PDA

View Full Version : ZIP Projections


Dadawg_77
01-20-2005, 11:16 AM
Baseball Think Factory, has their ZIPS Projections for the 2005 season. While not as good as PECOTA system, ZIPS has a decent track record. For you fantasy players and anyone else it is available here.
http://www.baseballthinkfactory.org/szymborski/

ZIPS has the Sox pitching staff posting a 5.09 ERA, which is worse then Sox 4.91 which was 25th. Since they project the entire 40 man Roster, the pitcher who I used to come up to the 5.09 number are Marte (3.73), Buehrle (17-11, 4.04), El Duque (6-5, 4.08), Shingo (4.19), Politte (4.19), Garcia (14-11,4.39), Adkins (4.95), Vizcaino (4.96), Garland (12-12, 5.03), Hermanson (5.14), Cotts (5.27) and Contreras (9-12, 5.56).

As for hitting, ZIPS has Scott Podsednik hitting .271 and .339 OBP. While better then what I am expecting, still not good for a leadoff man. ZIPS is projecting a major power drop off from Thomas and Konerko. Their ISO Power last year was .292 and .258 respectively. ZIPS has it at .227 and .221. ISO power is SLG - AVG. Also ZiPS project the Sox to score 807 runs.

Here is the list:

Konerko .276/.352/.497
Podsednik .271/.339/.396
Thomas .242/.378/.469
Crede .258/.313/.462
Dye .262/.333/.462
AJ .292/.338/.447
Rowand .276/.329/.491
Uribe .267/.314/.443
Gload .298/.339/.468
Everett .258/.331/.447
Harris .260/.335/.341
Perez .272/.311/.382
Davis .233/.291/.380.

Palehose13
01-20-2005, 11:25 AM
:hawk

"You're what hurts?"


Well, with this information the Sox might as well not even play this season. :rolleyes:

Flight #24
01-20-2005, 11:27 AM
Baseball Think Factory, has their ZIPS Projections for the 2005 season. While not as good as PECOTA system, ZIPS has a decent track record. For you fantasy players and anyone else it is available here.
http://www.baseballthinkfactory.org/szymborski/

ZIPS has the Sox pitching staff posting a 5.09 ERA, which is worse then Sox 4.91 which was 25th. Since they project the entire 40 man Roster, the pitcher who I used to come up to the 5.09 number are Marte (3.73), Buehrle (17-11, 4.04), El Duque (6-5, 4.08), Shingo (4.19), Politte (4.19), Garcia (14-11,4.39), Adkins (4.95), Vizcaino (4.96), Garland (12-12, 5.03), Hermanson (5.14), Cotts (5.27) and Contreras (9-12, 5.56).

As for hitting, ZIPS has Scott Podsednik hitting .271 and .339 OBP. While better then what I am expecting, still not good for a leadoff man. ZIPS is projecting a major power drop off from Thomas and Konerko. Their ISO Power last year was .292 and .258 respectively. ZIPS has it at .227 and .221. ISO power is SLG - AVG. Also ZiPS project the Sox to score 807 runs.

Here is the list:

Konerko .276/.352/.497
Podsednik .271/.339/.396
Thomas .242/.378/.469
Crede .258/.313/.462
Dye .262/.333/.462
AJ .292/.338/.447
Rowand .276/.329/.491
Uribe .267/.314/.443
Gload .298/.339/.468
Everett .258/.331/.447
Harris .260/.335/.341
Perez .272/.311/.382
Davis .233/.291/.380.

Wait a minute, something doesn't appear to make a whole lot of sense here. Ther only pitchers higher then the team average of 5.09 that you cite are Contreras, Hermanson & Cotts, all within about a half-run of the avg. Everyone else is less than 5.09, and the high guys should be more than offset by the 40-odd starts of Buehrle/Hernandez who post an ERA more than a run better than the average.

It seems mathematically impossible unless Contreras, Hermanson, & Cotts are pitching a ridiculous # of IP (i.e. way more than Buehrle/Hernandez combined).

WhiteSoxFan84
01-20-2005, 11:30 AM
Baseball Think Factory, has their ZIPS Projections for the 2005 season. While not as good as PECOTA system, ZIPS has a decent track record. For you fantasy players and anyone else it is available here.
http://www.baseballthinkfactory.org/szymborski/

ZIPS has the Sox pitching staff posting a 5.09 ERA, which is worse then Sox 4.91 which was 25th. Since they project the entire 40 man Roster, the pitcher who I used to come up to the 5.09 number are Marte (3.73), Buehrle (17-11, 4.04), El Duque (6-5, 4.08), Shingo (4.19), Politte (4.19), Garcia (14-11,4.39), Adkins (4.95), Vizcaino (4.96), Garland (12-12, 5.03), Hermanson (5.14), Cotts (5.27) and Contreras (9-12, 5.56).

As for hitting, ZIPS has Scott Podsednik hitting .271 and .339 OBP. While better then what I am expecting, still not good for a leadoff man. ZIPS is projecting a major power drop off from Thomas and Konerko. Their ISO Power last year was .292 and .258 respectively. ZIPS has it at .227 and .221. ISO power is SLG - AVG. Also ZiPS project the Sox to score 807 runs.

Here is the list:

Konerko .276/.352/.497
Podsednik .271/.339/.396
Thomas .242/.378/.469
Crede .258/.313/.462
Dye .262/.333/.462
AJ .292/.338/.447
Rowand .276/.329/.491
Uribe .267/.314/.443
Gload .298/.339/.468
Everett .258/.331/.447
Harris .260/.335/.341
Perez .272/.311/.382
Davis .233/.291/.380.

how'd u get that team era with no innings pitched per player?

Tekijawa
01-20-2005, 11:33 AM
no 300 hitters? this could be a long season if this it true?!?!?

Flight #24
01-20-2005, 11:33 AM
Baseball Think Factory, has their ZIPS Projections for the 2005 season. While not as good as PECOTA system, ZIPS has a decent track record. For you fantasy players and anyone else it is available here.
http://www.baseballthinkfactory.org/szymborski/

ZIPS has the Sox pitching staff posting a 5.09 ERA, which is worse then Sox 4.91 which was 25th. Since they project the entire 40 man Roster, the pitcher who I used to come up to the 5.09 number are Marte (3.73), Buehrle (17-11, 4.04), El Duque (6-5, 4.08), Shingo (4.19), Politte (4.19), Garcia (14-11,4.39), Adkins (4.95), Vizcaino (4.96), Garland (12-12, 5.03), Hermanson (5.14), Cotts (5.27) and Contreras (9-12, 5.56).



If you use a pitching staff of: Buehrle, Hernandez+Diaz, Garcia, Garland, Contreras, Marte, Takatsu, Politte, Hermanson, Cotts, Vizcaino, the ERA they project comes out to 4.66. That's just summing their projected totals and using (ER/IP)*9 = ERA.

EDIT: 4.66 would have ranked them 6th in AL last year. Also, using the staff DD77 outlined (subbing Adkins for Diaz), the ERA would be 4.65. However, since they have Hernandez making only 21 starts, I figured the "extra" pitcher would more likely be a starter, not a reliever.

Palehose13
01-20-2005, 11:36 AM
no 300 hitters? this could be a long season if this it true?!?!?

Um...Thomas is projected to hit a .242, but Crede a .258?!?!?! That right there tells me this is crap. I want to remember this thread at the end of the year and see how right it is. Cause right now:

:bs:

voodoochile
01-20-2005, 11:40 AM
Um...Thomas is projected to hit a .242, but Crede a .258?!?!?! That right there tells me this is crap. I want to remember this thread at the end of the year and see how right it is. Cause right now:

:bs:

Who are they projecting to make up the starts that El Duque is clearly missing? 6-5 implies he is injured for most of the year (roughly 60%) yet I see no 6th starter projections to make up for it.

I agree about the Thomas projections. I would like to know what they based that on...

Mickster
01-20-2005, 11:47 AM
no 300 hitters? this could be a long season if this it true?!?!?

Are you serious? Do you actually believe this crap?

Flight #24
01-20-2005, 11:52 AM
Um...Thomas is projected to hit a .242, but Crede a .258?!?!?! That right there tells me this is crap. I want to remember this thread at the end of the year and see how right it is. Cause right now:

:bs:

Some other tidbits from this "analysis"....

Jon Adkins makes 16 starts, Neal Cotts 15, Hermanson 14. one of those guys could be doing that in the minors, but that would still leave about 30 starts form 2 guys who are almost guaranteed to be chained to the bullpen....They also have the combined starts from Buehrle, Hernandez, Garcia, Garland, Contreras = 147. So there's only 15 more to go around, unless they're including projected playoff starts in the totals.

PH13, I think you have it right.....

jabrch
01-20-2005, 11:56 AM
ZIPS - DIPS - Potato Chips...


what a pile of mokeypoo. Hocus Pocus - dominocus - make my baseball season appear on a spreadsheet!!! Good thing baseball isn't played on spreadsheets huh? (It's actually played inside of TV sets).

Tekijawa
01-20-2005, 11:59 AM
Are you serious? Do you actually believe this crap?

No... I was just saying it would be a long year if this were true...

Palehose13
01-20-2005, 12:07 PM
(It's actually played inside of TV sets).

Wow. So Mini Me of Austin Powers fame must be accurate. Clones are 1/8 the size of the original. That's how they get all the little players inside the TV sets! :D:

Dadawg_77
01-20-2005, 12:35 PM
Um...Thomas is projected to hit a .242, but Crede a .258?!?!?! That right there tells me this is crap. I want to remember this thread at the end of the year and see how right it is. Cause right now:

:bs:

The projections are heavily based on three year projections and estimated effect of aging players. He might even be degraded by his injury. I am not exactly sure how they will turn out, but just Zips has a decent track record. Personally, I am waiting to see what PECOTA says. Oh, just realized I used RA not ERA by typing in the wrong column.

Dadawg_77
01-20-2005, 12:40 PM
Who are they projecting to make up the starts that El Duque is clearly missing? 6-5 implies he is injured for most of the year (roughly 60%) yet I see no 6th starter projections to make up for it.

I agree about the Thomas projections. I would like to know what they based that on...

I didn't look at games started since the numbers show 192 for the staff I outlined, but I rather looked at innings pitched. Which ZIPS is showing 1485 for the staff I outlined, 30 more then 162*9. The only real hard projections is the ratio stats, since counting stats (w-l) depend mainly on playing time which is very tough to predict. Ratio stats are easier to predict since you can say based on a players past history and past history of player like him, you can expect this level of production,

mweflen
01-20-2005, 12:44 PM
I do believe that stats can predict general trends when taken with age, but these stats as posted seem to only assume degredation with age. What about improvement with age? One would assume that Crede and Harris, for instance, would see marginal increases in average simply based on experience and coming into physical prime.

Dadawg_77
01-20-2005, 12:51 PM
Are you serious? Do you actually believe this crap?

So why is it a lie, because it doesn't conform with what you expect the Sox to produce? I don't know the methodology behind ZIPS, what I do know it has a decent track record for projecting what players will do in the upcoming season. Personally I think it is undervaluing the power production. Frank's average is about 20 points from where I expect it to be. As for .300 only Gload and Rowand are only hitters on 2005 Sox who hit over .300 last year. So it isn't too much of a reach to say the team may not have one this year.

ZiPS is only a piece of information nothing more, nothing else.

Brian26
01-20-2005, 12:57 PM
no 300 hitters? this could be a long season if this it true?!?!?

If this is true? Give me a break!

There's a great thing about baseball that happens every single year:

The games actually get played on the field. That's the whole point. You can play strat-o-matic baseball all winter, but it doesn't mean a darn thing. Players have to play the game, and I'm pretty confident we've got a good team this year.

santo=dorf
01-20-2005, 12:58 PM
Just curious, do these people keep track of how accurate they were with their predictions from previous seasons? :?:

maurice
01-20-2005, 02:08 PM
It seems mathematically impossible

This was my first impression also. I didn't run the numbers, but it seems that the team ERA would actually be pretty good -- assuming around 200 innings each for MB, FG, and JG; less than 80 each from Hermanson and Cotts; and less than 180 from Contreras. Could you explain this please, DD?

Dadawg_77
01-20-2005, 02:14 PM
Just curious, do these people keep track of how accurate they were with their predictions from previous seasons? :?:

Yes. http://www.baseballprospectus.com/article.php?articleid=2515
The ZiPS projections are Primer in the graphs.

Mickster
01-20-2005, 02:37 PM
So why is it a lie, because it doesn't conform with what you expect the Sox to produce? I don't know the methodology behind ZIPS, what I do know it has a decent track record for projecting what players will do in the upcoming season. Personally I think it is undervaluing the power production. Frank's average is about 20 points from where I expect it to be. As for .300 only Gload and Rowand are only hitters on 2005 Sox who hit over .300 last year. So it isn't too much of a reach to say the team may not have one this year.

ZiPS is only a piece of information nothing more, nothing else.

Well, Flight made an excellent point:

Some other tidbits from this "analysis"....

Jon Adkins makes 16 starts, Neal Cotts 15, Hermanson 14. one of those guys could be doing that in the minors, but that would still leave about 30 starts form 2 guys who are almost guaranteed to be chained to the bullpen....They also have the combined starts from Buehrle, Hernandez, Garcia, Garland, Contreras = 147. So there's only 15 more to go around, unless they're including projected playoff starts in the totals.

Care to explain how the Sox pitchers will end up starting 192 games? I have no problems with stats, but when an outfit is projecting the ERA of our pitching staff using numbers as they have above, it calls in question the results that they have obtained. The old saying...garbage in, garbage out.

Dadawg_77
01-20-2005, 02:44 PM
Well, Flight made an excellent point:



Care to explain how the Sox pitchers will end up starting 192 games? I have no problems with stats, but when an outfit is projecting the ERA of our pitching staff using numbers as they have above, it calls in question the results that they have obtained. The old saying...garbage in, garbage out.

It is hard to project playing time. What I would look at more then games started is the ratio stats of the players and apply those over the playing time you expect from the palyer.

From the creator of ZiPS, Dan Szymborski (http://www.baseballthinkfactory.org/files/member/2/)

Also, you get into the territory at which predicting playing time can detract from the usefulness of the results. Say you have a minor-league player like Jose Leon of the Orioles. What is more useful? A projection that he'll go 4-13 with 2 RBI for the season, or a full season line tempered with knowledge of his chances of making the team? I've opted for the latter choice, but nothing's going to please all parties.

http://www.baseballthinkfactory.org/files/main/article/dszymborski_2003-03-04_0/

Mickster
01-20-2005, 02:56 PM
It is hard to project playing time. What I would look at more then games started is the ratio stats of the players and apply those over the playing time you expect from the palyer.

From the creator of ZiPS, Dan Szymborski (http://www.baseballthinkfactory.org/files/member/2/)


Taken from the quote above....

.......I've opted for the latter choice, but nothing's going to please all parties.

Nuff said. I choose to believe his projections are crap.

Flight #24
01-20-2005, 03:32 PM
It is hard to project playing time. What I would look at more then games started is the ratio stats of the players and apply those over the playing time you expect from the palyer.



That's kind of what I guesstimated on to realize the 5.09 team ERA was off. Looking over their project IP, I'd say everything's pretty dead on with a couple of exceptions: 1)Cotts & Hermanson pitch way too many innings given that I doubt they'll start many (if any) games. 2)There's no "backup" starter for El Duque since he's projected to pitch only 119 IP. So I picked Diaz, but it could easily have been Grilli. The net is that you're reducing some IP at a high ERA and adding some at a slightly lower ERA, so it's not going to appreciably change the totals, which come in at a team ERA right around 4.6 - that's pretty solid, and significantly improved over 2004.

OEO Magglio
01-20-2005, 04:05 PM
I can make projections too, if you think the sox are going to have a bad year because of this, well......you're nuts.

Ol' No. 2
01-20-2005, 04:16 PM
That's kind of what I guesstimated on to realize the 5.09 team ERA was off. Looking over their project IP, I'd say everything's pretty dead on with a couple of exceptions: 1)Cotts & Hermanson pitch way too many innings given that I doubt they'll start many (if any) games. 2)There's no "backup" starter for El Duque since he's projected to pitch only 119 IP. So I picked Diaz, but it could easily have been Grilli. The net is that you're reducing some IP at a high ERA and adding some at a slightly lower ERA, so it's not going to appreciably change the totals, which come in at a team ERA right around 4.6 - that's pretty solid, and significantly improved over 2004.The problem with all these projections is that they never account for the uncertainty in the projection. Hernandez has pitched 141, 214, 196, 94, 146 and 85 IP over his career. Anybody with a calculator can punch up the numbers and "project" 119 IP next year. But it's BS, because the variability is so large. That's 119 plus or minus 50. What good is a projection like that?

Dadawg_77
01-20-2005, 04:18 PM
The problem with all these projections is that they never account for the uncertainty in the projection. Hernandez has pitched 141, 214, 196, 94, 146 and 85 IP over his career. Anybody with a calculator can punch up the numbers and "project" 119 IP next year. But it's BS, because the variability is so large. That's 119 plus or minus 50. What good is a projection like that?

Thats why I wouldn't use this for items like inning pitched but for playing time "independent" items.

Ol' No. 2
01-20-2005, 04:35 PM
Thats why I wouldn't use this for items like inning pitched but for playing time "independent" items.But it makes a huge difference if Hernandez starts 32 games or if he starts 20 and Grilli starts the other 12. Other statistics have the same problem. Players' performances vary a lot from year to year. How do you know whether he's going to have a good year or a poor one? If you'd projected the 2000 team you would probably have projected 81 wins (maybe less). They may be interesting to look at, but take them all with a LARGE grain of salt.

Mickster
01-20-2005, 04:45 PM
But it makes a huge difference if Hernandez starts 32 games or if he starts 20 and Grilli starts the other 12. Other statistics have the same problem. Players' performances vary a lot from year to year. How do you know whether he's going to have a good year or a poor one? If you'd projected the 2000 team you would probably have projected 81 wins (maybe less). They may be interesting to look at, but take them all with a LARGE grain of salt.

A grain? Dont you mean:

http://www.cdc.gov/nccdphp/dnpa/immpact/images/salt.jpg

Flight #24
01-20-2005, 04:52 PM
But it makes a huge difference if Hernandez starts 32 games or if he starts 20 and Grilli starts the other 12. Other statistics have the same problem. Players' performances vary a lot from year to year. How do you know whether he's going to have a good year or a poor one? If you'd projected the 2000 team you would probably have projected 81 wins (maybe less). They may be interesting to look at, but take them all with a LARGE grain of salt.

Look at it from this perspective: Their assumptions are IMO fairly pessimistic: Hernandez with 20 starts, Contreras with a 5.56 ERA, Garland with a 5.03ERA. And the Sox team ERA still improves from 4.91 to 4.61. If you get 32 starts from El Duque, or an ERA even just around 5 from Contreras, the team's going to have a top pitching staff.

When your staff likely ranges from good to great, that's a great change from poor to average in 2004.

FWIW - using their #s for the hitters: We get 171HR from the 9 likely starters: Pods, ARow, Frank, Koney, Dye, Uribe, AJP, Crede, Harris. Thus if you can get 29HR from the bench, you're at 200HR, which would be top 10 in MLB.

santo=dorf
01-20-2005, 05:28 PM
Yes. http://www.baseballprospectus.com/article.php?articleid=2515
The ZiPS projections are Primer in the graphs.

I'm a little confused. Was that article just comparing ZiPS porjections to other programs? I want to see how their numbers compare to the actual numbers for each player in the MLB

From the article;

Take these results with a grain of salt; one year's worth of data may not be enough to establish the superiority of one projection system over another, especially in the hitting categories, where the results were tightly bunched. But for the time being at least, it looks like PECOTA walks the walk.

Dadawg_77
01-20-2005, 05:34 PM
I'm a little confused. Was that article just comparing ZiPS porjections to other programs? I want to see how their numbers compare to the actual numbers for each player in the MLB

From the article;

The article was comparing the four projection system to each other by their error of projection of overall MLB numbers. So the system with the lower error is the better system. The actual MLB number are show under the ACTUAL heading. I haven't seen anything that goes that in depth of comparing each player to a projection.

Ol' No. 2
01-20-2005, 05:53 PM
Yes. http://www.baseballprospectus.com/article.php?articleid=2515
The ZiPS projections are Primer in the graphs.The proof of the pudding is in the eating. Comparing their projected OPS with the actual gave a root mean squared error of about 80 pts. So they were able to predict most players' OPS to within a 160 pt window? That doesn't seem too impressive. They predicted ERA to a RMS error of 1.0. A 2.0-run window? How hard can that be? What they should have included in their comparison was projecting just by averaging a player's last three years stats. I'll bet it would have done just as well.

TheTwinsMustDie
01-20-2005, 06:02 PM
The proof of the pudding is in the eating. Comparing their projected OPS with the actual gave a root mean squared error of about 80 pts. So they were able to predict most players' OPS to within a 160 pt window? That doesn't seem too impressive. They predicted ERA to a RMS error of 1.0. A 2.0-run window? How hard can that be? What they should have included in their comparison was projecting just by averaging a player's last three years stats. I'll bet it would have done just as well.


:LTP Math hard. Joe no like math. Joe swing hard in case Joe hit something.

Whitesox029
01-20-2005, 06:07 PM
It seems mathematically impossible unless Contreras, Hermanson, & Cotts are pitching a ridiculous # of IP (i.e. way more than Buehrle/Hernandez combined).
Besides that Buehrle will in all probability throw more innings than Hermanson and Cotts combined.

MRKARNO
01-20-2005, 07:07 PM
The projections are heavily based on three year projections and estimated effect of aging players. He might even be degraded by his injury. I am not exactly sure how they will turn out, but just Zips has a decent track record. Personally, I am waiting to see what PECOTA says. Oh, just realized I used RA not ERA by typing in the wrong column.

IIRC ZIPS was pretty pessimistic about the Sox last year. I'd put more stock in PECOTA which can be a bit pessimistic at times, but it's a better tool for looking at trends than the actual stats.

surfdudes
01-20-2005, 07:26 PM
Thomas bats .242 if they stab out his eye with a pencil and make him wear leg shackles. These stats are screwy.

soxwon
01-20-2005, 07:48 PM
These Guys Know "zip" About Baseball.

jordan23ventura
01-20-2005, 11:10 PM
Baseball Think Factory, has their ZIPS Projections for the 2005 season. While not as good as PECOTA system, ZIPS has a decent track record. For you fantasy players and anyone else it is available here.
http://www.baseballthinkfactory.org/szymborski/

ZIPS has the Sox pitching staff posting a 5.09 ERA, which is worse then Sox 4.91 which was 25th. Since they project the entire 40 man Roster, the pitcher who I used to come up to the 5.09 number are Marte (3.73), Buehrle (17-11, 4.04), El Duque (6-5, 4.08), Shingo (4.19), Politte (4.19), Garcia (14-11,4.39), Adkins (4.95), Vizcaino (4.96), Garland (12-12, 5.03), Hermanson (5.14), Cotts (5.27) and Contreras (9-12, 5.56).

As for hitting, ZIPS has Scott Podsednik hitting .271 and .339 OBP. While better then what I am expecting, still not good for a leadoff man. ZIPS is projecting a major power (http://69.42.87.196/v40/ezlclk.fcgi?id=3711) drop off from Thomas and Konerko. Their ISO Power last year was .292 and .258 respectively. ZIPS has it at .227 and .221. ISO power is SLG - AVG. Also ZiPS project the Sox to score 807 runs.

Here is the list:

Konerko .276/.352/.497
Podsednik .271/.339/.396
Thomas .242/.378/.469
Crede .258/.313/.462
Dye .262/.333/.462
AJ .292/.338/.447
Rowand .276/.329/.491
Uribe .267/.314/.443
Gload .298/.339/.468
Everett .258/.331/.447
Harris .260/.335/.341
Perez .272/.311/.382
Davis .233/.291/.380.

^ That's why it makes no sense to pay attention to stat heads.

Frank Thomas hitting .242?? LMAO. More ridiculous garbage.

jordan23ventura
01-20-2005, 11:12 PM
:LTP Math hard (http://69.42.87.196/v40/ezlclk.fcgi?id=3876). Joe no like math. Joe swing hard in case Joe hit something.

LMAO.

He does look like a caveman, doesn't he?