of the 7th inning with the score tied and both starting pitchers get taken out to start the 7th. I give both teams the same offensive lineups and adjust all stats/projections to a neutral park. Thanks for all the good questions/suggestions folks. Probably some of the best feedback I've gotten in a while.
vr, Xei

Author: Dodger Sims

by Xeifrank on Feb 6, 2008 4:08 PM CST up 0 recs

Here's some tools

You probably saw them on BTF/Transaction Oracle, but I thought I drop them here.

SG, of Replacement Level Yankees posted some tools to run DMB for 100 iterations.

http://www.replacementlevel.com/stuff/sg/RunDMB.zip

AIM: SouthSideCheat

by The Cheat on Feb 6, 2008 10:16 PM CST up 0 recs

Seasons...

Ok, I've added some code to the simulator. I can now create a schedule for any team and play any number of seasons. I will do so for the White Sox in the near future and post the results. If it's too slow I won't run too many seasons.

vr, Xei

Author: Dodger Sims

by Xeifrank on Feb 6, 2008 11:20 PM CST 0 recs

Seasonal sim update

Ok, I ran 14 seasons so far for the White Sox. I just played 162 games for the White Sox, using their 2008 schedule, so there are no standings for the other teams, only the White Sox. I used a five man rotation and skipped the 5th starter around days off.

The average totals so far are...

Wins: 72.4
Losses: 89.6
Runs Scored: 710.0
Runs Allowed: 778.1
Win Percentage: .4469

Best season: 85-77
Worst season: 61-101

To get more exact/accurate numbers I'd need to run many more seasons, but I have a feeling that the averages won't change much.

vr, Xei

Author: Dodger Sims

by Xeifrank on Feb 7, 2008 2:18 AM CST up 0 recs

not too far off

from most of our best case scenario predictions and most people have been saying anywhere from 72-77 wins is the most likely win total for the sox. pretty cool to see.

12/12/07- We'll miss you Andy Gonzalez

by The Deacon on Feb 7, 2008 7:16 AM CST up 0 recs

Excellennt job, xei

I called for 72 wins back in Nov, I think, and this confirms the range. With the cannibalizing of the farm to prop up the present I think we might hit .500 if all goes well, but still have the capacity to suck very, very hard if injuries are realized and potential isn't.

Thanks for your work. What do we owe you?

Jeff Cox: I'm going to call you "FN".

by winningugly on Feb 7, 2008 8:29 AM CST up 0 recs

excellent

unlike my typing.

Jeff Cox: I'm going to call you "FN".

by winningugly on Feb 7, 2008 10:53 AM CST up 0 recs

thanks for that

obviously we can all sit here and pick holes in what you did but you've actually done some projections and it's probably pretty close to the ballpark of what we're looking at.

2/6/1958: flowers of manchester

by larry on Feb 7, 2008 10:42 AM CST up 0 recs

ugh

72-90? Ouch. And 710 runs scored? - that's only 17 more runs than last year with Andy, Alex and Jerry. That spells one long, brutal season. Thank gawd The Hawk has been extended, it will take all his skills and all his powers to put a positive spin on that kind of season.

"Swisher's versatility allows us to give Owens a chance to come back and show he's ready to be that sparkplug like Podsednik was a couple of years ago." - KW

by ChicagoPete on Feb 7, 2008 1:05 PM CST up 0 recs

Thanks for these

I think the next step would be to find an optimistic projection...

Install Ramirez as the starting 2B with his PECOTA line of .295/.342/.452
Limit Owens at-bats. Quentin projects much better.
Drop the 5th starter in favor of Egbert, who has great projections. (This effectively gives one of our 3 question marks in the rotation a good season.)

AIM: SouthSideCheat

by The Cheat on Feb 7, 2008 2:58 PM CST up 0 recs

I could do that if you could provide for me the following for Ramirez and Egbert.

For Ramirez.
Projected OBP,AB,H,2B,3B,HR, some kind of speed rating from 1 to 100 and does he bat L,R or B.

For Egbert
Projected IP,K,BB,H,HR, LH or RH and if possible a GB/FB ratio.

And given the lineup I used near the top of the diary, re-arrange it using the same players I did (I already had Quentin in there over Owens), also plugging Ramirez in there. I will wait for that info.

vr, Xei

Author: Dodger Sims

by Xeifrank on Feb 7, 2008 3:17 PM CST up 0 recs

I don't have the full PECOTA for Ramirez

But I'll point Larry this direction.

Egbert (according to ZIPS, PECTOA is similar): 160IP, 180H, 16HR, 67BB, 101K. He's right handed and has had plus GB rates throughout his MiLB career. 55% GB, or a 1.93 GO/FO ratio according to MiLB.com

AIM: SouthSideCheat

by The Cheat on Feb 7, 2008 4:27 PM CST up 0 recs

Given the new players

I think the lineup should go something like this:

Ramirez
Swisher
Thome
Konerko
Dye
Fields
Quentin
Pierzynski
Cabrera

AIM: SouthSideCheat

by The Cheat on Feb 7, 2008 4:35 PM CST up 0 recs

I love that lineup...

but while I can actually see Ozzie leading off Ramirez, I don't see him batting Swisher second over a speedier guy who can bunt in Orlando Cabrera. Even though Swisher is supposedly a good bunter, but he shouldn't be bunting anyway.

by SSH2005 on Feb 7, 2008 6:10 PM CST up 0 recs

ramirez as 2B

OBP .342
AB 463
H 137
2B 28
3B 2
HR 14

no idea for speed rating. he's not slow but he's not ridiculous fast, either. 9 SB is what they have him at. bats righty.

by larry on Feb 7, 2008 4:36 PM CST up 0 recs

I'd say he has above average speed

on a scale of 1-100, where Konerko, AJ, Thome, and Crede can't crack 10, Swisher is about 40-50, Owens is about 90, I'd put Ramirez around 70.

AIM: SouthSideCheat

by The Cheat on Feb 7, 2008 4:54 PM CST up 0 recs

Got it.

Thanks for the info. I will try to rip thru a few seasons. I will post the results and how many runs I did within the next 2 hours.

First season just went 73-89 .451
RS-721
RA-751

Curious to see if the jumps in RS and RA are outliers.

vr, Xei

Author: Dodger Sims

by Xeifrank on Feb 7, 2008 4:56 PM CST up 0 recs

Thanks for running through all the iterations

I appreciate you tinkering and re-doing sims with requested changes

-Jeeves Life in the Cell

by jeeves on Feb 7, 2008 5:28 PM CST up 0 recs

Run #2

Ok, with Ramirez in there, Egbert as the 5th starter and the lineup shuffled the new and improved White Sox after 15 full seasons simulated now look like...

Wins: 79.4 (Low 70, High 88)
Losses: 82.6
RS: 737.4 (Low 673, High 791)
RA: 765.7 (Low 724, High 826)
Win%: .4901
Standard Dev(Wins): 4.15
Sample Size: 15 Seasons

From watching some of the game logs/scores scroll by the big difference in the improvement (which makes sense) is that the White Sox seemed to have more success against the weaker teams. There is one stretch of the schedule, I believe it is games 12 thru 19 where their record always seems to improve. The first 3 or 4 series are tough.

If anyone wants to take this a step further and run their own simulations I can email you a zip file that has the setup.exe in it. The GUI is nice but you kind of have to know what you're doing. If you let me know what you want to simulate, I can send along some simple instructions. The projections are ascii files that you can edit, the schedule too. Or you can just run a sim with two teams playing each other as many times as you'd like. Or you can manage one of the teams (Visitors option only now) and play for fun.

vr, Xeifrank
(xeifrank@yahoo.com)

Author: Dodger Sims

by Xeifrank on Feb 7, 2008 6:33 PM CST up 0 recs

egbert didn't seem to help much

RA remained pretty high. shuffling the lineup, however, did seem to pay some good dividends.

by larry on Feb 7, 2008 6:44 PM CST up 0 recs

RA

keep in mind that RA could've been slightly inflated due to the White Sox winning more games, which meant their opponents in road games won by the White Sox would've batted an extra inning. Not sure how much of a difference that would make. RS/27 and RA/27 outs would do a better job of showing this, but I don't have that programmed in (yet).

vr, Xei

Author: Dodger Sims

by Xeifrank on Feb 7, 2008 6:51 PM CST up 0 recs

good point

since you mentioned that earlier, i should have remembered that.

by larry on Feb 7, 2008 6:54 PM CST up 0 recs

That jibes with the off-season

improvements, IMO. A .500 club sounds about right.

Jeff Cox: I'm going to call you "FN".

by winningugly on Feb 7, 2008 8:01 PM CST up 0 recs

Simulations with PECOTA

Xeifrank shared his program with me, very cool easy to use. Been playing with it this morning, and ran through ten seasons.

I used this starting lineup with pecota:

"Orlando Cabrera",322,516,139,27,2,7,80,1
"Nick Swisher",373,531,141,30,1,31,45,3
"Jim Thome",379,397,106,18,0,29,20,2
"Paul Konerko",354,516,138,28,0,27,40,1
"Jermaine Dye",332,438,116,23,1,24,55,1
"A.J. Pierzynski",303,357,93,18,0,10,30,1
"Josh Fields",332,460,118,25,2,22,75,1
"Carlos Quentin",349,317,83,20,1,11,60,1
"Danny Richar",314,534,139,30,4,15,70,2

This is the bench:

"Jerry Owens",319,385,102,14,2,3,75,1
"Juan Uribe",298,421,105,21,1,16,60,1
"Pablo Ozuna",304,97,26,4,1,1,75,1
"Toby Hall",279,153,37,7,0,3,25,1

SP is using ZIPS (I don't have pecota pitching):

Vazquez
MB
Contreras
Floyd
Danks

AVG W-L is 77-85, with H=85, L=69
RS AVG = 747, H=812, L=678(gags)
RA AVG = 787, H=843, L=767

Interesting is that the first 11 games set the tone for the season, even in a simulation. If they get through the first 11 going at least 6-5 then the season goes ok, if they get whacked early then the whole season goes in the shitter.

"Swisher's versatility allows us to give Owens a chance to come back and show he's ready to be that sparkplug like Podsednik was a couple of years ago." - KW

by ChicagoPete on Feb 9, 2008 8:14 AM CST 0 recs