THT Mailbag: We’re Back

So now that the season is underway and all is once again right with the world, we should be getting back to weekly-ish mailbags. I’ll also use this space to once again pimp the A’s-Angels rivalry. I heard during the broadcast today that today’s 2-1 A’s victory was the 21st one-run game in the last 42 between the teams, who have combined to win the last five division titles. I’ve been lucky enough to attend a couple of these one-run thrillers, and I can’t imagine that they are any less exhilarating than a Red Sox-Yankees game. If nothing else, they are definitely tighter games with more close plays and old school pitching and defense. From Jason Kendall diving over the plate to squelch the would-be game-tying run at the plate on Thursday or Vlad Guerrero terrorizing the A’s all weekend, this rivalry is what baseball is all about.

He Just Hits

> I probably just missed it, but has anyone noticed that Juan Pierre had over 200 hits and an OBP of .330? This seems difficult, or am I way off base? Just wondering how often this combination occurs.

– Rich P

Matthew Carruth: Simple answer? Not often. With those exact parameters (200-plus hits, .330 or less OBP) has happened five other times from the one you mentioned (Juan Pierre in 2006). The other five:

• 1967 Lou Brock, 206 Hits, .327 OBP
• 1970 Matty Alou, 201 Hits, .329 OBP
• 1973 Ralph Garr, 200 Hits, .323 OBP
• 1979 Buddy Bell, 200 Hits, .327 OBP
• 1985 Bill Buckner, 201 Hits, .325 OBP

Expanding the OBP threshold to .350 shows five occurrences prior to 1960 and a whopping 28 overall, with repeat offenders in Garry Templeton (1977 and 1979), Bill Buckner (1982 and 1985) and Cecil Cooper (1982 and 1983).

W.O.A.T.

Neifi Perez was fast moving “up” the ranks toward worst-ever hitter, according to Lee Sinins’ Runs Created formula.

That was before ’06, during which he stank, even by his own standards.

Is he No. 1 yet? The suspense is killing me.

– Mike C.

Dave Studeman: I don’t know how Lee compiles his lists, but here are the ten worst batters in major league history, according to the most recent version of Lee’s Sabermetric Encyclopedia (which includes 2006):

RCAA                           RCAA   
1    Tommy Corcoran             -442  
2    Ski Melillo                -355  
3    Neifi Perez                -354  
4    Germany Smith              -352  
5    Tommy Thevenow             -351  
6    Malachi Kittridge          -343  
7    Joe Quinn                  -333  
8    Bones Ely                  -317  
9    Royce Clayton              -314  
10   Bill Bergen                -312 

Tommy Corcoran was a shortstop in the 1890s and 1900s, and Ski Melillo (what a great name) was a second baseman in the ’20s and ’30s. Corcoran’s lifetime OPS+ is 74, and he finished in the top 10 in outs made 10 times. Melillo’s lifetime OPS+ is 63, the same as Neifi’s. You’re right about last year; Neifi’s OPS+ was 46 last year, the second lowest of his career.

By the way, Lee’s Sabermetric Encyclopedia is a great resource for any stats-oriented baseball fan.

Gambler’s Fallacy

Is there any evidence to suggest that there is a negative correlation between spring training performance and regular season performance? For example, if you look at the season as seven months (six plus spring training and you saw Tomo Ohka’s 2.84 ERA in the first month, would you not expect his regular season to be worse that originally expected as a regression to the mean? Or if you saw Ryan Howard’s .218/.313/.382, would you not expect his next six months to be even more awesome than originally thought? To expand on this, in a situation where two or more similar players are competing for one roster spot, would it not be a better strategy to use the pitcher putting up a 5.00 ERA over a 2.00 ERA?

– Peter S.

John Walsh: It seems to me like an example of what you’re suggesting is the following:

Players A and B both have projections for 2007 of OPS .900. In spring training Player A has an OPS of .700, while Player B manages an OPS of 1.100. So, since both players are projected have an OPS of .900, shouldn’t we expect Player A to play much better than Player B during the regular season, if both are going to meet their projections?

This is known as the Gambler’s Fallacy, which is typically illustrated by considering flipping a coin. If you flip a coin five times and get tails each time, what is the chance of getting tails again on the next toss? It is exactly 50%; in other words, the previous outcomes have no effect on what happens next. Or, getting back to Players A and B, once the spring training stats are in the books, they won’t have any effect on what comes during the regular season.

Unless, of course, you want to use the spring training performance to update the projection. This has been studied many times, though, and it’s generally found that spring training stats do not help in projecting the regular season. One possible exception was found by John Dewan of Baseball Info Solutions, who found that players who exceed their career slugging percentage by at least 200 points in spring training have a moderate tendency to show improved power during the regular season.

I hope this answers your question.

Tiger, Tiger

The most remarkable missing team from your 2006 predictions is, of course, the Detroit Tigers. Nobody figured they’d do anything, and in fact if it wasn’t for the Kansas City Royals, I suspect they’d have been picked for last in their division.

Therefore, since past history is a strong indicator of future reality, let me hereby predict that the Kansas City Royals will win the AL Central, and will represent the AL in the World Series. The NL will be won by the Phillies, and they will be MLB Champs.

The Royals are unanimously picked by everyone to do nothing, just like the Tigers were last season. The Phillies are picked by four prognosticators to finish first in the NL, but no one seems to think that they can possibly win it all. In 2006 exactly the same disappointing future was predicted for the St. Louis Cardinals. Subsequently, they are the obvious choice to emerge as champs.

Phillies beat the Royals in six games for the World Series championship.

– David R., Shawnee, Kansas

John Beamer: That’s not quite true.

Sure, few, if any, thought that the Tigers would the American League’s representative in the fall classic. Diamond Mind pegged the Tigers at a 79-83 team, which is close to .500—my recollection is that that was typical of the various projection efforts.

Sure, they were a lot better than that but if you think about the team a helluva lot went their way. Justin Verlander turned in an ace-like performance, Nate Robertson outperformed in the first half, and even relatively obscure hurlers like Zach Miner contributed. Add that to the emergence of Joel Zumaya and a strong pen you can see why they did well in the first half. In the second half they struggled and played more like a .500 team.

Anyway, the point is that while the Diamond Mind got it wrong, it wasn’t as spectacular bust as it first seems. Interestingly THT’s own Nostradamus, David Gassko, picked the Tigers for the 2006 AL Central.

So, what about the Royals. THT has them as a 67 win team (ie, nowhere near .500). Not withstanding their opening win against the Sox, they would need to be exceptionally fortunate (or Alex Gordon would have to be hugely inspirational) to mount a challenge. The Phillies? I’ll give you that one—there is no reason why they don’t have a genuine shot at the NL East. And we all know that once you’re in the playoffs it is a crapshoot

In case you were wondering David Gassko went for Cleveland in the AL Central, with the Tigers in fourth and the Royals in fifth. Based on that alone , if you feel like a wager on the Royals playing October baseball then let me know.

Proven Winners

We all know that wins are a poor judge of pitching talent. We also all know that many people continue to insist that they are a good judge of pitching talent. What’s interesting to me is that this divide seems to be largely generational—those who grew up watching starters pitch more innings (and get more wins/decisions) believe won-loss record is useful, and younger folks don’t. This is a crude description; Bill James is no one’s idea of young, but you get my point.

Here’s my question: Was there some point in time at which wins were correlated with pitching skill? In other words, what is the historical, year-by-year, correlation between wins and some decent measure of pitching skill, such as ERA+ or mERA? I’d be interested to see the graph of this correlation to find out if the oldsters ever had a reason for valuing wins or even describing them as they currently exist. I’m familiar with the moralism involved in original scoring rules and decisions, but when we describe things that way, I think we lose the context of the times (yes, I’m an historian).

So, can someone run the correlation between ERA+ and Pitching wins for leagues from 1900-present? I’d predict that it slopes downward with a few peaks, but I’m really just not sure. Would that correlation track with percentage of team innings pitched by starters? Or a percentage of total innings (innings/start) pitched by individual pitchers? I guess I’m asking for a thought experiment in which we treat “wins” as a skill, and we try to figure out what characteristics a “winner” has. Wins have become a straw man these days, and I think it’d be interesting to try and figure out why people ever thought they meant something in the first place. They weren’t just uneducated schmoes; they had reasons.

– Christopher M., Milwaukee, Wisconsin

David Gassko: That’s an interesting question. Perhaps because pitchers threw more innings in the earlier days, and had more room to pace themselves and gas it up when necessary, wins correlated better with pitcher quality back in the old days. So I looked at the correlation between wins and ERA+ (corrected for league, but not park) for all pitcher seasons with at least 100 innings pitched since 1901:

As you can see, the graph does show a slight downward trend, but it’s not anything worth writing home about. However, what you can also see is that the correlation between wins and ERA+ is pretty high; historically, it’s around .60. That’s the thing: Wins are not a poor judge of pitching talent; they just aren’t the best judge, either.

After all, look at last year’s leaders in wins: Johan Santana, Chien-Ming Wang, Jon Garland, Freddy Garcia and Kenny Rogers. Are any of these guys bad pitchers? No, they’re all very good! It’s just that these days, we have better ways of evaluating talent, just like we have better ways of evaluating car safety, flight routes, etc. Society advances, but that doesn’t mean that it’s old methods are useless.

Would I rather have ERA+ than wins? Yes. But would I rather have wins than nothing? Of course.