This is a great book for answering a lot of classic baseball questions using data and sabermetrics instead of speculation. I enjoyed it tremendously, but the book definitely had its shortcomings - it was written for a VERY lay audience in an occasionally aw-shucks goofy tone, it didn't include a lot of the formulas for their statistics so you had to take the reliability of their measurements on faith, and the thing that annoyed me the most, it is in the end a collection of articles: the editing was very spotty, so sometimes great references were made to other chapters, but at other moments the text explained what a correlation coefficient was for the 17th time. Those faults (and a bit of a rhythm problem with having all the "boring" baseball financial chapters packed in the middle) aside, this was a very solid read, and confirmed a lot of the rumors I've heard from the sabermetric crowd as well as making my own thoughts more precise. Here's the question by question breakdown from this thoroughly enjoyable book (written in a mock "9 inning" format)
Batting Practice - Is Barry Bonds better than Babe Ruth? This is fairly complicated argument, coming down to the fact that the Babe was better relative to his own league, but Bonds is as great as he is against far superior (athletic, integrated league, superior equipment / nutrition, and drawing from the world instead of just the States) competition. Among other things, this article points out that the men's Olympic record for the 100 meter freestyle in 1913, the year Ruth started, was 63.4 seconds; now it's 48.2 seconds, and girl's high school swimmers often break 60 seconds. (The article mentions how ludicrous it is to say that expansion has diluted baseball talent; in a vacuum, yes, but now the league is drawing from a population almost 7 times larger (700 million v. 110 million)). The article is very interesting in that it devalues Ruth's stats for having played in his relatively smaller-competitive times, but then must add value to them for not having access to the science that Bonds has had. In the end, Ruth edges out Bonds, barely, and it actually comes down to his pitching contribution from his early career.
1.1 - What's the matter with RBI? Everything, it turns out. Unsurprisingly, RBI stats are strongly tied to RBI opportunities, and usually managers put their best hitters in the spots where they will accrue the most RBI opportunities. So it's not surprising that the big bashers end up with lots of RBIs, but the cart is before the proverbial horse. The article picks like mad on Joe Carter, who managed to have 100 RBI seasons where his hitting VORP (value above replacement player) was near zero, meaning they could have gotten equally good hitting from a minor leaguer off waivers. In short, RBIs ain't all that.
1.2 - Is David Ortiz a clutch hitter? In short, no - clutch hitting is not a repeatable skill, meaning that people who hit well in the clutch in one year have no increased statistical probability to do so in the following year; their year to year clutch stats have no correlation, and getting hits in the clutch is often a matter of luck (i.e., good timing). This article defines clutch hitting by late situations where a hit stands to greatly increase your team's chance of winning the game - for instance, if your team is down by one run with one out in the bottom of the ninth and you hit a double, you increase your team's chance of winning from 3.9% to 11.2%, a +7.3% change. After heavily crunching the numbers, it turns out that if there is a clutch hitting skill, it has about 2% of the impact (70% of it is general hitting skill, 28% is blind luck). And it turns out that even this small amount of clutch shows up best in hitters like Mark Grace, the career leader in "clutchiness," meaning that this probably has something to do with your batter's eye or ability in situational hitting. SO, sorry, Big Papi.
1.3 - Was Billy Martin crazy? by James Click - Billy Martin once picked his lineup out of a hat, and James go on to discuss that optimal batting orders pretty much just descend in order of OBP, but even that said, optimizing a batting order would have very little effect on the runs a team scores. Batting order, it turns out, is nowhere near as important as people think.
2.1 - Why are pitchers so unpredictable? A much ballyhooed stat discovered recently is that pitchers have very little, if any, control over what a ball in play does - they can have a tendency to produce groundballs instead of flyballs or vice versa, some pitchers show a small ability to induce pop-ups, but otherwise its the defense that determines the percentage of balls hit in play that become outs. Pitchers DO have control over preventing homeruns, striking batters out, and not hitting / walking batters. So on the one hand, there is simply a lot of variability from year to year on what happens to balls in play, AND pitchers hits/walks can get clustered somewhat randomly, so pitchers are less consistent than you would think. On the other side, the problem is in our measurements - ERA is an unforgivably dumb stat that hinges on what "would have happened" and is at any rate susceptible to a bad defense; a simple RA is better but similarly susceptible to the defense's ability, and a stat like wins and losses is absurdly out of the pitcher's control in a lot of cases. Add on top of this that pitchers get injured easily and - well, at best you can say that pitchers are relatively consistent at an inconsistent job, and we have been looking at them incorrectly for a long time.
2.2 - Are teams letting their closers go to waste? Yes - if the closer is really that good, he should be brought in the situations where he can have the most impact, not just the ninth inning when you have a lead. Baseball managers, notably Joe Torre, tend to behave more properly in the playoffs. Bringing in a closer in the ninth with a three run lead is a waste; bringing him in as the home team down by one in the top of the ninth is not. It turns out that dire situations where it would be appropriate to use your closer start in the sixth inning and beyond, not that anyone's going to start doing this any time soon.
2.3 - Five starters of four? Your top four pitchers should be better on average than your top five. There is no evidence that three days of rest is worse injury or performance-wise than four; the injury evidence points towards too many pitches within a given start as predictors of injury and poor performance. So if you started only four pitchers and had them throw fewer innings / pitches per game (but had them making more starts), you would actually gain in total innings pitched by each of your four best starters. Given that the fifth guy is usually a replacement level type, this sounds like smart thinking.
3.1 - When does a pitcher earned an earned run? - Another article investigating the correlations of ERA from year to year and pointing out that there are a dearth of factors that are out of the pitcher's control when it comes to his ERA.
3.2 - Did Derek Jeter deserve the gold glove? by James Click - Nope. Despite the flash, Jeter is notoriously slow-footed, low range defensive shortstop. Actually, the flash is probably due to his lack of range - he has to dive, leap, and otherwise show off on plays that better short stops would make routinely. In the best line of the book, James said that Jeter's nickname should be "Pastadiving." Sound that one out. The year he won a golden glove, though, Jeter made a big improvement in his D - unfortunately, this improvement was from terrible to mediocre. So he didn't deserve it, and just to rub salt in, James showed how obvious it is that, given the normal continuum of defense, that Jeter should be the one playing 3B and A-Rod short - Jeter is costing his team runs.
3.3 - Is Mike Matheny a catching genius? Guess what, no. There's no statistical evidence that certain catchers will reliably improve their pitcher's ERA. Their effect on the opponent's baserunning game is palpable, but past that, there are huge swings in the measures of other aspects of a catcher's performance from year to year (notably the article does not explicitly mention pitch-blocking, but one can probably guess that this means that the base level of blocking pitches in the dirt at the major league level is not far enough from the top performers in this area to produce big differences in runs or wins).
4.1 - What if Rickey Henderson had Pete Incavilia's legs? by James Click -or really, how much is the running game worth? Extensive article pointing out just how little Ricky's running ability positively contributed - about 5 wins over the course of a 25 year career. Largely due to the fact that while stealing bases is great, making outs on the basepaths is god-awful. And all of those ideas about a good baserunner distracting the pitcher and the defense have proved untrue. So speed makes the game exciting (and let's not forget helps a lot in the field), but its impact is very small.
4.2 - When is one run worth more than two? by James Click After pointing out how routinely futile and more often than not, detrimental bunting is to a team's scoring chances, this article compares big inning strategy v. small inning strategy. The most interesting this it pointed out is that when thinking of run expectancy, don't forget to include the probability of scoring "at least one" as a factor - there are lots of situations where it turns out that playing for one run is appropriate, even though over the long haul you would score more on average by doing other things. (The obvious one - tied in the bot 9th - those extra runs, or any strategy that aims at them at the cost of the probability of scoring one, is bad).
4.3 - Is Joe Torre a Hall-of -Fame manager? by James Click - Yep, but you're going to have an impossible time finding any numbers to back that up. All of the things attributed to managers do nothing tangible for wins - that's pretty much the summary of this one. Good, detailed article, and nailed the idea that it's largely the tools and not the blacksmith.
5-1. Why is Mario Mendoz so important? This is Mario of "Mendoza Line" fame, and he pretty much exemplifies the concept of a replacement player, the very bottom level of offensive output you would tolerate in the big leagues (in his case, actually, his defense did enough to make him slightly above replacement player, but you get the point). Players like Mendoza point out just what a rare commodity an "average" player is, and doubly points out that even the Mendozas are in the 99th percentile of baseball players. The whole concept of VORP has been vital to shifting the concept of statistics; not just performance in quantity, or relative to the average, but relative to what you could easily acquire for no money.
5-2. Is Alex Rodriguez overpaid? Yes. This goes into excruciating detail of the amount of money that wins actually bring in for the club, and A-Rod's 25 million per year is probably around twice what his ability would be expected to bring in on average. Egads. (I actually found this article to be the most statistically suspect - they seemed to derive a lot of figures from certain teams and extrapolate to others; I just don't know if you can get at the smoky financials of the trust league). I tuned out a little bit here (sorry).
5-3. Do players perform better in contract years? In short, yes, but barely. This article stresses the point that most players in contract years are at or slightly past their prime, so basing any future predictions on that year's performance is dubious anyways. Again, I don't like the money side of things.
6-1. Do high salaries lead to high ticket prices? No, and if anything, the other way around is more likely. The ticket prices follow classic supply-demand economics,a nd usually increasing the price of the tickets will cost you money - so doing something that costs money in addition to paying the higher salaries is not too good.
6-2. Are new stadiums a good deal? See what I mean about the economics section in the middle of the book? Turns out that a publicly funded new stadium is a fantastic deal for the owner, a bad one for the taxpayers, and that even privately funded stadiums are often publicly funded via tax breaks and such. And most teams do seat a boost (if a short lived one) after a new stadium goes up.
6-3. Does baseball need a salary cap? Mildly interesting article pointing out that all of this "competitive imbalance" is not what it seems - true, the Yankees and Braves are there pretty much every year, but year-in and year-out, then winners and playoff contenders (and the world series winners) have varied a lot. Actually, this article was a solid look at the details of revenue sharing and such, and points out that the size of a team's market has very little to do with its on-field performance.
7-1. What happened to Todd Van Poppel? This article addresses the question of the relative merits and risks of drafting players out of high school or college. Turns out a high school super-prospect, one you would take with the #1 overall pick, is a legitimate choice, but after that, college prospects, particularly college pitchers v. high school pitchers, are a much safer bet. College hitters are the safest bet; all of this revolves around "averages," so it's hard to argue when to take chances.
7-2. Is there such thing as a Quadruple-A player? Yes, but the argument is that hitting is part ability (say 90%) and part experience (10%). So take one inexperienced but skilled young AAA player and one older experienced but not as skilled AAA player who are performing the same (both well in AAA) and bring them up to the majors. Now, both of their experience is wiped out (because of the huge leap from AAA to MLB) and suddenly the kid with skill has a lot more to fall back on than the older player. This is theoretically where you find the AAAA guys - ones that don't perform well enough in the minors to be noticed until experience is playing a too large role, also keeping in mind that these guys would be more likely to be on the decline anyways.
7-3. Why was Kevin Maas a bust? Kevin Maas was a slow, strike-out and walk heavy heavy player who came up to the majors and hit about 21 homeruns in his first less-than-half-of-a-season. He then flopped badly. The idea professed in this article is that there are "young skills" (average, speed) and "old skills" (power, drawing walks). Young skills, obviously, deteriorate eventually, and you must rely more on your old skills to compensate. So someone who comes up with only old skills, like Mr. Maas, is already performing at max capacity, and the only thing that's going to deteriorate is his already old skill set. Cool article, especially if you look at case study like Geroge Brett who started his career utilizing his young skills and finished off adapting and using his old skills.
8-1. Can a team have too much pitching? Theoretically, no, and according to the pythagorean theorem, pitching runs are actually more valuable than hitting runs. But given that pitching is a finite resource purchased with a finite resource (wampum), you can't just add pitching ad nauseam, and it turns out that, shockingly, a good balance of hitting and pitching and defense, and correctly identifying your situation and making the appropriate adjustments, is the way to win. This may well have been the "well, duh" chapter.
8-2. How much does Coors Field really matter? A whole lot.
8-3. Is Wayne Huizenga a genius? That's the GM of fire-sale fame for the 1997 Florida Marlins. Another money chapter; they said yes, and had some pretty decent, though heartless, reasons.
9-1. What do statistics tell us about steroids? Pretty weak argument here - but it first pointed out that the number of "sudden spikes" in the so-called steroid era were no more than other eras (actually, the "Greenie Era," the 1970s, had more spikes). Then it pointed at the players who have been caught, and noted their performance before and after their suspensions, and showed a very small (if any) decline in performance after they stopped. It then went on to question why marquee players would bother... blah. This article very blandly said "we couldn't find anything" about a topic that is hard to find anything out about. Boo-urns.
9-2. What does Mike Redmond know about Tom Glavine? by James Click. The basic argument is sound; that the batting samples oft-cited to claim that a certain batter owns a certain pitcher are absurd. And James correctly points out that you can't apply a normal distribution to an individual batter's 48 at-bats, because they're not a random sample; so even if you said something like "if Mike Redmond's actually a .284 hitter, there's only a 0.5% chance that he would have hit .438 or better against Glavine," you are pulling some goofiness there; yeah, you would reject that .284 hypothesis, but of course you would - he hit .438 against him in the entire (small) population! SO while that logic is good, I don't like the following logic - if you put all batters who have 50 plate appearances against Glavine in a distribution, that distribution is normal, and Redmond merely falls in one of the places that someone inevitably would have. I feel like that's a slight of hand, a "looks like a bell curve, smells like a bell curve..." statement. The fact remains that Redmond IS at the high end of that curve, and the others are not - true, the fact that it's a bell curve means that we don't KNOW that Redmond owns Glavine, but we certainly wouldn't reject the notion based on this either. I think James leans too hard the "no one owns anyone" way. Additionally, he brings up the fact that pitchers who have faced people 50 times are likely to be good pitchers, but he didn't mention that batters who have faced a pitcher 50 times are also likely to be good batters.
The article goes on to make some fine points - that teasing out these differences is really the key, like the famous lefty-batter v. righty pitcher advantage. If someone can figure out the other trends, like who hits ground ball pitchers better, etc., then you may get some serious platooning going on.
9-3. Why doesn't Billy Beane's shit work in the playoffs? A long examination of the small underperformance that the A's have suffered in the recent playoffs - of course, the playoffs are a small sample, so a lot of their losses are due to bad luck, but even in the small sample, the A's have performed worse than expected. It turns out that in a competition fought in the cold weather between the best pitchers and hitters, it turns out that defense, pitching and especially relief pitching have disproportionately large impact. So Beane's slow-footed starting pitching heavy teams were not ideal for the conditions, and thus lost more than they should have. But remember, in 2003, they were one nasty Derek Lowe curveball away from a ALDS win - so it's not like they purely failed.
Extra Innings - Can stats and scouts get along? They need to - after an entire book of sabermetric articles, I started to notice a lot of the weird assumptions that stats make - that players are unfluctuating, that a player is *really* a .270 hitter even if he hit .290. It's the old one foot in boiling water one foot in ice argument, but the general disdain some scouts show for the stats is highly understandable. Plus, there's an entire aesthetic aspect to baseball that sabermetrics ignores - and that may be obvious, but hear me out. In hockey, the neutral zone trap can be most effective, but it's stunningly ugly. In basketball, an overbearing in your shorts physical zone defense may be very effective, but it would reduce things to a jump shot contest. In football, a team that could just run dives over left tackle for six yards a pop would be impossible to stop, but... you get the idea. A sports efficiency ideal could be an aesthetic nightmare, and the rules get changed accordingly - usually to "improve the product" - gag - but it's a real problem, that scientific pursuit of the ideal offense could expose the fundamental monotony of the game. I sometimes get the feeling that the logical end of the sabermetric argument is a robotic game - and while that awesomely serves front-office purposes, the scouting side is vital to that which keeps the game beautiful.
No comments:
Post a Comment