For those who don't want to tell the Providence Journal the names and addresses of the their last four sex partners:

email: .(JavaScript must be enabled to view this email address)

I set up an account for everyone: username .(JavaScript must be enabled to view this email address), password bselig.

Doh! Floyd, can we call it a tie?

No. I win.

Nyahh.

just out of curiosity, does anyone know the identity of the Ba-h-ston Globe sportswriter who called James' hiring "troubling"?

it ain't Ryan,unless he's done a 180 from what he said on the radio last week

Remmeber when that guy Voros used to post here? Those were good times.

I would guess that the Globe sportswriter who called James' hiring "troubling" was Dan Shaughnessy.

Can someone post the link to Shaughnessy's article?

Martone's a huge James fan and usually quite reliable. I'd be surprised if he flubbed that quote and less surprised if James just said the wrong thing in haste.

Let's forget about what Voros says and James says and what anyone says about $H (nonHR hits / ball in play) for the moment.

The fact is that 90% of all pitchers with at least 3000 BIP, have a $H that is within 5% of their teammates $H, for their careers.

So, when James says "pretty much constant", I think we can accept that it is descriptive enough of what I just said, without having to quote the numbers I just did.

One interesting thing is that James created DER, which is nothing more than 1 - $H. And in his 15 things, he's got a couple of important pitching/fielding references as well. James was very close to stumbling onto DIPS.

Anyone else impressed with the "15 things I learned..."? Me too. Anyone care to add what they've learned?

The day the Sox announced the hiring of Bill James, the Boston Channel 5 Sports guy (Mike Lynch?) expressed reservations. He mentioned Mike Gimbel and seemed to think Bill James was pretty much the same weirdo stat-head type. He wasn't the guy referred to in this article, but speaks to the mindset of some of the Boston Sports media.

Good point Mike.

Of all pitchers with at least 3000 BIP (758), 18 were at least 5% worse than their teammates, while 44 were at least 5% better.

That works out to 6% of them are better and 2% are worse. If the "bad guys" who didn't make the 3000 cut were allowed to come in, they would have probably been up to the same rate as the "good guys". So, if 6% are better, then maybe 6% are worse too. Again, the reader can decide if that is "small" or not.

For the statisticians out there, perhaps you can interpret this for me. Of all pitchers with at least 1 BIP, 60% were within 1 SD, if we assumed random variation. If there really was no control over the $H, we would expect 68.6% of them to fall within 1 SD. So, there is some control, and I'd just like to know how to interpret that 60% number.

Ben, feel free to email me if that is a better avenue. Let me tell you exactly, step-by-step, what I did, and then you tell me if I'm doing this right, or how I can do it better.

Greg Maddux (through 2000 I think) bad an estimated 10,414 BIP. His $H is .270, while his teammates' $H (weighted by the BIP of Maddux, year-by-year), is .278.

I took his teammates' .278 as the "true mean", and calculated that 1 SD = sqrt(.278 x .712 / 10414) = .0044

Maddux is .0080 better than his teammates. He is .0080/.0044 (1.8) SDs from the mean.

I did this for all pitchers, using their BIP as "n".

The results were that for all pitchers (whether I looked at n>0, or n>500, etc), the number of pitchers within 1 SD (as calculated above) was 60%. As I understand it, a truly random $H would have resulted in 68.6% pitchers.

Ok, now this is the extent of my statistical knowledge, more or less. Please feel free to correct me, and suggest other things that need to be addressed (like the spread of hitters' $H).

Thanks....

Ben, thanks for your comments.

I remembered the issue with the distribution of hitters from a comment you made to me last year, but that we are saved by the Central Limit Theorem (or whatever it is you said).

As I noted earlier, the 60% applies whether BIP > 0 or > 500 or > 1000. Let me publish the results of number of pitchers within 1 SD, assuming various breakdowns....

HEre are the 5 different runs I did

Thanks for taking the time to illuminate me (us?). To follow up on your last comment, can we determine, based just on these numbers, the DEGREE to which pitchers do control $H? I'm 95% sure that at least/most 30% of pitchers have control? I'm 99% sure that at least/most 10% have control? I'm ??x sure that at least/most ??y pitchers have control.

John, normally you are funny... but that's a good point. Maybe I'll present it within bands, say, how many pitchers have an SD within 1, for BIP 1 to 1000,1001 to 2000, etc,etc... If the "quality pitching" is a factor, we should see it. I might make the bands such that I have the same number of pitchers, now that I think about it. I'll do this tomorrow.

Here's the data, based on a breakdown suggested by John Cleese.

DIPS Bands