Comments on: What the Good Book says: Yankee pitchers

By: Benjamin Kabak

Benjamin Kabak — Mon, 22 Feb 2010 19:54:25 +0000

You disagree with Mo’s projection. Either:

1. You don’t understand how PECOTA, on the whole, works.

2. The projection system is bad.

I know what I’m picking here.

By: Dan

Dan — Mon, 22 Feb 2010 19:52:54 +0000

Mo’s Projection means the projection system is probably worthless.

By: DCBX

DCBX — Mon, 22 Feb 2010 18:51:25 +0000

Chock full of win!

By: Jud

Jud — Mon, 22 Feb 2010 06:51:29 +0000

I’m willing to concede the point that the more pessimistic projections are more accurate. My point is that I believe that these projections are not a “valid projection”. In order for a projection to be valid in my mind, it needs to be similar in it’s structure to the real life. By “similar in structure”, I’m referring to the distribution of the results.
To take it to an extreme, both a projection that would have A-Rod with a WAR of 8 and a projection that would have Yuniesky Betancourt with a WAR of 8, are valid projections. One of them of course is reasonable, and the other extremely horrible, horrible – but valid.
A projection system with NO ONE with a WAR of 8 is not valid as it does not model real life MLB results.

By: Dalelama

Dalelama — Mon, 22 Feb 2010 03:27:05 +0000

The new stadium threw them off.

By: bexarama

bexarama — Sun, 21 Feb 2010 23:44:33 +0000

Am I up in arms? No, Lanny.

By: Accent Shallow

Accent Shallow — Sun, 21 Feb 2010 23:28:03 +0000

Oh, I'm sorry, did you ask me to link this article again? /one of my favorites

By: AndrewYF

AndrewYF — Sun, 21 Feb 2010 23:26:13 +0000

The numbers don’t work because the methods used are clearly wrong. When statistical analysis contains significant and obvious errors, then yes, it’s fairly worthless.

The game isn’t played on spreadsheets, but spreadsheets can predict with fair accuracy what’s going to happen, given past performance and historical trends. But if the spreadsheets are wrong, like BP’s are, then it’s akin to the game being played and runs not being counted arbitrarily. That’s what BP’s spreadsheets are projecting, currently, and that’s why everyone is ignoring them. Man, do they miss Nate Silver.

By: Joey H.

Joey H. — Sun, 21 Feb 2010 23:11:55 +0000

I get what you mean. But you get what I mean as well. Only 15 wins out of CC and 11 out of Burnett is a bit chincy. I dont expect a 25 win projection for Sabathia. Maybe 17. Do these projections take run support into account? If so, they really think the offense fell off a cliff.

By: Lanny

Lanny — Sun, 21 Feb 2010 23:11:41 +0000

The same people that say w-l don’t matter are up in arms over a statistical projection of w-l’s. Comedy