August 13, 2003
Aim For The Head
Pete Rose and the Hall of FamePete Rose is, in fact, reinstated after the season, as previously reported, he becomes eligible to be placed on the Hall of Fame ballot. For many fans, his on-the-field qualifications are a foregone conclusion. As baseball's all-time hits leader, 17-time All Star, the 1973 MVP, and key member of the Big Red Machine, it's hard to deny that Rose has some impressive credentials. And indeed, baseball fans voted him onto the All-Century Team as one of the finest players of the 20th century.
However, there's been a reassessment of Rose's value as a player over the past 15 years, as sabermetrics has advanced our understanding of how offenses work, and how teams win. As the importance of On-Base Percentage has been recognized, and measures such as OPS (On-Base Percentage Plus Slugging average) have become popular, Rose has become a poster boy for the overrated star--one whose game consisted of hitting a lot of singles, and posting a high but empty batting average.
Some have gone as far as to say that Rose doesn't deserve to be in the Hall of Fame on the merits of his playing career, even excluding any gambling controversy. But is this a revisionist history by the statheads, or an honest, updated assessment of a former star?
While there are other ways of looking at his qualifications (including Bill James's "Keltner List"), long-time readers of Baseball Prospectus know that we have developed a variety of statistical tools to help evaluate player performance. And in particular, I developed and use a system called VORP (Value Over Replacement Player), which includes a rate statistic called MLVr (Marginal Lineup Value rate), to estimate the value of a player's on-field performance. For newer readers, I'll give a brief introduction to these statistics below.
MLVr is the number of runs per game a player would add to a league-average lineup, if he was replacing a league average player. For example, as of 8/11, Albert Pujols had a .619 MLVr, which means that if an average team scored 4.5 runs per game, replacing an average player with Pujols would improve their run scoring to about 5.12 runs per game, or about 100 runs over a full 162 games. A zero MLVr does not mean a player has no value--rather it means that he is exactly average. Negative MLVr is not only possible, but quite common. Jeff Cirillo currently has a -.281 MLVr, dragging our 4.5 R/G team down to about 4.22 R/G, or costing his team about 45 runs per 162 games *compared to an average hitter*.
MLVr is based on a formula for estimating how many runs a team should score, based on its total number of hits, double, triples, home runs, walks, steals, strikeouts, caught stealing, and other batting outs. By incorporating walks, extra-base hits, and outs, the full value of on-base percentage and slugging ability is captured. It also has adjustments for the league offense level (the 1960s were a low run-scoring era, whereas the late 1990s were a high run-scoring environment) and park factors (pitchers in Colorado are better than their raw stats would indicate, and the hitters worse, because the park is easier to hit in, while the reverse is true for Pac Bell Park).
VORP extends MLVr in two ways. First, it is a cumulative statistic, and thus incorporates the value of playing time, whereas MLVr is a rate statistic (as in the difference between batting average, a rate stat, and RBI, a cumulative stat). Second, instead of comparing all players to league average, VORP compares them to a typical backup/fringe major leaguer (or "replacement level player") who plays the same position. This way, shortstops, who play a tougher defensive position, are not compared directly to first basemen, who play a relatively easy position. Also, setting the bar to replacement level instead of average recognizes the value of average players, since all teams can easily find replacement-level players, but average players are more scarce (and costly).
Let's look at Pete Rose through the lenses of MLVr and VORP. We'll examine: (a) whether Rose is a Hall of Famer, based on his on-field production, and (b) where he ranks among the all-time greats.
We'll start with MLVr, or the rate of offensive production (in the same way that batting average is the rate of getting hits per time at bat). Rose had a lifetime MLVr of .1568, well above average, and worth about 25 runs per 162 games to a typical team. However, that is not an exceptional career total. Players with higher career MLVr include Danny Tartabull, Bill Madlock, Bob Watson, and Joe Torre. In fact, some of the players with comparable career MLVr are names like Rusty Greer, Cesar Cedeno, and Lyman Bostock--fine players in their own right, but not Hall of Fame quality. Rose's career MLVr ranks 315th on the all-time list (for players with at least 2000 plate appearances).
Rose has one thing that his "comparables" lack, however. Astonishing longevity and durability. Appearing in over 3,500 games in your career will allow you to rack up a lot of value. So, for our next step, we turn to VORP, which is designed to recognize the value of playing time (which Rose had in abundance), as well as adjusting for position played. For the years Rose played second base, he is compared to other second basemen, seasons at third are compared to third basemen, and so on.
Rose's career VORP was 713--he produced more than 700 runs for his teams beyond what a typical backup could do. That's a very high total, good for #50 on the all-time list. And this places him in the company of several Hall of Famers. He sandwiched in between Johnny Mize and Willie Keeler, with Harry Heilmann and Al Kaline nearby. In fact, there are very few players above him that are not Hall of Famers, or active/recently retired players with solid cases such as Barry Bonds, Rickey Henderson, and Cal Ripken. The highest rated eligible non-HOFer is Bill Dahlen at #60 with a 670 VORP.
At this point, the naysayers could argue that longevity was Rose's only qualification, and that above average but never really stellar production for an absurdly long time shouldn't be a Hall credential. Some made similar arguments against Don Sutton and his membership in the 300-win club, though Sutton was voted into the Hall.
To see whether Pete, at the pinnacle of his career, was a Hall-worthy player, we can look at his peak value. There are many ways of defining peak value, but for our purposes, we'll use the total VORP accumulated during a player's best five seasons
Rose's best five seasons were 1965, 1969, 1975-76, and 1979--ironically his MVP season of 1973 just missed the cut-off, and did not rank among his five best, mostly because he was compared to better-hitting left fielders instead of 2B/3B as he was in several other seasons.
In Rose's five peak seasons, he racked up 275 VORP, or about 55 VORP/season. That's a very good peak; a 55 VORP season is about what Jose Vidro or Mike Piazza posted in 2002. It is not, however, a historically high peak. Cal Ripken had a five-year peak of 361 VORP. Roberto Alomar, 378 VORP. Wade Boggs, 410 VORP. And that's without touching the rarified air of Babe Ruth, Ted Williams, Lou Gehrig, Rogers Hornsby, Ty Cobb, and Barry Bonds, whose five-year peaks exceeded 500 VORP (averaging over 100 VORP per year).
Rose never led the league in VORP, but he did place in the top 10 seven times, with a highest ranking of second in 1976. And the only time between 1963 and 1981 where he didn't place among the top 50 position players in the NL was 1980, at the age of 39. He rebounded to rank #12 in the NL the following season.
Rose's five-year peak VORP of 275 ranks #159 all-time. Players with similarly valued peaks are: Dwight Evans, Tim Salmon, Julio Franco, Babe Herman, Rudy York, and Eric Davis. There are Hall of Famers with lower peaks, including Roberto Clemente, Roy Campanella, Al Kaline, Tony Lazzeri, Pee Wee Reese and Rose's Big Red Machine teammate Tony Perez. Clearly, you can earn your way into the Hall of Fame without a stratospheric peak, if you have longevity (such as Kaline) or are associated with some of the best teams in history (Lazzeri). Rose, of course, has both.
Is Rose then an "inner circle" Hall of Famer? No, not based on his on-field production. He did not hit for enough power, walk enough, nor steal bases well enough to be compared to Mickey Mantle, Eddie Collins, Stan Musial, or the man whose hit record he broke, Ty Cobb. Nor is he really of the solid, but less-than-mythic Hall members like Rod Carew, Carl Yastrzemski, Mike Schmidt, or Mel Ott, who combined longevity with higher rates of production.
But it seems pretty clear that Rose has a Hall-of-Fame quality career, contrary to the view of his most ardent critics of his on-field performance. While never quite dominating his league, nor posting an exceptional career rate of production, Rose mixed solidly above-average production with several great seasons and unparalleled durability and longevity. To put things in perspective, just half of Rose's plate appearance total would rank among the 250 longest careers in MLB history, and is more PA than Fred Lynn, Cecil Cooper, Norm Cash, or Carney Lansford had in their careers.
In fact, Rose's career was so long, that it becomes useful to view his overall value as the sum of two other players with significant careers. Looking at his career MLVr and VORP, Pete Rose can be roughly compared to Tim Raines' career (without the speed) plus Lou Boudreau's career (without 2/3rd of his managing career, including the World Series title). Boudreau is in the Hall, and Raines has a good case to get in.
Factors beyond what he did on the field may ultimately determine whether Rose ends up in the Hall, but his production as a player should not hold him back, even to the sabermetrically-inclined way of thinking.