Speed, Position, Walk Rate, and Skin Color on Career Arcs.

Who is Scoresheetwiz?
Archived Commentary
Most Useful Sites
Archived Analysis
Quick Links (revised)

to be studied further . . .

Someone asked in the Scoresheet baseball forum mcscoresheet whether they should keep Chipper Jones or Carl Crawford. Jones is coming off a couple of monster seasons of a .968 and 1.005 OB + Slg. Crawford merely hit .830 last year. Brian Fawcett responded, “Isn’t this really a trick question to discover whether the person answering is a Roto player?” alluding to Crawford’s 58 stolen bases which would indeed be valuable to a Rotisserie team, but not coming close to making up the 150 point O+S difference in Scoresheet – or even real baseball. Yet, Jones is nearly 35 and has only been in shape to play about 110 games each of those two incredible seasons – and not with impressive defense either. Surely, a serious decline in hitting is around the corner. Crawford, who has played over 150 games in each of his Major League seasons, is just 25 - still room for more significant improvement. Noting Crawford’s incredible speed and the adage that speedy players develop power later and last longer than slower players, I thought the choice between them isn’t so easy.


Do speedy players eventually outperform their slower colleagues of the same age and overall ability?


Do they peak later or progress further? That is part of a premise I recall reading from Bill James a couple decades ago. A top notch producer of baseball player projections Ken Warren doesn’t buy this theory. He sees speed as a skill unrelated to the development of other skills such as power, contact rate, and strike zone command. Speed, in fact, is the skill which peaks at the earliest of these major hitting skills, therefore, if a player compared to others of the same age and overall ability derived more of his value from speed, he would have the less progress after the peak age of speed - 24. Of course, we can all name speedy no hit wonders who faded from the major leagues before they reached 30.


There is a logical leap here, though, I’m not prepared to make. There are also many examples of very speedy players who had their best years well into their 30s. As Mitchel Lichtman co-author of The Book – Playing the Percentages in Baseball retorted to Ken in the this article "Projections Roundtable" from the Hardball Times in 2006:


Just because speed peaks early  . . . does not mean that players with speed . . . will “fizzle out” early. For whatever reason, players with speed . . . may retain all of their skills better as they age than players who are not fast.


If having speed has no effect on the other skills, then Ken’s progress structure would make sense. His hitting projections have been shown to be among the most accurate, so, perhaps, he’s right. However, perhaps, they could be better.


My recollection of the rationale behind James and others who say speed does have a relation to the other skills is that as we lose speed, we are also losing quickness. Hitters get some of their power and their ability to make contact with their quickness. If a player is extremely quick to begin with, then when they begin to lose it, they still have enough left to hit a major league pitch, while when the slower guys lose it, they cannot catch up to a major league pitch.


When I finally found the essay where James did his study on fast players vs. slow players I was reminded the essay was more directly about players with young players skills – speed and high batting average vs. players with old player skills – power and plate control. I seem to have taken James' rationale a bit further than he had in the sense that I found nothing about relating running speed to bat speed quickness. I’m not sure if I just haven’t found that theory, yet, or it was an explanation for the longer career arc of speedier players that I reasoned in my own head.


James takes it further in that he includes high batting averages vs. walks and power as part of what to look for in a player with younger skills which will take him further than a player with older skills.


First, in 1986, James even includes position. From his 1986 Abstract page 88 under the Minnesota Twins:


[Latest projection system] Brock6 tends to be more conservative than Brock2, particularly with respect to certain types of players - power hitters, slower players playing the positions at the left end of the defensive spectrum (1b, lf, dh). Brock2 clearly tended to over-project the careers of power hitters. One study that I did focused on young players with old players' skills - that is, young players who were slow but hit for power and had good strike-zone judgment. I identified 12 such players from the years 1960-1972 and ran career projections for them. Eleven of the twelve fell short of their projections. 



It was in the 1987 Bill James Abstract Chapter III “Evaluating a Rookie”. In the sub-section headed "Young Player's Skills" and "Old Player's Skills" (pg.66) Bill James:


All players tend to follow this pattern to one degree or another, and all players are finished when they reach the point at which the gains in power and strike zone judgment can no longer offset the losses in speed and batting average. It makes sense, then, that the further along in this progression the player is (regardless of age) the closer he is to the end of his career (or conversely, the earlier he is in the progression, the longer he can be expected to play).


But a lot of things that make sense nonetheless are not true. . .


James then goes on to describe how he tested this theory. I don't see any flaw in it depending on how he came up with his 36 matching pairs of very similar 22 and 23 year old rookies each of all-around equal quality and similar positions expect one had young player skills (run scoring, high BA, SB, and 3B), the other had old player skills (RBI, HR, BB producers). The result was "the players with young player's skill did play 33% more games in their career (48,583 to 36,530). The player with young player's skill played more major league games than the comparison player in 23 of the 36 cases."


Meanwhile, I found another similar study called The Value of Speed by Nate Silver of Baseball Prospectus. Silver’s 2003 study takes all of the post war hitters who had Isolated Power of 75% [percentile?] or better through the age of 24 with at least 300 at bats. He then splits this into a fast group with 75% speed or better and a slow group with 25% speed or worse, throwing out all the hitters in-between. Then he does several comparisons and graphs of these two groups. The fast guys seem to break away from the slow guys in their mid 20s, then the slow guys rapidly catch up, before gradually splitting apart again with the faster players having ever more productive careers starting from the age of 29.


My qualm with this study is that the two groups are not the same in overall ability. They may have similar ages and power, but speed is an asset in baseball that is a very good thing both offensively and defensively. You would expect that teams would give more playing time and more chances to speedier ball players than to slow ones who are otherwise the same. Arguably, though, collectively any group of players given 300 major league at bats through age 24 are as valuable as any other group of major leaguers with 300 at bats through age 24. Speed or not, they wouldn’t have been given such a chance if they weren’t worthy of it. Hence, you would expect the average speedy player in this group to have less power than the average slow player of equal overall value. So, if you have the same baseline for power (75%), you are excluding the less valuable of the speedsters from the desired subset of players with equal overall ability.


I would like to see more investigation in this area. It does seem that if speed is the only skill a player has of solid major ability, he will disappear fast. Will he disappear faster than a player who only has power? I don’t know. If the speedy player has ample talent in the other skills, then those skills will likely last longer. I would like to know if they can be expected to develop to some degree further, too. How much longer, how much further, and how much of which skill you need for all of this to happen is still out there for us to figure out.


What about high walks rates?


One decade ago the weekly fantasy baseball must-have was Baseball Weekly (now Sports Weekly). They made their own prospect lists with in depth articles. One mantra they often repeated was that players with high walk rates – or, at least, high BB:K rates are more apt to improve than those who didn’t. This seems to be generally agreed upon by baseball experts, now, but I haven’t seen any hard evidence of this. I made a two part study on “Young Stars and their Improvements” and a quick check finds no such correlation among Major League 22 year olds. You can look at the data yourself, but the articles need to be downloaded from the Scoresheetwiz Analysis page. If anyone knows of a more serious study on this, please, share it on scoresheet-talk or mcscoresheet.


Bill James did study the issue and reported in the same 1987 Abstract article. He investigated the issue with similar methodology as his young skills vs. old skills investigation. He compared 47 sets of matched rookies except for their BB and SO. Actually, the players with the poorer K/W had overall slightly better careers than their matches. However, the difference was too slight to be significant. His conclusion confirms my skepticism:


Strikeout to walk ratio [in rookie hitters] is apparently not an indicator of potential growth or development.


. . . Eddie Epstein argues that what is significant about strikeout-to-walk ratios is how the change over a period of years; if a player, after being in the league a few years, not only fails to improve in this area but actually goes backward, that may be a signal that something is amiss in his development as a hitter. That may well be, but that’s another study. Looking just at a rookie, strikeouts and walks give no significant information about future development.


Other growth indicators in rookies


James does many such comparisons in this ’87 Abstract study finding, for example:


Age of rookie


The largest growth indicator of a rookie is the age at which he gets his start. This is pretty much taken for granted now by the baseball intelligencia. I don’t know when the rotisserie players took notice, but James concluded in ‘87:


If the performance of a 21-year-old and a 23-year-old rookie is just the same, the 21-year-old has a dramatically better chance to become a major star.


The rookie’s team


As a group, the rookies who played on good teams did go on to have distinctly, if not dramatically, better careers. The 112 rookies who played for poor teams played 105,179 major league games. The rookies who played for good teams played 133,271 – 27% more. The rookies who played on good teams eventually hit 33% more home runs (11,842-8,886), had 33% more hits and did almost everything else 33% more times.


In this study there could be a serious problem with quality leakage. The players who played on good teams could very well have been better players or better prospects in ways too subtle for the similarity method to pick up. Nonetheless, the way that the method looks at players is not that different from the way that you and I look at rookie, and it seems like a worthwhile thing to keep in mind when considering who to daft in your rotisserie league or your Stat-O-Matic league.


What could this leakage be? Poorer teams rush their prospects more out of need no doubt. Do a higher percentage of the rookies on weak teams achieve a level of competent statistics by luck, just because a higher percentage are rushed? Or is there a real effect here, such as better teams having better coaches and/or better peer leadership?

The rookie’s position


Bill James on catchers:


As a generalization, a good-hitting player loses 25-30% of his career and 40-50% of his career production if he is a catcher. A great hitting catcher loses more than that.


Right now, before the 2007 season has started, there has been much debate over what order should be the three perceived most outstanding players for an A.L. start-up Scoresheet league: Mauer, Sizemore, and Santana. I wonder how many minds would change after reviewing that study.


Bill James on secondbasemen:


On the whole, the pattern of evidence would indicate that second basemen lose almost as much of their offensive potential as do catchers.


Bill James on thirdbasemen:


. . . the biggest surprise was the poor development of third basemen as hitters. In some studies, third basemen performed almost as poorly relative to their rookie seasons as did second basemen and catchers . . .


James on other positions:


A rookie center fielder who can hit will play about 15-20% longer in the major leagues than a right or left fielder of the same offensive ability, and nearly 40% longer than a first baseman.


The rookie’s ancient ancestry


Here is a politically incorrect shocker. I don't understand why the fast player vs. slow player or young skills vs. old skills thing stuck in my head while this didn't. It must be because I didn't start playing in a continuing fantasy league for another 4 years and the media never dares to bring this up. However:


The difference between the development of a black player and the development of a white player was greater than the development as hitters of an outfielder and a catcher.


The difference between the career expectation of a black player and a white player was greater than the difference between that of a slow player and a fast player, or between a player who had young player's skills and one who had old player's skills.


The difference between the development of a black player and a white player was greater than the difference between a player on a poor team and a comparable player on a good team.


The only greater difference that I could find would be between a player who came up at 20 and one who came up three of four years later.


James also made a study on Latin players concluding they seemed to do better than whites, but not nearly as well as blacks. (Please, excuse the 80s terminology). However, James felt he didn't have enough Latin players to make a conclusive call on this. (How times have changed!) 


The final section of the Evaluating Rookies chapter is speculation regarding the reason blacks (afro-americans) succeed more than whites (euro-americans?) beginning at equal points. He delves into the possibility of due to few opportunities outside of baseball, the blacks might have greater determination to succeed. What his research uncovered is that  blacks retained their speed longer than whites. He draws a dramatic comparison between Gus Bell (father of Buddy, grandfather of Dave) and Henry Aaron.


As rookie, Bell hit .282 and slugged .443; Aaron hit .280 and slugged .447. Bell at that point seemed to have better speed; he hit 12 triples to Aaron’s 6 and stole four bases to Aaron’s two. The next year Bell led the league in triples. When Cincinnati acquired Bell they moved him to center field, and for four years he was a terrific player, hitting .300 with 25-30 home runs and 100 RBI and doing a good job in center field. He played in four All-Star games, competing for the job with people like Willie Mays, Duke Snider, and Richie Ashburn.


Yet if you check in on them at 29, Henry Aaron is stealing 31 bases in 36 attempts. Bell, fight for playing time, hits two triples and steals two bases. Five years later, and Aaron is still stealing bases, 28 in 33 tires. Bell is out of the game.


However, it was brought to my attention from a scoresheet-talk regular Garth Hewitt who apparently has professional expertise on genetically transferred diseases, that James’ study on black rookies doesn’t hold up. The most damning reason is that most of James’ pairings came from the 50s and much of the 60s. At that time “black” kids did not have the same coaching and training opportunities “white” kids did. There was prejudice that they didn’t have what it took to be successful. Upon reflection, unless they were super talented, it seemed many weren’t considered as marketable. So, of course, when a black kid and a white kid of the same age first reach the majors and have the same success initially, the white kid is already closer to his maximum ability, while the black kid has much more to learn and more talent to develop and will likely far surpass the comparable rookie white kid in the long run.

John Carter