About Us  |  Analysis  |  Stats  |  Stuff  |  News
Baseball Primer
Baseball Primer Sections

  You are here > Baseball Primer > Articles  > Voros McCracken  > DIPS Version 2.0

DIPS Version 2.0
by Voros McCracken

Well, it’s finally here. After over two years of debate, argument, re-debate and re-argument, I’ve actually gotten to the next step in the process of developing pitching statistics that are not dependent on the quality of defense behind the pitcher.

Now I’m sure a good number of you have followed the bouncing ball on this so far. But for those Johnny-come-latelys, here’s the state of things (in my opinion at least) as they stand right now:

  1. The amount that MLB pitchers differ with regards to allowing hits on balls in the field of play, is much less than had been previously assumed. Good pitchers are good pitchers due to their ability to prevent walks and homers and get strikeouts in some sort of combination of those three.
  2. The differences that do exist between pitchers in this regard are small enough so that if you completely ignore them, you still get a very good picture of the pitcher’s overall abilities to prevent runs and contribute to winning baseball games.
  3. That said, the small differences do appear to be statistically significant if generally not very relevant.

The finer points and how this all came about, and the ramifications have been discussed before. With the first step of DIPS as it was used in 1999 and 2000, dealt with points 1 and 2, it did not deal with point 3 just yet. Now what I will try to do is outline a necessary first step in trying to get at the very small differences between pitchers in this regard, while still keeping the stats independent of the defenses behind them.

One of the biggest problems I’ve had with Defense Independent Pitching Statistics is a misunderstanding about what it is I am trying to do. Many people argue that since I agree that there are small differences between pitchers in this regard, I should count the stat in the numbers just to a lesser extent. You run into problems there, however, in that you can no longer safely assume that the quality of the pitcher’s defense didn’t help or hurt him. Quite the contrary, that possibility is as likely as any other.

For example, say you have a pitcher who allowed a .275 rate of hits per balls in play while his team in total posted a rate of .285, which would indicate a well above average defense. The argument, as it has been made, is that I should keep the difference between the pitcher and the team’s defense, and simply adjust for the defense upward some.

The problem is that this is an inconsistent application. By using that .285 as the benchmark for the team’s defense, you are assuming the pitchers on the team are not any better at preventing hits per balls in play than anyone else. But by then using this adjustment, to show one pitcher being better in this regard, you’ve just contradicted yourself. In other words, the adjustment is only valid if we can assume that there’s no difference between pitchers in this stat or that they always even out on any particular pitching staff. But the latter argument is obviously shaky, and if the previous argument is true, there’s no need for the adjustment anyway. It becomes an entangled “chicken and the egg” scenario at the end of which we no longer have a pitching stat that is independent of defense.

Better, I have always argued, to try and get at these differences from a different way. A way that retains the defense independence of assuming no difference between pitchers on hits per balls in play, but can fine-tune it to pick up small differences here and there, while still avoiding the stats directly related to team defense. I’m hoping that what I’ll detail here is an improvement in this direction. Here’s a point by point breakdown:

IN THE NEW SYSTEM TO FIND DIFFERENCES BETWEEN HITS ON BALLS IN PLAY

Knuckleballers – Well after two years of wondering on this score, I’ve run through the numbers, and the biggest slam-dunk, “no doubt about it” difference to be found is here. Conclusively, knuckleballers have an advantage with regards to hits per balls in play.  Some of the very lowest rates of the last 50 years have been by these guys. Chralie Hough, Tim Wakefield, Phil Niekro, Hoyt Wilhelm, they’re all a decent bit better than their teammates in this regard. Using Craig Wright’s list from the Diamond Appraised of knuckleball pitchers, and adding Wakefield, Dennis Springer and Steve Sparks, I found a definitive difference between knuckleballers and normal pitchers. The knuckleballers tend to have an advantage of anywhere from .008 to .012 depending on how I choose to slice the pie. 10 points is a nice round number and should work well enough.

Lefty/Righty – A small difference appears here, and I’m not really sure where it comes from, but it appears to be statistically significant. It’s only around a .002 difference and since I’m not distributing prescription drugs to orphans here, we can put it in, and come up with the “whys” and “wherefores” along the way.  It might be a ground-ball/fly-ball issue. Maybe lefties tend to be more likely ground-ballers than righties. But at this point that’s really only a guess.

Strikeouts – Here’s where the murkiness starts. Looking at the numbers over and over and over again, it becomes clear that a pitchers strikeout rate during a single season is a bit better predictor of his hits in play the following year than his own hits per balls in play. This is there and it’s real. Why? I can only come up with two explanations: the obvious and the hard to show. The obvious explanation is that the more a hitter swings and misses at a pitcher, the more he also makes poor contact and therefore doesn’t hit the ball hard. Most would favor this, though if this is the case, there are some questions as to why the differences between pitchers aren’t greater.

To propose another possibility, the more players you strikeout, the more often hitters hit with two strikes. Since this is a noticeable ability for hitters, if a particular situation causes a hitter to change his approach, it could effectively make the hitter a worse hitter in this regard. And if a particular pitcher creates more of these situations than others, it would effectively reduce the quality of hitters he faces in this regard. This dovetails a bit better with what we already know, but it also seems like a little bit of a reach.

Draw your own ideas and conclusions here.

Home Runs – I wasn’t going to include this until the very end. I asked myself if the numbers said more with it, than without, and I decided on “with.” While a shaky relationship, it appears that the more Home Runs a pitcher gives up, the fewer hits per balls in play he gives up.

The problem here is that the “why” of this is a minefield. Clearly there’s an obvious problem that if there were pitchers who gave up lots of homers AND lots of hits per balls in play, that might greatly reduce their chances of remaining in the league and therefore in the study. So this is a real problem.

So I ran through every way I knew how to minimize this problem (using only established pitchers; running the numbers over various times; etc.) , and at the end of which, it was still there. Balancing these problems, is that Home Runs work as a decent proxy for ground-ball/fly-ball tendencies, especially when we don’t have this data available. It also lacks some of the problems inherent in those numbers (which I’ll get to below).

This stat does not affect the numbers much at all, but I felt it was worth putting in there. I will provide the details on the system on my web-site, and this is the only one of the above I think you could leave out if you so chose, and not effect the other numbers badly at all.

NOT IN THE SYSTEM TO FIND DIFFERENCES BETWEEN HITS ON BALLS IN PLAY

Walks – If there’s any relationship here, it’s that the more walks a pitcher allows, the fewer hits. But it suffers from the same problem Home Runs suffer from above, but in this case, I think the problem dwarfs any effect that may or may not be here. Maybe if there’s some additional refinements in the future, this may go in. For now, I think it’s best if we leave it out.

Height – Here’s something to ponder: why exactly do I get a statistically significant correlation between the pitcher’s height and his hits per balls in play? After racking my brain for days, I decided the best possible explanation is that the shorter pitchers tend to be much better fielders than the taller pitchers (say Greg Maddux versus Randy Johnson as an example). So maybe the shorter guys stop more line drives up the middle, field bunts better and cover first better thereby reducing the hits off them a bit.

Of course, realistically, this we should measure this in a different way other than how tall the guy is, and technically it’s not pitching anyway but rather defense. I can’t justify including this one as it’s weird and probably prone to a pretty decent error rate. I still think it’s interesting though.

Ground-Ball/Fly-Ball Ratios – Well, this one ain’t going away anytime soon, so whatever I decide, it certainly is by no means final. It has been proposed for some time that fly ball pitchers tend to have an advantage here over ground ball pitchers. In the end, I’m pretty sure this is right, but the problems currently here are tough to overcome.

For starters, I’m not convinced that these numbers are defense independent. There are really three types of batted balls as they are counted: ground-balls, fly-balls and line-drives. It isn’t hard to imagine a batted ball where it is difficult to determine whether it was a line-drive or a fly-ball. Since humans are making subjective judgments here, there is some question as to whether the eventual determination of fly-ball or line-drive might be largely influenced as to whether the ball falls in for a hit. Now granted, this would be a small effect, but we’re talking about trying to find small effects to begin with. The problem with skinning this cat is that even small biases need to be taken into account. So the problem might be that an increase in hits per balls in play might cause a decrease in ground-ball/fly-ball ratio, rather than the other way around. Nothing can be more ugly to deal with than reverse-causation.

Also, because the differences here are so small, we need lots and lots and lots of data to be able to be sure of significant tendencies. Because ground-ball and fly-ball data are a relatively new phenomenon, it greatly reduces the amount of data we have to examine. So without the extended data, this sort of analysis complicates the whole shooting match. If we did use the ground-ball and fly-ball data, would it then remove the effects of Home Runs and Left Handed pitchers? Even though we can get the data now, it isn’t exactly published in every sports page or on every web-site with pitching statistics. The use of ground-ball and fly-ball data would then make it difficult for those with a desire to run numbers to be able to do so.

For now, I’ll keep working on how to make this work (if indeed at the end it really does), and continue to use Home Runs allowed as it’s proxy. I’ll probably try and make a few phone calls to stats as to the nature of the numbers and see if I can’t maybe count all of the balls in play not currently in the fly-ball or ground-ball category as fly-balls.  This would once again make them defense independent, but in my opinion probably lessen the relationship some as well. When I (or someone else for that matter) can work this out, we’ll have DIPS Version 3.0.

--

Anyway, before anybody gets carried away here, I should note that of the 100 pitchers with the most batters faced, 93 of them eventually rated within .006 of the league average hits per balls in play rate using the above adjustments. As you would guess by my comments, it comes as no surprise that the two lowest estimated rates go to Steve Sparks and Tim Wakefield, both knuckleball practitioners (.279 and .284 respectively with a league average of .292). The highest estimate went to Jimmy Anderson (.300) of the Pirates, who is lefty who didn’t strike much of anybody out.

I’m fairly sure that in many instances, the estimates are actually going to be further from whatever the pitcher’s actual ability may be, than simply assigning him a league average rate. But I think on the whole this does get us a bit closer to the true rates. The system was developed independent of this years’ results, so this year would be a decent test to see if it helps any. Comparing the correlation between the pitcher’s hits per balls in play rate (using pitchers with 100 or more innings pitched) and that of his team’s was .415, which is essentially the way the old DIPS worked. Using these new adjustments, and estimating the differences as being from his teammates rather than some “league average” the correlation is .428 a very modest improvement, but improvement nonetheless. That doesn’t address the issue above regarding the teammate’s rate containing some level of pitching. So, by then looking at how this system estimates pitching affects this stat for each team, and then adjusting by team with that adjustment made, the correlation then goes up a little more to .433. All of this said, the difference between doing it this way, and simply using the pitcher’s team rate is fairly minimal, but then again the differences between pitchers in this regard is probably minimal anyway, so every little bit helps.

The final adjustment here is to use a new method to come up with the earned run estimates, to avoid problems with possible differences in extra base hit ability. This problem can be sidestepped for now by a simple reduction of the relative run value of a non home-run hit and a slight increase in the relative value of a home run. Multiple regression analyses confirm this.

Obviously there’s a lot to be discussed about this, and, as always, I’m hoping this is the beginning of a discussion and not the end of it.

Here is the link to this year’s DIPS numbers:

http://www.baseballstuff.com/mccracken/dips2001.html

Here is the link to the explanation on how to calculate this new version of DIPS:

http://www.baseballstuff.com/mccracken/dipsexpl.html


Related Web Pages and Articles


Reader Comments and Retorts
Post a comment | E-mail us directly | Report an inappropriate comment

January 25, 2002 - Warren

Nice work, Voros.

I'm not sure I understand how "the final adjustment here... [avoids] problems with possible differences in extra base hit ability." Given that pitchers have varied percentages of (non-homer) extra-base hits per hit allowed, does DIPS 2.0 take this into account? For example, each (non-homer) hit that Greg Maddux allows is more likely to be a single than each (non-homer) hit that Rick Helling allows. I realize there are problems with this, since extra-base hits are defense-dependent.

One more comment: after the defense-independent hits total (dH) is found (along with the other defense-independent stats), you figure ERA in a manner similar to component ERA - that is, using a linear system to predict runs allowed. Component ERA (or linear-weights ERA) certainly has its place, since it's been shown to predict future ERA better than plain vanilla ERA. But since you've mentioned that DIPS isn't designed with prediction in mind, wouldn't it make sense to adjust the pitchers' true ERA based on $H?

For example, imagine a pitcher who allowed 100 runs in 200 IP (4.50 ERA). For whatever reason ("clutch" pitching, good pickoff move, etc.) this ERA is lower than his component ERA, which is 4.80. Say he allowed 5 more hits than he "should" have, given his estimated $H rate. Your method essentially adjusts the 4.80 number to account for this, but why not adjust the real ERA (4.50)? This would create a DIPS ERA that has less predictive value, but that's not really the point of DIPS, is it?

All that being said, very nice work. I wonder if you'd share a few more details about how you determined that handedness, strikeout rate, etc. play a role in $H - can we see some of that data? Peer review and all that :)

January 25, 2002 - Sean Forman

January 25, 2002 - Mike Emeigh (e-mail)

Two quick comments:

-- The "knuckleballer's advantage" is most likely related to two phenomena: errors and K+WP/K+PB (knuckleballers have more of these). I recall reading a study (don't remember where) that showed there tended to be more errors behind knuckleballers, on average, than behind other pitchers; I'm in the process of checking this using PBP data. The K+WP/K+PB will definitely factor into the equation when you use Voros's estimator for BFP, because you're assuming there that all strikeouts are outs. -- The lefty/righty issue, like many similar issues related to pitching, is actually a hitting issue. LHP face fewer LHB than do RHP, and LHB tend, on average, to be better hitters than RHB, and also to have a higher percentage of hits on balls in play, whether they are facing LHP or RHP.

-- MWE

January 25, 2002 - Sean Forman (e-mail)

Voros,

How would you recommend making the park adjustments. Say from 1933, when all you have is the PPF and also, would new values need to be computed for older data?

January 25, 2002 - Colin

Voros, Could you make a few of your relationships a little clearer here? I think I understand them, but when you mention lefty/righty (and to a lesser extent, though more easily figured out, height) you don't make clear which direction the difference is.

January 25, 2002 - Aaron (e-mail)

I have a request. How about a report on the 20 or so pitchers whose stats were most affected by defense?

Aaron

January 25, 2002 - Sam Hutcheson

Voros scares me. Not that that makes him a bad person or anything, nor does it in any way invalidate his work here (though I think you should have left the "height" factor in, just for giggles).

But he still scares me.

January 25, 2002 - Voros McCracken (e-mail)

Okay, I realize this was a bunch of material to cram into a short article, so there's plenty of questions, let me _briefly_ give a few responses.

Warren on ERA: Warren, one problem with this is that theoretically the difference between the pitcher's ERA and his Component ERA (or whatever) is also defense dependent. If the pitchers defense turns more double plays, throws out more runners on the bases, makes catches in so called "clutch" situations, the credit for those kind of events, if given at all, should go to the defense. Since they have a direct effect on the difference between ERA and a component ERA, that difference is defense dependent.

Mike, the differences I calculated for knuckleballers wasn't in runs, it was done directly to hits allowed (I believe Tim Wakefield still has the largest career difference between his $H rate and his teams among active pitchers), so the factors you mentioned (except maybe the error rate but only slightly) probably aren't the driving force there.

As far as the 20 biggest differences, sure I could do that. Most reasonable work requests I get, I'll do and simply post on my site. If someone here finds one particularly interesting, they can "clutch hit" it. I'll whip up a list today.

January 25, 2002 - Voros McCracken (e-mail)

Colin,

Righties have a small advantage over lefties in the $H dept. Exactly how much is open for debate, and the causes are as of yet unknown (to me at least), but it's around .002 to .003 higher for lefties. Not much.

Sean,

A good question about changing values for older seasons. It should work, but you'll probably need new coefficients, that actually can be derived from the existing ones. (I'll e-mail you with my thoughts).

As far as park factors go, I think you could almost get away with using strictly the home run park factor on the pitcher's home run total and moving on since the hit rate is irrelevant as long as we stick strictly to pitching. Parks, in my experience, don't affect walks much at all so that leaves strikeouts as the only missing link. I'm guessing you'll get closer to what you need with the Homer factor than the run factor. You could run studies using the current run and home run factors, and see how they relate to the component stats, and then extrapolate backward. I'm sure that would have some error involved with it, but would still be valid, I suppose.

Sam,

I just don't know what I'm going to do with you. You see one bad work badge picture of me, and now I'm Ted Bundy. Now, I have to go. I have to get this human head in the freezer before it goes bad.

January 25, 2002 - Arvin Hsu

Voros said: As far as park factors go, I think you could almost get away with using strictly the home run park factor on the pitcher's home run total and moving on since the hit rate is irrelevant as long as we stick strictly to pitching. Parks, in my experience, don't affect walks much at all so that leaves strikeouts as the only missing link. I'm guessing you'll get closer to what you need with the Homer factor than the run factor. You could run studies using the current run and home run factors, and see how they relate to the component stats, and then extrapolate backward. I'm sure that would have some error involved with it, but would still be valid, I suppose. "

But parks do affect $H, do they not? The most obvious example, of course is Coors, and it's inflation of BA. But other parks should also affect $H, such as the long, deep gap in right-center of Pacbell that creates a lot of triples. To get away from defense affects, I presume it might be possible to use a historic park factor that spans different defenses and 10's of years, but then we have all these new parks, to go along with places like Wrigley, where the PF changes each year.

-Arvin

January 25, 2002 - Colin

re: righties vs. lefties:

I wonder if there could there be a variation with the percentages of balls that go to right field against RHP, with RHp getting more balls hit there than LHP (and LHP getting more balls hit to LF)? If so, and assuming teams put better defenders in RF than in LF to begin with, then maybe lefties are subjected to poorer defense in relying more on their LFs?

January 25, 2002 - Charles Saeger

Righties-lefties

You won't believe this, but a right-handed batter is 30% more likely than a left-handed batter to hit the ball to RIGHT field. Same story for a left-handed batter and left field. Apparently, there's no tendency for most batters to pull flyballs.

They do pull groundballs. A right-handed batter is FOUR times more likely than a left-handed batter to hit a groundball to third base, and twice as likely to hit a groundball to shortstop.

I had a hard time believing this myself, so I understand your skepticism. IAE, even reversed, the number of lefty and righty batters a team/pitcher faces won't make a huge dent in the fielding statistics of its/his outfielders.

January 25, 2002 - Alex

"You won't believe this, but a right-handed batter is 30% more likely than a left-handed batter to hit the ball to RIGHT field. Same story for a left-handed batter and left field. Apparently, there's no tendency for most batters to pull flyballs. They do pull groundballs. A right-handed batter is FOUR times more likely than a left-handed batter to hit a groundball to third base, and twice as likely to hit a groundball to shortstop." - Charles

I haven't seen your data, but this does make sense given the following assumption: Groundballs and flyballs are mostly mistakes due to misjudging the speed of the pitch.

Then a batter who thinks a pitch is slower than it is, will be both behind it and under it (expecting it to drop more), sending it as a flyball to the opposite field.

On the other hand, misjudging a pitch in the other direction produces pulled groundballs.

Not a convincing argument, perhaps, but a possible one.

January 25, 2002 - Nate Silver (e-mail)

Voros,

I might be full of Selig here, but are you accounting for the interaction effects between the various predictors of $H?

For example, strikeout rate and home run rate are positively correlated. The correlation is slight, but statistically significant. Seems like a pitcher in the Burt Blyleven mold might be receiving double credit, for example. Also, since WWII, lefties have tended to strike out a few more hitters, and give somewhat up fewer home runs, than have righties.

These interactions create an especially interesting impact as it respects walk rate. One thing first: my model is different than yours; I go ahead and look at a pitchers' $H relative to that of his teammates in conducting macroscopic analyses. I think you're right to say that this can be deceptive when looking at a particular pitcher's performance in a particular year, but I'm aggregating data across a pitcher's entire career, and I think a lot of valuable information (about a pitchers' defense, his ballpark, the league environment) is lost when you throw that information away.

At any rate, when I stuck K rate, BB rate, HR rate and Lefty/Righty into a single regression equation, looking all at pitchers with debuts after WWII and weighting by career IP, BB Rate turned out to be a statistically significant predictor of $H. A higher BB rate, all other things being equal, means more hits on balls in play. If you look at BB rate alone, the result isn't statistically significant, largely because the effect of BB rate is hidden by K rate, which it is correlated with. But if you look at all the factors in tandem, BB rate is in fact at least as significant a predictor of $H as is K rate.

Specifically, my model estimated the following equation:

$H rate = (-.023 * K rate) + (-.057 * HR rate) + (.04 * BB rate) + (.011 * Lefty)

Where:

* $H rate is the number of hits on balls in play a pitcher gave up per inning above and beyond the average of his teammates.

* K rate, HR rate and BB rate are simply K, HR, BB per inning pitched.

* Lefty is a dummy which equals '1' if the pitcher is a southpaw, and zero otherwise.

This is tricky stuff. I'd like to include GB% in the model as well, but I don't have that data.

In fairness, most of this debate is probably academic; the four factors described above are all statistically significant because of the size of the data set, but accounted in combination for the whopping total of 2.33% of the variance in $H.

January 25, 2002 - Voros McCracken (e-mail)

Nate,

I did use multiple regression. In fact, the rate I got for walks was actually in the other direction than yours (I got more walks, fewer hits). And the data set did compare the pithcers $H with that of his team, it is just that once you've found your relationships, then it is time to discard that measure since it is not defense independent.

Also I would be worried about using career rates WRT to $H since there's a selection problem.

Finally, you need to era adjust every stat category in order to make such comparisons useful. I'm not sure if you did, but if not you wind up with serious problems comparing results for pitchers from the 50s with those of the 90s. I believe I have my regression results if you wanted me to send them to you.

January 25, 2002 - Michael

This is really interesting research, so keep it up!

I'd be most interested in which pitchers show a persistent trend over the last three years (minimum 50 IP per season) to have H% above or below the predicted percentage after you've made the various adjustments. I think that would allow us to look at the exceptions, see if they have anything in common seemingly or if they instead are a bunch of random outliers. That might give your assertion that you've made all of the small adjustments that can be found.

On a related note, in the 2002 Baseball Forecaster by Ron Shandler, which you might not have read if you're not interested in future predictions, he mentions several pitchers as seeming to disprove your theory, but (i) Shandler might be using a simplified version of H% and certainly didn't make the smaller adjustments you mention, and (ii) random distribution would still produce a few pitchers that give the appearance of persistance when examined in isolation.

January 25, 2002 - Nate Silver (e-mail)

Voros,

Yeah, it looks like my results are different because I'm weighting based on career IP. If I use a threshold instead -- for example, look at all pitchers with greater than 1000 IP, but don't weight, then the results are quite diffrent (in fact, only handedness retains its statistical significance). Of course, that correction doesn't really resolve the issue of selection bias in the first place.

I think my point really is that, given the very small degree of variation in $H that seems to be predictable, and the profound degree to which the coefficients are subject to change as a result of different methods applied to the same question, DIPS v1.0 might in fact be the more useful stat. It's a slippery slope once skill-based metrics are included as independent variables; DIPS v2.0 might be a little more accurate, but it's a lot less elegant. I'm not saying you'd disagree with any of this.

I also think that $H prevention "ability", to the extent it exists, is probably not very linear. In other words, I believe that a very large percentage of the marginal $H-prevention ability is owned by a very small percentage of the pitcher population. I believe I made this comment on another thread some time back, but something like 35 out of Bill James' top 35 pitchers all fared better than their teammates with respect to career $H.

Anyway, I'd be happy to have a copy of your regression results at the e-mail address above.

-Nate

January 25, 2002 - Mike Emeigh (e-mail)

Further explanation of the knuckleballer's advantage:

An average pitcher, pitching 200 innings, will have somewhere around 500 balls in play converted into outs by his fielders. That figure doesn't vary much for most pitchers. For the knuckleballer, though, it will vary. His fielders will have to get more outs, because some of the strikeouts that he gets that would usually be outs for other pitchers won't be outs for him. If his teammates make more errors than the norm as well, that's extra non-hits put into play.

A knuckleballer, to get 500 outs from his fielders, might have something like 550 non-hits in play over 200 innings, where a normal pitcher might only have 525 with the same quality of defense behind him. The flutterballer in that case is starting at a 25-runner disadvantage - before considering hits. If the team isn't better at preventing hits behind the knuckleballer than behind a normal pitcher, the knuckleballer will be at a further disadvantage - if the team $H is .300, the normal pitcher will allow 225 hits, the knuckleballer about 236 in the same number of innings. (And that's also before considering extra outs by the fielders needed to make up K-WP/PB).

What I'm saying is that the knuckleballer's advantage *has* to exist; for a knuckleballer to survive, he has to allow fewer hits on balls in play than the average non-knuckleballer, because (a) he needs more outs by his fielders than an average pitcher because of the K-WP/PB factor, and (b) there will be more balls that should be outs that aren't outs. That's probably one reason why there aren't more of them in the majors, because you need better defense behind them to get the same results.

-- MWE

January 25, 2002 - Greg

I'm very interested in Mike Emeigh's points about knuckleballers but have one question: If his ideas are right (that knuckleballers need better defense, etc.) then how do we explain the relative success of knuckleballers on bad teams? It's true that some have famously piled up losses--especially Phil Niekro--but still with good ERAs, which suggests that either there's an inordinate number of unearned runs causing those losses or else the losses are caused by the offense, not the defense. Have those teams that have had bad records really been good defensive teams? Those Niekro Braves teams, for example? Or, if not, then doesn't that suggest knuckleballers do well regardless of defense?

January 25, 2002 - Mike Emeigh (e-mail)

In 1979, when Niekro was 21-20 with an ERA of 3.39, he allowed 31 unearned runs in 342 innings - 0.82 per nine innings. That's a lot of unearned runs. His $H that season was .254. Eddie Solomon, the #2 starter on that team, posted an ERA of 4.21 with a $H of .266 - but allowed just 11 unearned runs in 186 innings, 0.53 per nine innings. The average NL pitcher that year allowed 0.49 unearned runs per nine innings with a $H of .281. All data quoted above are taken from Retrosheet event files - the $H are exact totals of H/BIP.

Actually, I may have jumped the gun when I made the comment that knuckleballers need good defense to be successful. We're looking at the results of a ball in play as being a function of something other than the pitcher. The automatic assumption is that, if it's not the pitcher, it's got to be the defense. But there's another factor in this mix - the hitter. Suppose the hitter has the largest effect on the results of balls in play. Most pitchers have fairly predictable movement on their pitches; the fast ball moves one way, the cuve moves another, the slider and the splitter move in relatively predictable ways. The hitter therefore can get his timing down, and can guide the ball more easily. But the knuckleball *doesn't* move in a predictable pattern - that's why pitchers use it. Thus the hitter can't anticipate what the ball is going to do, and thus has less control as a result. Perhaps it is this reduction in the hitter's control over the result of the ball in play that provides the knuckleballer with an advantage.

Or perhaps not. It's a thought, to be pursued. Knuckleballers often end up with bad teams, as Bill James noted, and often pitch much better than one might expect. Bad teams don't often have good defenses, so there's probably something else going on. The knuckleball pitcher may be less affected by *bad* defense than other types of pitchers.

-- MWE

January 25, 2002 - Craig Burley (e-mail)

Mike and Greg,

I'm wondering, though, if knuckleballers are winding up on bad teams that are just bad, or bad teams that specifically are bad because don't have enough pitching.

If knuckleballers are, as they seem to be, used primarily as a last resort for teams that don't have enough pitching, those teams could be bad enough to be bad teams without the need to be poor defensively.

When I say "not enough pitching" I mean, often, literally don't have enough major league pitching to throw the 1450 innings they need to pitch... and then you will be happy to let a Phil Niekro throw 330 innings for you, or Joe 250.

The "knuckleballer-as-last-resort" idea might mean that in fact the defenses that play behind knuckleballers are not bad, but that those teams are bad for a wholly different and understandable reason (no pitching).

I think I need to do a study...

Voros,

What are the historical numbers like for pitchers who throw primarily sidearm or underhand?

If it's the dancing of the knuckler at work, perhaps other unusual pitching styles would give similar results as well...

January 26, 2002 - Art

Voros, I think this has been suggested before, but is there a difference in the slugging percentage allowed per balls in play? This could either be a function of the pitcher's ability or the defense (a quick outfield might prevent more balls in the gaps rolling for doubles or triples). Just curious as to whether the results differ from treating all hits equally.

January 26, 2002 - Mike Emeigh (e-mail)

SLG on balls in play is not independent of $H, so you don't want to evaluate that - you want to look at isolated power on balls in play. I looked at this on the team level for 1999 and 2000, and ISO/BIP tends to vary in the same way as $H - teams with high $H tend to have high ISO/BIP, and vice versa. I suspect you'd find the same thing, as a general rule, on the individual level as well.

-- MWE

January 26, 2002 - Dean Carrano (www) (e-mail)

I see that a similar argument to Neyer's case for Glendon Rusch (3.97) could be made in favor of Texas' Doug Davis (4.07). Does anyone think there's hope for Mr. Davis, or am I in the land of wishful thinking here?

January 26, 2002 - Mike Emeigh

Doug Davis, in 2000, had an ERA of 5.38 with a $H of .307. In 2001, his ERA dropped to 4.45, while his $H rose to .329. His ERA fell primarily because he cut his walk rate significantly.

Texas has had an awful defense the last two years, and they haven't made a significant effort to address that in the offseason, from what I can see - although neither Blalock nor Perry can be much worse at 3B than Lamb, to be sure, and Everett isn't a bad defender, I just don't see where they're going to get a significant upgrade. The team $H was .314 in 2000, .312 last year. Davis should come back to the league some on the $H, but he still walks more guys than the league average and strikes out quite a bit fewer hitters than the average. I don't consider it likely that he'll get much better (not that a mid-4s ERA in the AL is too shabby). If he were pitching in Seattle, on the other hand, you might see another Jamie Moyer-type season.

-- MWE

January 26, 2002 - Colin

Charles - Wow, thanks for the info on righties and lefties. Very interesting indeed.

January 27, 2002 - Herlin

I think the reason that lefties allowed a higher hitting average when balls put into play than righties is that lefties face more right-handed batters and thus having more groundballs pulled to 3rd base/shortstop. And 3rd basemen are more prone to errors than 1st basemen.Maybe some of this errors will actually be recorded as infield hits?

What inspires me to think about this is that I've read an article on BP discussing another topic: do the error ratio of opposing fielders depends on the particular batter who's at bat? Most of us would assume that the ultra-fast players putting pressure on infield will force opposing fielders make more errors. Well, it's not exactly so; some players topping the list are just the opposite -- right-handed (but not left-handed) sluggers that are not necessarily fast. It appears they frequently pull hard grounders to 3rd base and inevitably there'll be larger ratio of error on this kind of batted balls. Left-handed sluggers, if they hit a ground ball hard, will hit it to 1st base and 1st baseman have more room for mistake like bobbling the ball a little or such since they dont have to make the long throw like 3rd baseman.In fact, of course we all know that among infielders 3rd basemen generally have the lowest average FPCT while 1st basemen have the highest.

January 27, 2002 - Herlin

I think the reason that lefties allowed a higher hitting average when balls put into play than righties is that lefties face more right-handed batters and thus having more groundballs pulled to 3rd base/shortstop. And 3rd basemen are more prone to errors than 1st basemen.Maybe some of this errors will actually be recorded as infield hits?

What inspires me to think about this is that I've read an article on BP discussing another topic: do the error ratio of opposing fielders depends on the particular batter who's at bat? Most of us would assume that the ultra-fast players putting pressure on infield will force opposing fielders make more errors. Well, it's not exactly so; some players topping the list are just the opposite -- right-handed (but not left-handed) sluggers that are not necessarily fast. It appears they frequently pull hard grounders to 3rd base and inevitably there'll be larger ratio of error on this kind of batted balls. Left-handed sluggers, if they hit a ground ball hard, will hit it to 1st base and 1st baseman have more room for mistake like bobbling the ball a little or such since they dont have to make the long throw like 3rd baseman.In fact, of course we all know that among infielders 3rd basemen generally have the lowest average FPCT while 1st basemen have the highest.

January 28, 2002 - Mike Emeigh (e-mail)

And 3rd basemen are more prone to errors than 1st basemen.Maybe some of this errors will actually be recorded as infield hits?

The rate of errors that allow runners to reach base has fallen considerably. James used 71% in his original calculations for DER, and Clay Davenport used 70% in his defensive system in 1998. The rate over the last three years has been a lot closer to 60% - and more to the point, the rate of throwing errors on infield hits has risen to about 17% of all errors. I am absolutely certain that many of these would have been scored as straight-up errors in times past.

I would suspect that RH-hitters would, in fact, get some benefit from this effect, and that LHP would be hurt by it.

-- MWE

January 28, 2002 - Vinay Kumar (e-mail)

In fact, of course we all know that among infielders 3rd basemen generally have the lowest average FPCT while 1st basemen have the highest.

But first basemen have a high fielding percentage because most of their chances are on catching throws. If you eliminate catching throws (and the errors on those plays, like dropped throws), then what kind of FP do first basemen have? Obviously, they have an advantage because they have time to recover from a bobble (and also have the luxury of making unassisted putouts, rather than risking a throw). But 1B are usually the least-skilled infielders on fielding ground balls. I'd be curious what the fielding percentages would look like if catching throws was removed.

January 28, 2002 - Mr. Brett

Gratuitously Lazy Fantasy League Research Request

Which pitchers can we expect to improve their ERA in 2002 due to either:

1) An abnormally high hits per balls in play rate in 2001, due for "correction" in 2002

2) A pitcher with average or worse H% moving to a team with a better defense

In the interest of not being a complete cherry picker, I'll offer James Baldwin as a possibility. He'll be moving to Seattle, one of the better defensive teams in the league.

January 28, 2002 - Mike Emeigh (e-mail)

Using my 2000 play-by-play database (don't have accurate data for 2001), fielding percentages for 1B when you eliminate catching throws, with a minimum of 200 chances excluding catching throws, ranged from .988 (John Olerud) to .933 (Andres Galarraga). 11 of the 22 1Bs who qualified for this list were at .977 or above, and only 3 1Bs (Mo Vaughn, Kevin Young, and Galarraga) were below .958. 3B fielding percentages are generally lower, with the center being around .960 or so.

-- MWE

January 29, 2002 - tangotiger (www) (e-mail)

Mike, great list. Does the 1B information also remove the errors on catching throws? Does the pbp database show if the error was on a catch?

January 29, 2002 - Ray Kerby (www)

Voros, As usual, good job. Although DIPS is revolutionary for defensive analysis, I think that you are watering down a simple and important concept by niggling over details.

The notion that pitchers are not responsible for the rates of singles, doubles and triples for balls put in play is a big improvement in and of itself. However, I think that you unnecessarily complicate the concept and make it less accessible by trying to fine-tune its accuracy to the nth decimal.

DIPS stands on its own without you having to mangle it with statistical minutia. If you are doing this to respond to critics, it is unnecessary. No one is trying to tweak batting stats with adjustments for height, weight or batting stance.

It seems that simple, defense-independent "batting against" lines for pitchers is the simplest and most direct route to take.

January 29, 2002 - Mike Emeigh (e-mail)

I removed as many errors on catching throws as I could cleanly identify in the database. Usually on such plays, there is an error and an assist charged (e.g. 4E3). On some occasions, I couldn't make a judgment one way or the other from the information that I had, and I left those in although I suspect that they were probably missed throws; for example: 64/FO.B-2(E3) is almost certainly a missed throw, but I left it in because I couldn't be 100% sure.

Vinay, in his comment, was asking whether fielding percentages for 1Bs would be higher than 3Bs even when you eliminate plays that 1Bs get for catching throws, and they are; the center for 1Bs is around .975 (between .971 and .977, actually, for 2000) while the center for 3Bs is around .960. Removing a few additional errors would only push that 1B center even higher.

-- MWE

January 29, 2002 - Voros McCracken (e-mail)

The gratuitously lazy Mr. Brett,

Dave Burba looks like a decent bet despite going somewhere no better tahn Cleveland.

Jason Johnson ought to sink like a stone.

Joe Mays and Mark Buerhle are unlikely to be Cy Young candidates again this year but could be decent nonetheless.

Glendon Rusch looks like a very good choice.

Watch out for Kerry Wood if he remains healthy, could be a Cy Young candidate.

Derek Lowe as a starter looks to be a strong "value" play (though in a roto situation the lack of saves might mess with things).

I like Terry Adams behind a fairly good defense in Philadelphia.

I don't have a lot of confidence in Jon Garland.

Rick Helling should be better in Arizona.

--

I'll have projections for pitchers finished in as soon as a week and as late as two weeks. It depends on how much I get done this weekend.

Hope that helps.

January 31, 2002 - tangotiger (www) (e-mail)

Mike, good stuff.

Voros, I concur with Ray. There's something to be said for doing 90% of the work to gain an extra 10%. My suggestion is that "basic-DIPS" should always be presented, along with an "appendix" for V2.

Mike: I didn't realize that the 2B would get credit for an official assist on an error play.

February 1, 2002 - Mike Emeigh (e-mail)

Rule 10.11 states:

"An assist shall be credited to each fielder who throws or deflects a batted or thrown ball in such a way that a putout results, or would have resulted except for a subsequent error by any fielder."

The Rules of Baseball can be found at the MLB Web site in the Baseball Basics section.

-- MWE

February 5, 2002 - Doug (e-mail)

Just wondering if you checked whether plain old ball/strike pitch count ratio correlated to hits allowed on balls in play. This has some attractive features: easy to understand why it would be correlated (if it is), very strong defensive independence. Obviously there's not a ton of data to work with, but there must at least be several years worth by now.

Also, re: other comments on related effects, noticed that you're almost ready to include walks in your model and, with Ks already there, the name Nolan Ryan immediately came to mind. Same thing re: HRs and Ks for Bert Blyleven. I think you mentioned that this sort of double effect is all taken care of, but maybe you could explain simply how that would happen - are there thresholds that can't be passed or arbitrary rules that no two factors can contribute more than something, or what?

Interesting stuff, but is it really that much more useful than the basic version. I mean, are guys who were nowhere with version 1 now in the elite class, or is it just a case of adding a few points to your correlation coefficients as you briefly described in your piece?

February 11, 2002 - David Smyth

A few recent posters (Ray, Tangotiger, Doug) seem to be suggesting that Voros should have stuck with the simpler and cleaner DIPS version 1. I disagree. It is my belief that _perhaps_ Voros should have stopped at the point of presenting tables of $H for the pitchers, but once the decision was made to do DIPS, why not take it as far as possible so that we can get the best possible picture of how much of defense is pitching, and how much is fielding?

February 11, 2002 - Mike Emeigh (e-mail)

David Smyth:

once the decision was made to do DIPS, why not take it as far as possible so that we can get the best possible picture of how much of defense is pitching, and how much is fielding?

There are several reasons why a person might want to stop short of a full-dress treatment:

1. Introduction of additional adjustments that add minimal additional accuracy at the expense of ease of use tend to turn people off. How many people followed all of Bill James's gyrations through the various technical Runs Created formulae, and how many people used one of the simpler forms which were less accurate but easier to remember and use?

2. Like all sabermetric methods, DIPS is a model of reality, and not an exact representation of reality. There will always be areas of performance that the model can't explain, and you cannot logically force the model to try to explain them, so you shouldn't try. It's certainly useful to look at things that the model doesn't explain, such as the knuckleballer's advantage, but it's not necessary to try to account for each and every one of those effects in the model itself.

3. You run the real risk of decreasing the model's overall accuracy in an attempt to model a particular parameter that affects the accuracy of the model in some cases.

Taking into account accuracy, clarity, and ease of understanding and use, DIPS I was (IMO) about as good as one could hope to achieve. I'm not suggesting that Voros shouldn't TRY to make adjustments and attempt to model other aspects of pitching and defense not covered by the model, not at all. I do suggest that the finding that pitchers have little to do with the results of balls in play is FAR more important than nailing down the exact value of "little", and that for almost all aspects of defensive analysis, the assumption of 0% impact implicit in DIPS I does not introduce a significant error if the impact were later found to be some small positive value.

-- MWE

February 12, 2002 - tangotiger (www) (e-mail)

The most important feature, by far, of DIPS is $H. That is, the player's $H and his teammates $H are very similar. In fact, if you look at it throughout history, you will find that 90% of all pitchers with at least a certain number of balls in play (I think it was 1000, but I can't remember) will fall within 5% of his teammates' $H. That is the startling discovery, and I agree with David that just presenting this list of $H would have been enough.

This fact I find loses its impact in the refinements that Voros is bringing in.

February 13, 2002 - Neil deMause (www) (e-mail)

Chiming in late here, but I am again blown away by the work done by the ever-scary Voros. And 90/10 rule or no, we can hardly complain about having two excellent options for evaluating pitchers' value - the world has survived nicely having both OPS and EqA, for example, so I see no problem with letting a million flavors of DIPS bloom. (Voros' sleep patterns notwithstanding. He does sleep, doesn't he?)

Okay, all that aside: I'd really like to hear either Voros or Mike E. or someone elaborate more on whether SLG (or ISO)/BIP is really pitcher-independent. It seems completely counter-intuitive, for a couple of reasons: 1) there's always been an impression that certain pitchers, like Tommy John or Ramiro Mendoza, survive by giving up lots of hits but mostly singles, and 2) since preventing HRs *is* a significant factor in pitchers' success, you'd think fewer HRs would translate into fewer doubles off the wall as well. Of course, this whole discussion is deep into counter-intuitive territory, but I'd love to see some hard stats before I give up on yet another cherished piece of Baseball Common Knowledge. Anyone?

February 13, 2002 - David Smyth

Mike Emeigh: (BTW, how is that pronounced?)

Thanks for your thoughtful response to my post. But your first point boils down to personal preference, and your 3rd point essentially means that Voros must be careful in his evaluation and judgement of his research, which is true enough but is generally true anyway of everyone who deals in sabermetrics.

Your 2nd point assumes that the "model" is, or should be, exactly what your interpretation is. But if the model is simply to try to estimate pitching ability by using only defense-independent categories, then I think DIPS 2.0 qualifies.

And your final, un-numbered point that the difference is not "significant" is well taken by me (a champion of accurate simplicity), but is really, like your 1st point, just a matter of personal preference. There are probably plenty of people who do not see the differences between DIPS 1.0 and actual ERA as significant. And there are also, I'm sure, people who view the DIPS 2.0 refinements as not enough, in the search to pin things down to the 4th decimal place.

In a practical sense, if you look at the actual DIPS 2.0 procedure in an article on Voros' site, the extra steps are not really that difficult or time-consuming.

February 19, 2002 - Paul Westkaemper

My apologies if anyone else posted this and I didn't catch it as I read through, but it seems to me the LH/RH advantage is very straightforward. 2b/1b throws are substantially shorter than 3b/ss throws. This allows the defender to play a little further back, and gives him more time to make the play. So in the end a 2b can cover more ground than a shortstop and still get the ball to 1b in time, and so more GBs are turned into outs.

On the flyball side the throwing again works to the RH advantage. LFs have the shorter throw, so the Vince Coleman types (all speed no arm) end up in LF, while the strong arm, but slow OF will generally end up in RF.

This is all sophistry on my part at this point, but

February 19, 2002 - Mike Emeigh (e-mail)

Paul:

Actually, the explanation is a lot simpler than that.

1. When hitters pull the ball on the ground, they tend to hit it into the hole or down the line. 2. When hitters don't pull the ball on the ground, they tend to hit it away from the hole/line, and most often wind up hitting it directly at the fielder.

This is true on both sides of the plate. Because there are more RH-hitters than LH-hitters, RH-hitters have a net advantage in this category, even though the rate of conversion on balls hit into the hole on either side of the diamond is virtually the same. There are actually a (slightly) smaller percentage of balls hit into the 1B-2B hole converted into outs than there are in the SS-3B hole, which is somewhat unexpected, and there are "far" fewer balls hit up the 2B side of the middle converted into outs than there are on the SS side, which is expected since 2Bs are moving away from first on those plays while SS are moving toward 1B. 2Bs do convert more balls hit directly at them than do SS, which makes up for some of the difference. Furthermore, as you would also expect, 1Bs convert more ground balls hit down the line into outs than do 3Bs, which also disadvantages LH-hitters. The result of all of this is that, on ground balls, LHB make outs far more often than do RHB.

The argument on fly balls is on more solid ground, in that players who can throw but don't move very well play RF and players who can run but don't throw very well play LF (if they can do both they play CF). Having players who cover less ground in RF also tends to favor RHB, because fly balls are usually hit to the opposite field. Because RHB are favored offensively on both fly balls and ground balls, RHP would be favored defensively, the effect that Voros noted.

I should point out that this effect is, again, mostly independent of the type of pitcher. The distribution of balls in play by hitters does not change significantly based on the handedness or GB/FB tendencies of the pitchers.

-- MWE

February 19, 2002 - Mike Emeigh (e-mail)

David Smyth:

FWIW, my last name is pronounced as though it were spelled "Amy".

But your first point boils down to personal preference, and your 3rd point essentially means that Voros must be careful in his evaluation and judgement of his research, which is true enough but is generally true anyway of everyone who deals in sabermetrics.

Actually, it's not generally true that everyone who deals in sabermetrics takes that level of care, although it is most certainly true of Voros - but that's another thread.

I should clarify that when I talk about DIPS being a "model", and when David calls it an "estimator of pitching ability", we are really talking about the same thing. The model underlying DIPS is that you can gauge a pitcher's real level of performance by considering only his ability to prevent HRs and walks and to strike hitters out, and that because hit prevention is NOT primarily a pitching ability you do not need to look at that for an individual pitcher. The numbers that come out of DIPS are estimates of that ability level, based on the assumption that the underlying performance model is an accurate reflection of reality.

In evaluating a model, you need to ask two questions:

-- How accurately does this model reflect reality? -- How well does this model correlate with reality?

These are NOT the same question, expressed two different ways. The accuracy question asks "How close is the estimator of ability to the actual ability?"; it doesn't ask if the estimator overshoots on one player and comes up short on another, only if the net result is small. The correlation question asks "How likely is it that if the model ranks Player A as being better than Player B, that Player A is actually better than Player B"? - it doesn't ask whether the difference between the two players is 10 perfomance units or 1000, as long as A>B on both scales.

A model can correlate well with reality and be terribly inaccurate, if it fairly consistently underpredicts or overpredicts. A model can also be quite accurate, yet not correlate very well with reality if the errors tend to offset one another, or if players tend to cluster in performance groups with relatively small variations in performance between groups. You cannot rely on one or the other as your sole indicator of a "good" model.

And what I tend to see happening is that people forget to balance the two views of the data. People will either distort the model away from a viable correlation with reality in an attempt to get the model to "add up" or they will focus on things that make the model correlate better, by modelling some small additional performance feature, and forget to check whether the model becomes less accurate for players as a group when the small effect is added.

And that's before we even consider the effect of the "ecological fallacy", where in looking at the performance of players we make inferences on aggregate data (team-level data) that may be unreliable as a result of bias resulting from the aggregation of the data. See this link for a discussion of this problem. (If Walt Davis is reading this, thanks for steering me in this direction.)

The problem that I have with complex models is that they imply a level of accuracy that very likely does not exist, because of the problems that I outline above. It's why I think my first point, which David calls a matter of "personal preference", is really something more than that. Maybe the simplest models are the best not just because they are easy to understand, but because they capture most of the information people need to know without any pretense at being more accurate than they likely really are.

-- MWE

February 21, 2002 - Ben

Way back on January 25, 2002, Mike Emeigh posted: "The automatic assumption is that, if it's not the pitcher, it's got to be the defense. But there's another factor in this mix - the hitter. Suppose the hitter has the largest effect on the results of balls in play."

Has anyone followed up on this? Do some batters have consistently high or low $H?

-- Ben

February 21, 2002 - John

A little more on the knuckleballer...

If you ever watch a game with a knuckleballer involved...more times than not you'll see the hitters off balance when swinging. Obviously due to the speed and movement of the ball hitters have to try and adjust. In the end, most hitters can't make the necessary adjustment and will not make "solid" contact. Without solid contact, safe hits will undoubtedly drop.

In the same light, slap hitters like Ichiro would suffer. While hitters like Ichiro count on just making contact and having the pitcher provide the needed energy to put the ball in play...preferably on a line in the gap...knuckleballers do not provide enough energy to just spray the ball. Hitters have to be able to generate more energy on their own for the ball do what it would against a "normal" pitcher.

In my opinion...which doesn't really mean much, in my opinion...hitters have a big hand in how well knuckleballers DIPS numbers look.

With that said, catching also plays a role. I would venture to guess that the majority of unearned runs charged to a knuckleballer involve the catcher in some way whether it's PB, WP, or a botched third strike. It could even be throwing errors trying to catch would be base stealers because knuckleballers are more vulnerable to stolen bases. I wouldn't blame "poor" defense for some of their unearned runs, I would venture to suggest that they brought it onto themselves.

I've never done any research into this matter, this is coming from pure speculation based on experience playing and just taking in a game or two (or a million!!! hehe). I've had to joy of trying to catch for a lefthanded knuckleballer as well as hit against several knuckleballers...and I hate it!!!

Oh, one other thing...as a lefthanded hitter...growing up, in my district little league league...we only one a handful of lefthanded pitchers so my experience against lefties was very limited. And I still to this day feel uncomfortable against them.

Nowadays, lefties are revered...unlike the old days. So, I think you are going to begin to see more lefthanded pitchers come up through the ranks and as a result, I think they're effectiveness will begin to diminish in the next couple of decades...if it hasn't already.

Again...purely observation.

Thanks, John

August 21, 2002 - Nutzman (www) (e-mail)

I bet half the people here don't reveal their best research in hopes of making money off of it...

 

You want baseball news?  We've got baseball news!

Baseball Skeptic

Weblogs
Clutch Hits
 Last post 2:03am, Feb 09
Game Chatter
 Last post 9:36pm, Jan 27
Transaction Oracle
 Last post 11:20pm, Feb 07
Sox Therapy
 Last post 1:04pm, Feb 08
The Hall of Merit
 Last post 5:14am, Feb 09
Primate Studies
 Last post 12:24am, Feb 08
Al's Baseball Tidbits
 Last post 2:50pm, Oct 26
Columns

Dialed In
Visitors' Dugout
Cooperstown Confidential
Swinging from the Heels

Archive

Voros McCracken Archive





About Baseball Primer | Write for Us | Copyright © 2001-2003 Baseball Think Factory
User Comments, Suggestions, or Complaints | Privacy Policy | The Fine Print | Advertising