Bautista proves stat nerds wrong

No, he doesn't.

Juicy bait; fresh nightcrawlers.

Well, the question is, were the stat nerds uniquely wrong. IOW, did the anti-stats/scouting crowd predict what has happened? If not, then, the point is not anything about the stat nerds, it's about Bautista. You know, good for him and all. But unless the scouting crowd predicted it, so what?

Bautista proves EVERYONE wrong. Show me any article from before the season predicting 40+ homers and I'll revise that statement.

The headline was far more provocative than the article.

Clearly the Pittsburgh Pirates' brain trust saw this coming, and figured it would interfere with their master plan of finishing last every year.

Agree with Dan about the article. It mentions that all his homers are to the opposite field - I wonder if the scouting reports take this long to be updated. You'd think pitchers would be giving him a steady diet of outside pitches.

I don't know if there is a perception that statheads get frustrated being this wrong about a player. Not me. It would be a boring game if everyone just hit their projections. Bautista is a great, and fun story this year. I'm still skeptical he can do it again. I think this year will end up looking out of place with the rest of his career, like Brady Anderson or Roger Maris.

Bautista proves EVERYONE wrong. Show me any article from before the season predicting 40+ homers and I'll revise that statement.

SPORT magazine ran an article entitled "Is Jose Bautista The Next Babe Ruth?", but other than that, I can't think of any.

Yeah, I don't get what this proves. And it doesn't mention that those projection systems aren't personal opinions. Pretty worthless.

Bautista's year is a great story. There's no real lesson to take away from it though. Sometimes players have surprise seasons. What's more significant IMO is the year Jay hitters have had- Bautista, Gonzalez before he was traded, Buck, and Wells. But balanced by terrible seasons from Lind and Hill, along with a sort of lost season for Snider, the team is a wildcard going into next year. They did manage to turn Gonzalez into a pretty solid SS though, which was a nice move.

It seems it took Bautista a while to prove Cito wrong as well, batting either 1st or 7th in the lineup a quarter of the time.

The headline was far more provocative than the article.

This undersells it quite a bit. Other than his use of "stat nerds," he's quite complimentary of the work you guys do, as "even the most astute analysts have been caught looking by Bautista's breakout season," should demonstrate.

The headline, of course, is junk.

The headline was far more provocative than the article.

Agreed. He even writes this in the piece:

But you can hardly fault the baseball nerds for failing to predict Jose Bautista's remarkable season.

Assuming the author didn't concoct the headline, some editor is being unfair here. (Use of "baseball nerds" isn't offensive to me.)

The above snippet takes the piece out of context. The author also wrote:

But you can hardly fault the baseball nerds for failing to predict Jose Bautista's remarkable season. Here's what they had to go on:

EDIT: Cokes all around.

author didn't concoct the headline

I'm constantly surprised that people think authors do concoct the headlines. Certainly in my days in print media we kept authors as far away from the headlines as we could.

I read the headline as "McDonald and Bautista prove the stat nerds wrong"

And thought, ok, John McDonald is having a surprisingly strong season with the bat, but that's a bit much.

I'm constantly surprised that people think authors do concoct the headlines. Certainly in my days in print media we kept authors as far away from the headlines as we could.

I'm pretty ignorant on the subject, but I've alawys assumed one of the reasons for this was to have the headline written by someone who was putting together the whole page (ie. for formatting reasons). Is there a reason to continue this policy on the internet?

(This is an actual question, I have no idea if my original assumption is even correct)

I don't know if there is a perception that statheads get frustrated being this wrong about a player. Not me. It would be a boring game if everyone just hit their projections. Bautista is a great, and fun story this year. I'm still skeptical he can do it again. I think this year will end up looking out of place with the rest of his career, like Brady Anderson or Roger Maris.

The only "frustrating" thing to me about projections is that they're necessarily conservative. You can basically guess at a player's projection just by looking at his age and his stat line for the last three years, and basically end up in the ballpark; I can't recall being shocked at a player's projection, aside from some minor league prospects, but that's because I don't have a sense of park and league adjustments for the minors.

I think Gary Huckabay had a system (Wilton) that was designed to try to predict breakout seasons more. That was more fun, but he was probably wrong more than the more conservative projection systems.

This is why I like BP's projections, where they give you percentiles. That way you can get a real sense for how a player will do if everything breaks his way. Of course, even then, the percentile projections basically fall along a straight line.

How are the systems with pitchers these days? Is it still easier to project a hitter than a pitcher? I imagine pitchers are still harder because of the lack of predictability of BABIP, and because of pitchers pitching through injuries, etc., which would seem to affect pitchers more than hitters.

Anyway, obviously with Bautista the main points are:

* Everyone misses on a player like him, scouts and nerds alike;
* There are seasons like Bautista's sprinkled throughout major league history.

Shouldn't the writer have some say over the headline, so he doesn't look like an idiot?

I've always assumed one of the reasons for this was to have the headline written by someone who was putting together the whole page (ie. for formatting reasons).

Like everything 'here' was derived from the route from 'there'. In the days of hot metal you might rework the headline through several editions as the space available was adjusted by the editor's priorities or story developments.

Even on a web site that might apply if you've got the article linked from a main contents page and want identical headlines.

Shouldn't the writer have some say over the headline, so he doesn't look like an idiot?

If you don't want to look like an idiot in print, don't write for publication. The fact is, so many things can happen, even accidentally, in the publishing process that all writers have something in their past they want to hide.

I've never written a title for a non-BTF article. Which is awesome because I hate titles (and endings).

Back in creative writing in college, a favorite trick of mine was to simply use some random non sequitur as the title. Things like "Soccer Balls Are Just Postmodern Beehives" that had no connection with anything in the piece.

I've referred to myself and others as statnerds as well. Back in usenet days, SDCN (stat-drunk computer nerd) became a term of endearment.

"I'm pretty ignorant on the subject, but I've always assumed one of the reasons for this was to have the headline written by someone who was putting together the whole page (ie. for formatting reasons). Is there a reason to continue this policy on the internet?"

Yes, the reason writers traditionally did not write headlines was that there was no way for them to know how wide the headline or how many "decks" - maybe a large headline and a subhead, for example - there would be on the printed page.

NY Times columnists got to write their own headlines because the column always ran in the same place on a page, however.

Many longtime writers who have added blogs are getting their first experience at headline writing.

Back in creative writing in college, a favorite trick of mine was to simply use some random non sequitur as the title. Things like "Soccer Balls Are Just Postmodern Beehives" that had no connection with anything in the piece.

You would have done well in the Michael Nesmith School of Songwriting.

DB

Back in creative writing in college, a favorite trick of mine was to simply use some random non sequitur as the title. Things like "Soccer Balls Are Just Postmodern Beehives" that had no connection with anything in the piece.

I once knew a guy who wrote a final exam entitled "Kittens are Cute". Unfortunately that was the entirety of the essay, and it was supposed to be on Gulliver's Travels.

I write the dullest titles ever. They're pretty much just an apt description of what I'm writing about. For instance "Relgious Loyalism: Anglican Church Reactions to the French Revolution in the Winter of 1792-1793". Once I really let my creative juices flow and came up with "The Dawn of the War Correspondent: the Duke of Buckingham in the Isle of Rhe Campaign 1627"

Perhaps one day I'll get ballsy and include a quotation or something in the title.

This would be a good hijack. Dullest sounding academic paper you've ever written.

Shouldn't the writer have some say over the headline, so he doesn't look like an idiot?

McDonald's Newsmedia article was thrown around a bit ...but this was the MOST SHOCKING, so, natch, I ran with it.

Others...

"Is José Bautista for real?" (Montreal Gazette)

"What's up with Jose Bautista?" (The Ottawa Citizen)

Following sports IS nerdy no matter how you follow it, even if you're a big tough guy wearing your Joe Klecko vintage Jets jersey and screaming at runningbacks not to be such pussies.. I think, really, the act of being an audience is inherently nerdy while the act of participation would be anti-nerdy. It seems as I get older I'm becoming the audience way too much. Gotta do something about that.

Back in creative writing in college, a favorite trick of mine was to simply use some random non sequitur as the title. Things like "Soccer Balls Are Just Postmodern Beehives" that had no connection with anything in the piece.

I hated giving my stuff titles, too, which would lead to half hour discussions in workshop about my ####### titles. Did not enjoy the workshop experience. No, sir, I did not.

This would be a good hijack. Dullest sounding academic paper you've ever written.

Jeez, mine were so boring I can't even remember them.

This would be a good hijack. Dullest sounding academic paper you've ever written.

The Asymmetric Effects of Deflationary and Dis-inflationary Monetary Shocks

#17 Gary's system was called "Vlad". Wilton I believe came from Clay Davenport (and though there was no discussion of its design, the output sure looked regression based).

The thing about Vlad was that it was optimistic by design and saw breakouts everywhere. (Even then, it often missed the nature of the breakouts. It correctly forecast Todd Hundley's first good year except that it saw it as being basically driven by power. Missed quite badly on the OBP)

And yes, when I looked at the matter (can't find the post but it's on RSB) I found it was less accurate than the STATS projections with Bill James' name on it (which wasn't Brock2 based by then but still quite conservative). Not a whole lot better than simply guessing that a player would repeat his previous year. IIRC Vlad shaved maybe a half run off the standard error. It had a pretty decent hit rate (defining hit by getting the rate stats close enough that the projection would be within 6 runs given full time play) but crashed and burned a lot more than other methods.

Also, it had the endearing habit of trying to give Barry Bonds 900 or so plate appearances (unlike other projection systems it tried to predict playing time too). Clearly correct if you can swing it.

EDIT: There were actually two Vlads. Gary's house burned down and took VladI with it.

Assuming that Bautista's performance has not been due to PEDs, his extraordinary power burst was, statistically, predictable. I'm not suggesting that individually it was. But in a sample size as large as all major leaguers and all players who start the season on a team's 40-man roster, one or two Jose Bautistas or Andres Torreses is very likely to emerge. It's just that you can't say ahead of time with any precision who those outliers are going to be.

Somewhat analagous is predicting that one year I would be a brilliant stock-picker.

Back in the early-90s, I took part in a newspaper's stock-picking game. Anyone could enter and it cost nothing. You had to pick (if I recall correctly) at least 10 stocks and at most 50. You could weight the stocks however you wanted. Your goal was to have the greatest growth in your fake portfolio over the period of one year. The newspaper also had a few local "experts" participating. If you beat all of the experts you won a small prize. If you beat everyone else who entered, you won something like $1,000.

The first year I entered I did extremely well. The company which took off for me was Dell. I ended up in the top 2% of all stock pickers. (I didn't get a prize because I finished behind one of the "experts.") However, the fact that I did extremely well was a fluke. CHONE and ZiPS and B-Pro never would have figured me as a guy who would wind up near the top. They all would have pegged me for average.

The next year I entered and followed pretty much the same strategy of picking mostly tech stocks and financials and I finished around the 40th percentile. I stopped entering after two years. I didn't want a third year to show I could finish in the bottom of the heap, too.

Back in creative writing in college, a favorite trick of mine was to simply use some random non sequitur as the title. Things like "Soccer Balls Are Just Postmodern Beehives" that had no connection with anything in the piece.

The truth revealed: Dan Szymborski is actually Robert Pollard.

A large part of Bautista's success this year has been the lack of respect pitchers have shown him so far... throwing him fastballs on favorable hitter's counts. That has stopped recently and Jose has stalled on 43 HR (as have other HR leaders on their totals) while taking a lot of walks.

My point is that a projection system would have trouble taking onto account the lack of respect shown to a breakout hitter by pitchers.

Now, what do the projections by nerds say forBautista for next year?

All you needed to know to predict Bautista in '10 was to look at what he did in September of '09 (10 HRs in 30g, .600+ SLG).

But what stat nerd in his right mind would have put any credence in such a small sample size??

What about outliers? For any projection system, baseball or otherwise, there are going to be those that lie two or three standard deviations outside the mean, but you are still hitting 95% or so within one standard deviation.

If the rest of the projection system is pretty accurate does that mean it can be discarded simply by finding one or two misses?

Back in usenet days, SDCN (stat-drunk computer nerd) became a term of endearment.

I thought that was an RLM creation.

Like King Tut would refer to ostriches.

A large part of Bautista's success this year has been the lack of respect pitchers have shown him so far... throwing him fastballs on favorable hitter's counts. That has stopped recently and Jose has stalled on 43 HR (as have other HR leaders on their totals) while taking a lot of walks.

If this is true than pitchers must be kinda dumb. His hottest HR hitting spree was in mid-late August (at the beginning of which he already led the league in HRs).

I'm not sure what it means exactly, but Fangraphs runs per pitch type per 100 pitches has Bautista at

Fastball - 2.79
Curve Ball - 2.16
Slider - 1.86
Cutter - 1.07

He does seem to not hit the change up or splitter well.

Like I said, I don't know how to interpret those numbers exactly, but I would think it wouldn't have taken until the last week of August for pitchers to realize they shouldn't just feed him fastballs.

[27]

Shouldn't the writer have some say over the headline, so he doesn't look like an idiot?

McDonald's Newsmedia article was thrown around a bit ...but this was the MOST SHOCKING, so, natch, I ran with it.

Others...

"Is José Bautista for real?" (Montreal Gazette)

"What's up with Jose Bautista?" (The Ottawa Citizen)

From the Ottawa Citizen version at canada.com:

David McDonald is an Ottawa writer and a member of the Society for American Baseball Research.

- - -

Is José Bautista the new Ted Williams or the new Ned Williamson?

Some players with lengthy careers who have hit an inordinate percentage of their career homers in a single season.


Player         Age   {Best HR year} {Career HRs} {% of career HRs in best year}
José Bautista   29     43 (2010)*     102                   42.2
Ned Williamson  26     27 (1884)       64                   42.2
Dale Sveum      23     25 (1987)       69                   36.2
Felix Mantilla  29     30 (1964)       89                   33.7
Davey Johnson   30     43 (1973)      136                   31.6
Kevin Elster    31     24 (1996)       88                   27.3
Jim Gentile     27     46 (1961)      179                   25.7
Brady Anderson  32     50 (1996)      210                   23.8
Roger Maris     26     61 (1961)      275                   22.2

*And counting

This would be a good hijack. Dullest sounding academic paper you've ever written.

there's just not much spread between least and most dull.

boring: Effects of precursor composition on the local structure of Cu dispersed on mesoporous silica
not boring: Self-Aligned Micropatterns of Bifunctional Polymer Surfaces with Independent Chemical and Topographical Contrast

but they're both pretty boring sounding (not to me, but i assume to others.)

boring: Effects of precursor composition on the local structure of Cu dispersed on mesoporous silica
not boring: Self-Aligned Micropatterns of Bifunctional Polymer Surfaces with Independent Chemical and Topographical Contrast

Or another fun hijack.
Guess people's discipline by the titles of their work.

In your case it's pretty easy (Medieval Russian Literature), but others might be fun.

All you needed to know to predict Bautista in '10 was to look at what he did in September of '09 (10 HRs in 30g, .600+ SLG).

But what stat nerd in his right mind would have put any credence in such a small sample size??

In the SF Giants ZiPS thread, at least one poster was killing Dan (Szymborski) for not projecting Dan (Runzler) more optimistically, based on Runzler's work towards the end of 2009. I think the 2010 stats vindicate ZiPS a bit.

"The exception that proves the rule. Works every time."

--Billy Dee Beane

I'm sure all the straw hat guys with cigars knew he would hit about 50 homers.

All you needed to know to predict Bautista in '10 was to look at what he did in September of '09 (10 HRs in 30g, .600+ SLG).

But what stat nerd in his right mind would have put any credence in such a small sample size??

Hell, what stat nerd in his right mind wants to add 2010 to the sample size and project Bautista to repeat this performance next year?

I know I sure as hell don't.

I, and I alone, predicted Bautista's breakout season.
http://www.baseballprospectus.com/article.php?articleid=10585
http://fantasybaseball.fanhouse.com/2010/06/01/stud-or-shelton-jose-bautista/
I tossed his name out in my fantasy league's auction draft and kept bidding until there were no other bidders. He is not a one-year wonder either; he'll be a regular 30 HR guy for the next few years.

Greg K @37. No, I stand by my post. Bautista's home run hitting has been consistent all year, with 11 or 12 per month except June. Yes he hit 12 in August, whence player of the month, but he has only hit one in his last nine games as of this writing.

And yes pitchers/catchers/managers have been slow to keep the ball out of wheelhouse to find his weakness.

Heavyhitter @45. You are the only one to anticipate this. Got some lottery numbers for me?

I've referred to myself and others as statnerds as well.

It's the kind of word that other nerds can use, but if you portray yourself as a non-nerd and throw it at us, then it's a fightin word.

47, Nerd is like another n word.

If this is snark, it's pretty dumb snark.

EDIT: referring to #45.

How are the systems with pitchers these days? Is it still easier to project a hitter than a pitcher? I imagine pitchers are still harder because of the lack of predictability of BABIP, and because of pitchers pitching through injuries, etc., which would seem to affect pitchers more than hitters.

Depends a lot on what you're matching to. With pitchers, beyond sample size issues and stuff, there are two other key sources of error variance. First is that, even if you get the "true" rate stats right, how they'll play out in one season is highly variable (triply so for relievers ... OK, probably square root of triply for relievers). That is even if you peg them exactly right on K/9, BB/9, HR/9 and BABIP, the sample variance on those can be pretty big. Note, this is true for hitters as well.

But then you get into the mis-match between predicted FIP or component ERA or whatever and actual ERA -- which is the annoying thing that a single followed by an HR hurts a lot more than an HR followed by a single but FIP doesn't know that. We tend not to do that sort of comparison for hitters but still do it for pitchers. The hitters' equivalent would be something like seeing how well ZiPS does in projecting RBI (per 100 PA or something).

Thing is, even for hitters, individual-level confidence intervals are huge. The 95% CI from ZiPS is something like +/- 35 points of OPS+. (In one of the Swisher-Betemit trade threads, I discussed this using their ZiPS projections ... based purely on the projections, it wasn't a bad trade for the Sox ... although I recall Dan personally thinking it would be.) I recall in the early PECOTA days (and I doubt things are much better) the 90% or 95% CI on some young Red Sox pitcher was something like an ERA between 2.1 and 6.2 -- I coulda told you that without any information about the guy whatsoever.

In the SF Giants ZiPS thread, at least one poster was killing Dan (Szymborski) for not projecting Dan (Runzler) more optimistically, based on Runzler's work towards the end of 2009. I think the 2010 stats vindicate ZiPS a bit.

Not to mention all the grief Dan took over the Ben Zobrist projection ... and rightly so given he over-projected it by 22 points of OPS+! :-)

Not to mention all the grief Dan took over the Ben Zobrist projection ... and rightly so given he over-projected it by 22 points of OPS+! :-)

I wish I'd put more stock in that one...kept Zobrist over Carlos Gonzalez this year...

Groan. My fantasy team is being killed by Ben Zobrist and Cliff Lee. WTF?