Angels ZiPs projections are out

[Oz27]

Here is the link to the latest release of individual player projections, this one the FanGraphs ZiPS system - https://t.co/saYmNLrls7

So, firstly the good...

Trout: 9 WAR, 169 OPS+

Pujols 1.7 WAR, 114 OPS+ (this one feels a little optimistic)

Calhoun 3.0 WAR, 113 OPS+

Simmons 3.3 WAR, 87 OPS+

Bedrosian 119 ERA+

Kirby Yates 110 ERA+

 

Now the 'in the middle'...

Shoemaker 101 ERA+, 2 WAR

Skaggs 100 ERA+, 1.3 WAR

Richards 106 ERA+, 1.9 WAR

Maybin 97 OPS+, 1.1 WAR

Cron 111 OPS+, 1.1 WAR

Escobar 99 OPS+, 1.4 WAR (it seems unlikely he puts up a 1.4 WAR if he's a below average hitter, though)

Marte 101 OPS+, 0.5 WAR

Street 103 ERA+

Bailey 101 ERA+

 

And the bad...

Espinosa 76 OPS+ (!!!), but a 1.1 WAR

Perez 66 OPS+, 0.6 WAR

Nolasco 89 ERA+, 0.9 WAR

 

So what's good and what's bad? The starting outfield projects for 13 WAR and that is probably the best in baseball. The bullpen is projected to be worth 1 WAR and that is awful. We have five positions where our starters project to be below average (catcher, 1B, 2B, 3B, LF) while DH is on the bubble for that too. But the rotation doesn't look quite as terrible as it perhaps could.

[ettin]

Remember Oz that we have to caveat all of this by looking at the playing time ZiPs projects along with the position it thinks the player will be at. For instance Maybin is listed at 487 PA's (which I think will wind up being about 100 more than listed) and that he is a CF when we all know he will play LF (and have better defense at the position).

[AngelsLakersFan]

Honestly, these projections are pretty brutal. The system takes a conservative approach to expectations, and the conservative approach makes this look like a bottom feeder in the AL in 2017.

[AngelsLakersFan]

2 minutes ago, ettin said:

Remember Oz that we have to caveat all of this by looking at the playing time ZiPs projects along with the position it thinks the player will be at. For instance Maybin is listed at 487 PA's (which I think will wind up being about 100 more than listed) and that he is a CF when we all know he will play LF (and have better defense at the position).

It lists him a 487 because it is not confident in his ability to stay on the field for another 100 PAs. 

[tdawg87]

Seems pessimistic but fair. They've got Trout at his lowest OPS+ in 3 years, Cron regressing, Espinosa regressing. They also have Perez taking more at bats than Maldonado.

[ettin]

Just now, AngelsLakersFan said:

It lists him a 487 because it is not confident in his ability to stay on the field for another 100 PAs. 

Yes but that does impact the final WAR total which was my point.

[Oz27]

17 minutes ago, tdawg87 said:

Seems pessimistic but fair. They've got Trout at his lowest OPS+ in 3 years, Cron regressing, Espinosa regressing. They also have Perez taking more at bats than Maldonado.

Most project systems are a little pessimistic on Trout because he is such an extreme outlier. When PECOTA comes out in a month or so, Trout will be 1-2 wins lower than he should be and about 10 OPS+ points below what you'd expect.

As BP put it after PECOTA came out prior to the 2016 season: "It pegs the four-time earner and one-time winner of the AL MVP for 7.3 WARP and a .332 TAv, and beyond it being remarkable that a system could esteem a player so highly, I find it interesting how obviously and irrefutably those numbers sell him short. I mean these numbers sell Trout way, way short. That .332 TAv would be 20 points worse than his career mark, which is right about where he’s sat for the last two seasons. ... PECOTA underestimates Trout, because Trout has completed a practical transcendence of baseball." - http://www.baseballprospectus.com/article.php?articleid=28453

[Oz27]

17 minutes ago, ettin said:

Remember Oz that we have to caveat all of this by looking at the playing time ZiPs projects along with the position it thinks the player will be at. For instance Maybin is listed at 487 PA's (which I think will wind up being about 100 more than listed) and that he is a CF when we all know he will play LF (and have better defense at the position).

While you're correct with the defense bit (although the positional adjustment will limit the impact of that), if anything the 487 PA projection is a little optimistic. He has only made it past 400 in one of the past three seasons.

[Stradling]

I have issues with Cron being projected worse than last year. 

[Troll Daddy]

Worthless

[AngelsLakersFan]

10 minutes ago, Stradling said:

I have issues with Cron being projected worse than last year. 

Ya, I think they are just wrong. 

It probably has a lot to do with his poor performance after he returned from the broken hand. That put a pretty big dent into the improvement he had shown before that, which I think makes ZiPs more conservative.

[totdprods]

All these different projection models have been out long enough now to form a good sample-size. Has one proven to be the most consistently close to the mark?

[Oz27]

22 minutes ago, AngelsLakersFan said:

Ya, I think they are just wrong. 

It probably has a lot to do with his poor performance after he returned from the broken hand. That put a pretty big dent into the improvement he had shown before that, which I think makes ZiPs more conservative.

Cron is really projected to be much the same as he was in 2016 (115 OPS+ last year, 111 this year). The system uses DIPS and BABIP theory to forecast future BABIP for its projections so it's possible it doesn't believe a right handed hitter who hits fly balls at an above average rate will sustain a .300 BABIP. But that is me speculating on the reason and I could be wrong.

With these projections, there are always some which seem 'wrong'. There are some you look at and think they will obviously be better or worse than that projection. But if you took a random sample of 20 (or 40 or 100 or 200) player projections and guessed whether they'd be better or worse than that projection, you're unlikely to be more accurate than a coin flip would be. Basically, the projection systems are so much better at forecasting performance than any of us are.

[ettin]

I'm sure a lot of you just looked at the details of the article but Dan does detail many reasons why his projections won't be perfect down in the disclaimer at the bottom of the report so you should read it to understand it better.

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=4&cad=rja&uact=8&ved=0ahUKEwjAovDsp4bRAhUT3mMKHSLzA-0QFggsMAM&url=http%3A%2F%2Fwww.baseballthinkfactory.org%2Fszymborski%2Fzipsqa.rtf&usg=AFQjCNGYXDwrdkyzU6IC8G6Fz2erXagfyw&bvm=bv.142059868,d.amc

http://www.beyondtheboxscore.com/2016/2/22/11079186/projections-marcel-pecota-zips-steamer-explained-guide-math-is-fun

[AngelsLakersFan]

24 minutes ago, Oz27 said:

Cron is really projected to be much the same as he was in 2016 (115 OPS+ last year, 111 this year). The system uses DIPS and BABIP theory to forecast future BABIP for its projections so it's possible it doesn't believe a right handed hitter who hits fly balls at an above average rate will sustain a .300 BABIP. But that is me speculating on the reason and I could be wrong.

With these projections, there are always some which seem 'wrong'. There are some you look at and think they will obviously be better or worse than that projection. But if you took a random sample of 20 (or 40 or 100 or 200) player projections and guessed whether they'd be better or worse than that projection, you're unlikely to be more accurate than a coin flip would be. Basically, the projection systems are so much better at forecasting performance than any of us are.

No doubt, but there are many places where people paying close enough attention can point to cases where the system may be off. I think players who were injured or missed time are always a good place to start. The truth is that no one really knows what is going to happen, as the article points out Tropiano and Heaney are our 5th and 3rd best starters per the projection, but who knew they would get hurt? Were those who were bullish on Richards and Heaney last year wrong?

[Hubs]

I expect better from nearly everyone on this list. If someone dug up the 2016 projections and see how much better or worse they are than what actually happened, I'd be interested to see it. 

Why do I expect better? I think the playing time calculations are way off. Also who is John Lamb? Is he 6th on the depth chart?

Yet back to the projections Mike Trout as he gets closer to his peak years, is going to take a step forward. Is that silly to project as he won the MVP on a losing team and is still going to get better? I think he could be the best OF in baseball, by himself. He could have a ridiculous 12 WAR season. Silly I know, but he could end up with a .335+ BA, while getting 40+ 2b, 10-15 3B, 40 HR, and 40 SB. That would be around 12 WAR. ( I use BR's calcs which but last season at 10.6). Barry Bonds 2001 and 2002 seasons were around that number with a negative fielding WAR. I know Bonds hit 73 HR, which Trout won't do, but he had 32 2b and 2 3B. I can see Mike getting close to the same number of XBH, say around 100 MAX (it will be difficult to pass that 107) but more likely 90-95. And he will have more SB 40 to 2, and surely a positive defensive value, whereas Bonds was negative that year in LF. Finally WAR takes into account position, and the positional WAR takes away points for Left Field, but adds to Center Field.

Calhoun, Simmons, Shoemaker, Skaggs, Richards, Cron, Maybin, etc. are all not projected for much playing time, and I know some are coming off injury, but when healthy, should be worth about 1 WAR more than projected. 

Calhoun can have a season in the high 3 WAR low 4, as will Simmons, Richards 

Shoemaker should be around 3.

Skaggs, Maybin, Cron should be between 2 and 3.

Pujols should be around 2-2.5.

I hope Nolasco, Espinosa, Chavez, and Escobar actually end up between 1 and 2 WAR.

Perez, Maldonado combined will be close to 3.

I expect them to have 2-4 WAR from the pen total, including whatever swingmen actually end up starting games. I also think our pen is better than illustrated here, if only because their projected workload will be less so that should help.

The Bench is net neutral, for every Marte there is a Ji-Man Choi.

Which puts them between 33-53 WAR with the players I've listed and better than that overall. Which as they had just over 30 WAR last year total, that is a major improvement. 

 

[Oz27]

54 minutes ago, Hubs said:

I expect better from nearly everyone on this list.

 

Fans of every team says this and the majority of them are wrong. There is no justification for expecting "better from nearly everyone" besides overwhelming bias.

54 minutes ago, Hubs said:

I think the playing time calculations are way off.

Uhh, they're based off a player's previous performance. We'd need to play something like a million games to have as many plate appearances at the projections suggest we would.

54 minutes ago, Hubs said:

If someone dug up the 2016 projections and see how much better or worse they are than what actually happened, I'd be interested to see it. 

 

Looking at our projected regulars heading into 2016, Trout, Cron and Calhoun beat their projections (measuring by WAR). Escobar performed exactly as projected. Simmons, Perez, Giavotella, Nava and Pujols were worse. Looking at pitchers (and using FIP) Shoemaker and Richards were better. Tropeano, Weaver, Skaggs and Santiago were worse.

You can look for yourself using this link for the projections - http://www.fangraphs.com/blogs/2016-zips-projections-los-angeles-angels/

And this one for actual performance - http://www.fangraphs.com/leaders.aspx?pos=all&stats=sta&lg=all&qual=0&type=8&season=2016&month=0&season1=2016&ind=0&team=1&rost=0&age=0&filter=&players=0

But yeah, it sure looks like you're criticizing the projections because they don't say what you want them to. 

[Inside Pitch]

3 hours ago, totdprods said:

All these different projection models have been out long enough now to form a good sample-size. Has one proven to be the most consistently close to the mark?

PECOTA has been the most consistently close in the last few years. ZIPS has probably been the closest over the largest sample size.   Thing is there are roadmaps to follow when taking these projections into consideration.  Guys coming off injuries or shortened seasons are wild cards, as are players at the young and old points of the age spectrum.  Younger players seem to have the wildest range of performance and unpredictability due to breakouts.  Pretty much all the projection models come closest when looking at guys with 5 plus years of MLB service time and 150+ game seasons.

[Hubs]

1 hour ago, Oz27 said:

Fans of every team says this and the majority of them are wrong. There is no justification for expecting "better from nearly everyone" besides overwhelming bias.

Uhh, they're based off a player's previous performance. We'd need to play something like a million games to have as many plate appearances at the projections suggest we would.

Looking at our projected regulars heading into 2016, Trout, Cron and Calhoun beat their projections (measuring by WAR). Escobar performed exactly as projected. Simmons, Perez, Giavotella, Nava and Pujols were worse. Looking at pitchers (and using FIP) Shoemaker and Richards were better. Tropeano, Weaver, Skaggs and Santiago were worse.

You can look for yourself using this link for the projections - http://www.fangraphs.com/blogs/2016-zips-projections-los-angeles-angels/

And this one for actual performance - http://www.fangraphs.com/leaders.aspx?pos=all&stats=sta&lg=all&qual=0&type=8&season=2016&month=0&season1=2016&ind=0&team=1&rost=0&age=0&filter=&players=0

But yeah, it sure looks like you're criticizing the projections because they don't say what you want them to. 

I'm not criticizing them because I don't like them I think projecting a guy to pitch 114 innings instead of the 207 he had in 2015 before he was injured is asinine. The guy had a bad injury one year and a freak injury another. In his years as a starter he's had 207 innings when he was healthy. He was on pace for similar numbers both years he was injured. Saying he's going to pitch 114 is saying he's going to be injured again, and that's just not true. They can't know that. 

[AngelsLakersFan]

11 minutes ago, Hubs said:

I'm not criticizing them because I don't like them I think projecting a guy to pitch 114 innings instead of the 207 he had in 2015 before he was injured is asinine. The guy had a bad injury one year and a freak injury another. In his years as a starter he's had 207 innings when he was healthy. He was on pace for similar numbers both years he was injured. Saying he's going to pitch 114 is saying he's going to be injured again, and that's just not true. They can't know that. 

They can't know that, but I'm about as optimistic that Richards is going to throw 207 innings as I am pessimistic that he is going to throw 21 innings.

[Docwaukee]

So keep this in mind (as Ettin mentioned).  

WAR is cumulative.  

they have the following for out pitchers:

Richards 114 - 2

Shoe 146.1 - 2

Nolasco 126.7 - 1

Skaggs 100 - 1

Chavez 134.1 - 1

Street - 41, Bedrosian - 61, Bailey - 42.1, Alvarez - 66, Guerra - 55.1

That's 888 innings.  Almost 600 short of how many innings we will actually pitch.  Or 40% less.  

But they've also got Jon Lamb getting 1 WAR in 124.2 ip, Vicente Campos 0.6 WAR in 109.1 ip, David Huff at 0.8 in 81ip, Wright 0.4 in 141ip, pounders 0.4 71.3ip, yates, 0.4 in 56.3ip, and 0.2 in 60ip by justin miller.  Meyer 88.3ip and 0.8 war

Just by their math, that's an additional almost 5 wins.  

It's disappointing that they include Heaney, Trop, Weaver, Wilson, and others who have no chance of pitching.  It just clutters the projections and makes them look sloppy as well as less well thought out.  

The key thing that it will never account for even the smallest breakout.  

If there is any value in this it's to look at the individual players and give you a regressed estimate of that one guy.  It's also base on 4yrs of estimates or 3 for younger/older players. 

There is very little weight toward potential.  It's numbers based on previous numbers.  

These are not projections as much as they are mathematical probabilities based on past performance and where that fits in with conglomerated data from other past performances.  It's regressed in every possible direction.  What would be helpful is if they provided confidence intervals or standard deviation possibilities.  That would at least lend some level of credence to what we are seeing.  

The avg of 1,000,000 and 2 is 500,001. The average of 500,002 and 500,000 is 500,001.  The final result is the same.  How you got there is vastly differently.  Some insight on that might be nice.  

[Oz27]

35 minutes ago, Hubs said:

I'm not criticizing them because I don't like them I think projecting a guy to pitch 114 innings instead of the 207 he had in 2015 before he was injured is asinine. The guy had a bad injury one year and a freak injury another. In his years as a starter he's had 207 innings when he was healthy. He was on pace for similar numbers both years he was injured. Saying he's going to pitch 114 is saying he's going to be injured again, and that's just not true. They can't know that. 

What? Forecasting durability based on previous durability is a pretty obvious thing to do, especially with pitchers. Projection systems would be worthless if they didn't do that. Past injuries are as good a predictor as we have of future injuries.

[Hubs]

12 minutes ago, Oz27 said:

What? Forecasting durability based on previous durability is a pretty obvious thing to do, especially with pitchers. Projection systems would be worthless if they didn't do that. Past injuries are as good a predictor as we have of future injuries.

No it's a logical fallacy. Just because something was one way does not mean it continues. We are much more likely to gets projected number of IP for Skaggs for instance by examining other left handers with similar deliveries or are a similar age who have had TJ surgery in their second year back than by giving a mean or mode average for his number of innings thrown the last three years. 

How many innings does a young lefty starter throw two plus years removed from TJ? 

Let's look at those guys.

Richards I know is a crapshoot because not many have had this procedure and we don't know what to expect. But he's either going to throw less than 50 or 180. That's way more likely than 114.

And it doesn't account for an off year or an outlier year or a breakthrough. 

I expect Simmons to be better with the bat again in the AL and hopefully Espinosa does as well. I don't think he will play if he has a +76 OPS. So that's irrational. 

I like PECOTA way more and I too wish they wouldn't give projections of a guy throwing 124 IP if he isn't likely to even start the season in the rotation. Lamb isn't even in the top 8 of starters (Richards, Shoemaker, Skaggs, Nolasco, Chavez, Meyer, Smith, Pounders)

And they do mess with projections when they've got guys like Wilson or Gia on there. 

[Second Base]

Purely from an overall perspective,  they're predicting....

Trout - .290/.410 30 DB 35 HR 25 SB.

Calhoun -.260/.330 30 DB 20 HR

Pujols - .260/.315 20 DB 30 HR

Cron .270/.310 25 DB 20 HR.

That seems pretty accurate actually. The only one that I think I really take exception to is Danny Espinosa. An Obp of .280 with 15 DB 15 HR and 7 SB? Even the most pessimistic among us knows he has more to offer than that. A fair projection would likely be a .310 OBP 20 HR and 15 SB.

 

[Docwaukee]

'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. '

 

A quote from the disclaimer.  

I think the people who generate these 'projections' understand way more about what they are than the people who read them then and assume they are predictions.