2015 Twins ZIPS projections (FanGraphs)

[Seth Stohs]

FanGraphs posted their 2015 Zips Projections for the Twins today. 

 

Like when I post my player projections, it's all a guess. They use numbers. I typically use numbers and gut feel, adn generally, I'm guessing, we'll both be right and wrong on around the same number of players. I like reading projections to see what they say, but they have to completely be taken with a grain of salt.

 

This one is always funny to me because it has Byron Buxton getting nearly 400 plate appearances, Aaron Hicks getting 500+, Shane Robinson getting over 400, and Jordan Schafer too. They have guys like Jason Kubel and Jason Bartlett and Virgil Vasquez and Kris Johnson listed. It's purely mathematical... I do enjoy seeing the player comps...

 

Twins fans should really enjoy seeing the player comp for JO Berrios is JD Durbin!!

 

Enjoy and comment...

[nytwinsfan]

The plate appearances can't be what Fangraphs actually expects. That can't be right. They must be assuming if the player gets called up and put in a starting position or something like that.

 

Relatedly, Jeff Sullivan's XBABIP which takes into account how fast a runner is and how many hard hit balls he hits, to predict BABIP.  From this, it looks like Santana, Vargas, Escobar, Suzuki, Hicks, and Mauer are due for regressions, but Dozier and Arcia are due for improvements.  Interestingly, Santana still has a very high expected BABIP, just not as high as his absurd .405 last season.  Plouffe is expected to improve very slightly.  Especially for some of the young players, there is probably more likely to be improvement in the hard hit balls, as well as in K and BB rates (which of course are not taken into account by BABIP) but still, some reason for skepticism about continued offensive output on the same level as last season.

 

http://www.fangraphs.com/fantasy/2014-xbabip-values/

[jimmer]

The plate appearances can't be what Fangraphs actually expects. That can't be right. They must be assuming if the player gets called up and put in a starting position or something like that.

 

Relatedly, Jeff Sullivan's XBABIP which takes into account how fast a runner is and how many hard hit balls he hits, to predict BABIP.  From this, it looks like Santana, Vargas, Escobar, Suzuki, Hicks, and Mauer are due for regressions, but Dozier and Arcia are due for improvements.  Interestingly, Santana still has a very high expected BABIP, just not as high as his absurd .405 last season.  Plouffe is expected to improve very slightly.  Especially for some of the young players, there is probably more likely to be improvement in the hard hit balls, as well as in K and BB rates (which of course are not taken into account by BABIP) but still, some reason for skepticism about continued offensive output on the same level as last season.

 

http://www.fangraphs.com/fantasy/2014-xbabip-values/

From the article: 'Disclaimer: ZiPS projections are computer-based projections of performance. Performances have not been allocated to predicted playing time in the majors — many of the players listed above are unlikely to play in the majors at all in 2014.'

 

And ZIPS isn't actually from Fangraphs.  The guy who creates ZIPs does weekly chats for them and occasionally an article, but he's not staff

[Dance with Disco Dan]

Yeesh.  Old Zips doesn't give us much to dream on.  The most surprising projection to me was Vargas with a sub-.400 slugging percentage.  I'd like to bet the over on that one.

[nytwinsfan]

About the only positive projection is for Hughes, where they expect him to stay relatively close to the BB and K rates from last year and to be by far the most valuable player on the team.

[nicksaviking]

The only thing funnier than putting much stock into predicting statistics for a player is putting stock into predicting statisticas for 40 of them and expecting anything resembling accuracy.

 

It's not surprising that these ZiPS or Steamer numbers ultimately end up in the ballpark, as predicting how a team will fare can also be done fairly accurately by any John Doe who follows baseball even without the help of numbers.  Most teams' final finish can be determined fairly accurately using simple common sense.

[jimmer]

Most teams' final finish can be determined fairly accurately using simple common sense.

That's an interesting opinion since I imagine we could look on this site now, in the past and the future, and see predictions for each of the last three seasons (and the upcoming season) and see all sorts of differences.  Some in high 60s, some in low 70s, some predicting us to be at .500, some predicting us to be playoff contenders..

 

Whose simple common sense are you referring to?

 

The best thing about these projection systems is that they take the fans' bias eyes out of the equation and really, these are just projections.  In the stat world, there is a bit of a difference between projections and predictions.   For example, Dan isn't predicting Buxton actually has that many MLB plate appearances.

[nicksaviking]

That's an interesting opinion since I imagine we could look on this site now, in the past and the future, and see predictions for each of the last three seasons (and the upcoming season) and see all sorts of differences.  Some in high 60s, some in low 70s, some predicting us to be at .500, some predicting us to be playoff contenders..

 

Whose simple common sense are you referring to?

 

 

I'd guess that most of the folks on this site would have pegged the Twins 2014 victory total within 6 wins of how they ended and it would have nearly been unanimous that everyone would have been within 10 wins.  In fact, I think most people would have predicted the records of every AL Central team similarly.

 

Whether it's computer or random guy using top of his head evaluations, most people/stats are going to give teams like the Padres and Cubs a bit of a boost due to acquistions and a team like the Braves a knock due to losses.  Teams like the Red Sox and Rangers will both get boosts because they failed to meet expectations last year but still have talent. 

 

For indivdual players, both humans and stats are going to predict players' improvement/decline based on age, team, ballpark, last year's production, batting order protection, change of position, injury history etc.  It's not rocket science, it's common sense.  ZiPS can break it down if they want to, but the general concensus isn't going to change much.  There will always be outliers, but for the most part, these are going to line up with human expectations.

[ashbury]

predicting how a team will fare can also be done fairly accurately by any John Doe who follows baseball even without the help of numbers.  Most teams' final finish can be determined fairly accurately using simple common sense.

Chief ran a pool here last season to guess over/under on a sports book's line of win totals for major league teams.  He could weigh in on the ability of us John Does.

 

_{^{/ disclaimer: I won.  And I wasn't very darn good.}}

[jimmer]

I'd guess that most of the folks on this site would have pegged the Twins 2014 victory total within 6 wins of how they ended and it would have nearly been unanimous that everyone would have been within 10 wins.  In fact, I think most people would have predicted the records of every AL Central team similarly.

 

Whether it's computer or random guy using top of his head evaluations, most people/stats are going to give teams like the Padres and Cubs a bit of a boost due to acquistions and a team like the Braves a knock due to losses.  Teams like the Red Sox and Rangers will both get boosts because they failed to meet expectations last year but still have talent. 

 

For indivdual players, both humans and stats are going to predict players' improvement/decline based on age, team, ballpark, last year's production, batting order protection, change of position, injury history etc.  It's not rocket science, it's common sense.  ZiPS can break it down if they want to, but the general concensus isn't going to change much.  There will always be outliers, but for the most part, these are going to line up with human expectations.

Guessing the win total within 10 wins is a 20 win spread (10 too many, 10 too few).  That's huge! So yeah anyone who guessed between 60 and 80 wins last year would have met the criteria but that's hardly an accomplishment. I wouldn't say someone who guessed 60 or someone who guessed 80 came close to being right, nor could we say their analysis of the team was roughly the same just because they were both within 10 games. A 12 win spread is also huge. (which is what guessing within 6 wins is).

 

So I guess we just disagree on the degree of accuracy that meets the standard you are setting. I doubt it takes hardly any baseball knowledge at all to do that, when you consider most of the time every team wins a minimum of 58 games (and last year the lowest was 64), that would normally only leave 46 games to play with.

 

IMO, a plus-minus 3 games is a much better indicator and I doubt we could get anywhere close to a consensus that would end up within plus-minus 3 games.

[USAFChief]

From the article: 'Disclaimer: ZiPS projections are computer-based projections of performance. Performances have not been allocated to predicted playing time in the majors — many of the players listed above are unlikely to play in the majors at all in 2014.'

 

 

One would think an article written in 2015 should be able to accurately predict the 2014 playing time of everybody.

 

Real darn accurately in fact. ;-)

[jimmer]

One would think an article written in 2015 should be able to accurately predict the 2014 playing time of everybody.

 

Real darn accurately in fact. ;-)

yeah, I think he cut and pasted parts of that disclaimer :-)

 

If you read his stuff, that disclaimer shows up a lot. I know I've read the last part of it multiple times.  It stands out because it cracks me up. It is: 'Finally, Szymborski will advise anyone against — and might karate chop anyone guilty of — merely adding up WAR totals on depth chart to produce projected team WAR.'

[jokin]

 

 From this, it looks like Santana, Vargas, Escobar, Suzuki, Hicks, and Mauer are due for regressions,

 

but Dozier and Arcia are due for improvements. 

http://www.fangraphs.com/fantasy/2014-xbabip-values/

 

"Regressions" is a term to characterize both improvements and declines towards the career means. 

[jokin]

 

 

Twins fans should really enjoy seeing the player comp for JO Berrios is JD Durbin!!

 

Enjoy and comment...

 

I thought that ZiPS broke some surprising news about Matt Guerrier moving to TNT in 2015:

 

 

 

Player IP K/9 BB/9 HR/9 ERA+ zWAR No. 1 Comp

 

 

Matt Guerrier 30.3 5.94 3.56 0.89 83 -0.2 Ernie Johnson

 

http://cdn.bleacherreport.net/images_root/slides/photos/000/972/959/109351494_display_image.jpg?1306714734

[jimmer]

most of the time every team wins a minimum of 58 games (and last year the lowest was 64), that would normally only leave 46 games to play with.

 

 

That should read, 'most of the time a team wins and losses 58 games (and last year the lowest win total was 64), that would normally only leave 46 games to play with. 

 

I'm sure most understood what I meant and it doesn't really change the point of that overall post.

[jokin]

 They have guys like Jason Bartlett  listed. It's purely mathematical... I do enjoy seeing the player comps...

 

Twins fans should really enjoy seeing the player comp for JO Berrios is JD Durbin!!

 

Enjoy and comment...

 

Hey, not only is the Team Chemist making a double-secret second comeback... the "math" says he's apparently going to be revitalizing his career in a big way (now that Gardy is no longer around and having him patrol the OF, apparently?)... the Next Reincarnation of an 8-time Gold Glover @ SS.

 

 

 

Player PA RC/27 OPS+ Def zWAR No.1 Comp

 

 

Jason Bartlett 225 2.9 62 -2 -0.2 Mark Belanger

[jimmer]

Here is a good article that evaluates how last years projections turned out. It's a good read, but long. If you skip to the end it says, 'Given all of the above we can decisively say that in 2014 ZiPS did the best job. It performed the best overall and in most of our subsets while never stumbling into the bottom half.  You can’t go wrong with any of these systems, but as we look ahead to 2015, I’ll be using ZiPS.'

 

http://www.hardballtimes.com/evaluating-the-2014-projection-systems/

[nytwinsfan]

"Regressions" is a term to characterize both improvements and declines towards the career means. 

 

Ok, sorry. I understand the idea of regression to the mean, but in common parlance regression is usually meant as a negative only. In the future I'll be sure to use only the technical terminology approved for this blog. :-)

[lightfoot789]

FANGRAGPH  "ZIP"  RANKINGS  FOR  TWINS

Adam Brett Walker II has so much upside
 

HRs = 1st* 
Runs - 3rd (behind only Santana & Dozier)
Hits = 9th
RBI = 1st*
ISO = 4th (behind Sano / Arcia / Plouffe)

DEF = League Average (0+) Buxton / Mauer/ Polanco / Turner better

 

K% = 2nd  Worst
BB% - 4.9% but better than (Hunter/Rosario & Santana)

 

Despite his contact flaws he manages to remain productive in terms of RUN PRODUCTION

 

Not a sabermetric fan in terms of predicting the future - but curious to know if these (ZIP stats) still mean Walker probably won't be productive at the MLB level?

[jimmer]

 curious to know if these (ZIP stats) still mean Walker probably won't be productive at the MLB level?

No, this isn't suggesting that at all.

[JustinCB]

I've never even heard of Heiker Meneses.

[jorgenswest]

I do know that the projection systems do much better than using last year's data to project performance on this year. The Twins acquire someone like Stauffer or Hunter or Santana and the forum is filled with data from their most recent performance in argument of whether they will be helpful to the Twins. it is more likely that their projections will be closer to reality than last year's numbers. Rarely are the projections used in the arguments.