http://espn.go.com/blog/keith-law/insider/post?id=5068

 

2. Byron Buxton - might not hit immediately, but baserunning + defense will provide tons of value

 

4. Jose Berrios - Right now, better than Milone, perhaps better than Duffey as well

 

12. Max Kepler - Thinks Rosario is more of an extra OFer in the long run w/ obps in the .280s. Could have 400+ at bats in ML this year

 

Others to Watch:

Jorge Polanco

Top pitcher and hitter in the league, not bad.

 

I thought Law (and Buster Olney) made an interesting point on their podcast - despite some of the complaints we hear on the board and elsewhere in local media - the Twins really are among the most aggressive teams in getting prospects to the majors and in not worrying about service time. I wonder if that will make a difference with Berrios.

 

Top pitcher and hitter in the league, not bad.

 

I thought Law (and Buster Olney) made an interesting point on their podcast - despite some of the complaints we hear on the board and elsewhere in local media - the Twins really are among the most aggressive teams in getting prospects to the majors and in not worrying about service time. I wonder if that will make a difference with Berrios.

 

I'm all for getting Berrios in the rotation quickly, but I don't really see the point (and I don't think there's any chance it happens) of putting him on the Opening Day roster.  You don't even need a 5th starter more than once or twice in the first few weeks of the season.  I would wait until late April or early May to promote him. 

 

I'm all for getting Berrios in the rotation quickly, but I don't really see the point (and I don't think there's any chance it happens) of putting him on the Opening Day roster.  You don't even need a 5th starter more than once or twice in the first few weeks of the season.  I would wait until late April or early May to promote him. 

 

I would too - to manage innings as much as save future money.

His chat just completed, several Twins' questions

 

He'd play Buxton right now, even if the bat isn't ready, for the defense alone.

 

He'd play Buxton right now, even if the bat isn't ready, for the defense alone.

And yet Fangraphs somehow has Buxton projected to provide negative value on defense.

 

I thought Law (and Buster Olney) made an interesting point on their podcast - despite some of the complaints we hear on the board and elsewhere in local media - the Twins really are among the most aggressive teams in getting prospects to the majors and in not worrying about service time. I wonder if that will make a difference with Berrios.

This is true.  The greater fear is probably the same one holding back Berrios this spring (and arguably did last fall), that the Twins are having trouble sorting through their pile of lesser starters.

 

If for some reason we don't see Berrios until June despite good health and performance from him, that will be the issue, not service time.

 

And yet Fangraphs somehow has Buxton projected to provide negative value on defense.

Only marginally, -0.7 runs (not wins) over a full season compared to a league average CF.  Maybe 7% of a single win?  It's hard to create a system to project defense at all, let alone for such an inexperienced player.  I don't blame Fangraphs for not being able to capture that in a quantitative fashion.  Besides, they have Sano as a league average RF so that probably balances things out.

Klaw doesn't seem to be much of a fan of Eddie Rosario based on the two comments on him.

 

He also said Park should be the Twins starting first basemen... Not sure if there's anything to be read into that or if he just put mistakenly put first base instead of DH.

 

Top pitcher and hitter in the league, not bad.

 

I thought Law (and Buster Olney) made an interesting point on their podcast - despite some of the complaints we hear on the board and elsewhere in local media - the Twins really are among the most aggressive teams in getting prospects to the majors and in not worrying about service time. I wonder if that will make a difference with Berrios.

 

When they're talented, the Twins aren't afraid to push them. They handle them all on a case-by-case situation. 

I'm all for getting Berrios in the rotation quickly, but I don't really see the point (and I don't think there's any chance it happens) of putting him on the Opening Day roster.  You don't even need a 5th starter more than once or twice in the first few weeks of the season.  I would wait until late April or early May to promote him.

Berrios better not be a 5th starter!

 

Berrios better not be a 5th starter!

 

What does that mean?

 

Here is an interesting comparison:

 

Player A: 3.25 ERA, 1.255 WHIP, 9.6 K/9, 3.45 K/BB
Player B: 3.22 ERA, 1.112 WHIP, 8.2 K/9, 3.33 K/BB

 

Career AA stats.  Pretty similar.  No?

 

Player B is Berrios.

Player A is another current Twin.  Care to guess who?

A is Milone in Rochester?

I'm all for getting Berrios in the rotation quickly, but I don't really see the point (and I don't think there's any chance it happens) of putting him on the Opening Day roster. You don't even need a 5th starter more than once or twice in the first few weeks of the season. I would wait until late April or early May to promote him.

 

The Twins only have 3 off days in April (and 3 in May.)

 

They'll break camp with five starters, like most everybody, and use them, like most everybody.

 

The days of skipping fifth starters in April are virtually gone.

 

But I agree I doubt they keep Berrios when they break camp.

 

The Twins only have 3 off days in April (and 3 in May.)

They'll break camp with five starters, like most everybody, and use them, like most everybody.

The days of skipping fifth starters in April are virtually gone.

But I agree I doubt they keep Berrios when they break camp.

 

They have 2 days off in the 1st 2 weeks. The 1st time they would need a 5th starter is Monday April 18th.  Last year Kris Bryant for Service time issues was promoted on April 17th (I don't know the date it would be this year for the extra year of control).  

 

Of course they are going to break camp with 5 starters.  My point is, they would be foolish to put Berrios on the Opening Day roster when at this point they can so easily gain the extra year. 

 

A is Milone in Rochester?

 

No.  Those are career AA numbers of Jose Berrios compared to career AA numbers of Ricky Nolasco

 

Only marginally, -0.7 runs (not wins) over a full season compared to a league average CF.  Maybe 7% of a single win?  It's hard to create a system to project defense at all, let alone for such an inexperienced player.  I don't blame Fangraphs for not being able to capture that in a quantitative fashion.  Besides, they have Sano as a league average RF so that probably balances things out.

I don't blame them for having difficulty projecting players - particularly young ones - but I do blame posters on this board for putting too much stock in them.

 

Buxton having a negative defensive value is a great example of a projection veering off the rails. I think they have a lot of value in the aggregate but not a lot on an individual level.

 

No.  Those are career AA numbers of Jose Berrios compared to career AA numbers of Ricky Nolasco

Boo this man!

 

Just once I'd like to have a thread where we talk about projections, w/o the same 4 people saying math is dumb.....that would be nice.

I think it's going to be hard, Mike.  The math is unreliable - as you noted - yet people use the math to suggest X, others are going to point out the flaws.  We get the same problem on threads that use base runs or pyth w/l or WAR - people use the stat out of context or without stating the strong caveats that are needed and others are going to push back.  Saying "the math!" isn't going to fix that when the math isn't used correctly or reliably. 

 

I think it's going to be hard, Mike.  The math is unreliable - as you noted - yet people use the math to suggest X, others are going to point out the flaws.  We get the same problem on threads that use base runs or pyth w/l or WAR - people use the stat out of context or without stating the strong caveats that are needed and others are going to push back.  Saying "the math!" isn't going to fix that when the math isn't used correctly or reliably. 

 

So, every thread I should point out that the eye test is wrong?

 

I guess a better question is, why do the math haters think the math lovers believe the numbers with gospel, when MOST of us point out it is the most likely outcome of the outcomes, but is unlikely to be the outcome? Why do the math haters assume the math lovers believe that IS the number, when we all say it is a number, with error bars?

 

Whatever, if people enjoy typing the same thing in every thread where math is used, go for it. This is all for entertainment purposes anyway.

Edited February 26, 201610 yr by Mike Sixel

 

Math can easily become a 'garbage in, garbage out' activity. Assumptions need to be examined minutely.

 

Indeed. The debate on analytics is basically over, to me the next big step from the media and the stronger advocates will be to take a step back and better explain the limitations.

 

Drawing really strong conclusions from clearly flawed numbers is going to result in pushback and (wrongly) throw much of the rest of the data into question.

 

Indeed. The debate on analytics is basically over, to me the next big step from the media and the stronger advocates will be to take a step back and better explain the limitations.

 

Drawing really strong conclusions from clearly flawed numbers is going to result in pushback and (wrongly) throw much of the rest of the data into question.

 

What's your alternative? 

 

So, every thread I should point out that the eye test is wrong?

 

I guess a better question is, why do the math haters think the math lovers believe the numbers with gospel, when MOST of us point out it is the most likely outcome of the outcomes, but is unlikely to be the outcome? Why do the math haters assume the math lovers believe that IS the number, when we all say it is a number, with error bars?

 

Whatever, if people enjoy typing the same thing in every thread where math is used, go for it. This is all for entertainment purposes anyway.

 

I would say when the eye test is invoked and clearly ridiculous that your first point is taken care of plenty well.

 

If analytics are used well I don't think you see much pushback.

 

What's your alternative? 

 

Using analytics better. Being more humble in conclusions and emphasizing the limitations.

 

Probably depends on the individual case - if the math says Player A is going to be worth 1 WAR next year b/c he plays spectacularly bad defense, I think it's pretty fair to point out the math on defensive WAR is pretty bad.  

 

But taking the math as gospel - as many seem to - gets pretty annoying.  Last year the math people killed TD by telling us the Twins were over performing and they weren't a .500 team.  That got old, real fast (and, of course, didn't happen).  

Because the outcome was different than predicted, the model is wrong? Might as well just shut down everything based on statistics right now.

 

that logic, btw, is pretty much the same logic that predicted Danny Santana wouldn't repeat his performance, does that prove the model was good?

Edited February 26, 201610 yr by Mike Sixel

 

Using analytics better. Being more humble in conclusions and emphasizing the limitations.

 

 

so, every post that quotes WAR should have a footnote about it's limitations? 

 

so, every post that quotes WAR should have a footnote about it's limitations? 

 

Or at least a footnote acknowledging it is a silly stat to take literally.

 

Or at least a footnote acknowledging it is a silly stat to take literally.

 

What does "literally" mean, here?

 

If you mean as a 100% accurate, with no error bars, number, that tells you with certainty what will happen.......who believes that?

Edited February 26, 201610 yr by Mike Sixel

 

What does "literally" mean, here?

 

If you mean as a 100% accurate, with no error bars, number, that tells you with certainty what will happen.......who believes that?

 

Many more people than you apparently want to admit.

 

See, statements like that don't really help your cause.  I'm not dumb.  Most of us aren't.  You aren't the only person who uses or understands the limits of advanced stats.  But condescending comments like that just make most of us roll our eyes or mark down - "Mike's a jerk" and move on.  

No one is suggesting that because an outcome was different than predicted, the model was wrong (although it could be, depending on the model).  

 

What did this mean, then, since I didn't understand it in the context of is math useful:

 

Last year the math people killed TD by telling us the Twins were over performing and they weren't a .500 team.  That got old, real fast (and, of course, didn't happen).