[ Home ] [ Library ] [ Bookstore ] [ Contact ] [ Search ]
April 16th, 1997
John Hunt's Fantasy Beat column in Baseball Weekly recently ran an article claiming to disprove the widely held view in the sabermetric community that hitters, on average, reach their offensive peak around age 27. The author of the study claimed that there were no studies he could find that actually supported the "age 27" theory, and suggested that it had been passed around long enough that it simply came to be accepted without proof.
As it so happens, Bill James (and others) have looked at this systematically, which led them to the age 27 hypothesis. I myself have conducted two different investigations into hitters' peak ages, which I describe below:
The first study looked at all players who played 10 or more years in the majors (no AB minimum required). For each, I looked at the maximum number of adjusted batting runs they had in any season. Adjusted Batting Runs are defined by Thorn & Palmer in _Total Baseball_ and represent the number of runs contributed by a player above the league average, adjusted for park effects, using the Linear Weights methodology. The age at which the player attained his highest ABR total is defined as his peak year. If two or more years tie, each is counted as a peak year. The breakdown of the ages at which 10+ year players peak in ABR is as follows:
AGE: age of peak season
PEAK: number of players who peaked at this age
%Peak: percentage of players who peaked at this age
Cumul%: cumulative percentage, or the percentage of players
who have peaked at this age or earlier.
AGE PEAK %Peak Cumul% 18 1 0.1% 0.06% 19 5 0.3% 0.38% 20 15 0.9% 1.33% 21 29 1.8% 3.17% 22 55 3.5% 6.65% 23 87 5.5% 12.16% 24 90 5.7% 17.86% 25 148 9.4% 27.23% 26 160 10.1% 37.37% 27 163 10.3% 47.69% 28 155 9.8% 57.50% 29 138 8.7% 66.24% 30 141 8.9% 75.17% 31 107 6.8% 81.95% 32 91 5.8% 87.71% 33 68 4.3% 92.02% 34 46 2.9% 94.93% 35 25 1.6% 96.52% 36 22 1.4% 97.91% 37 13 0.8% 98.73% 38 9 0.6% 99.30% 39 6 0.4% 99.68% 40 1 0.1% 99.75% 41 1 0.1% 99.81% 42 2 0.1% 99.94% 43 1 0.1% 100.00%
Conclusion: For players who will have significantly long careers, the most common age to peak is age 27, with 25-28 being the most likely range for a peak year. By age 31, over three-quarters of all such players will have seen their best year come and go.
A second study used a different approach. I looked at all batters who had 400 or more plate appearances (AB+BB) in two consecutive seasons, and measured the change in production from year to year. For example, I looked at all players with 400+ PA's at age 25, and also had had 400+ PA's at age 26, to see what their overall change in AVG, OBP, SLG, etc. was. The change was measured as a ratio -- e.g. AVG(26)/(AVG(25), and multiplied by 1000 to eliminate the decimal. 1000 is average, 950 represents a 5% decline, and 1100 represents a 10% increase.
I measured several rates of production, including AVG, OBP, SLG, Home Run to Double ratio, and Home Run to Hit ratio (the latter two were part of an investigation into how the shape of a player's power output changes with age). I also looked at Runs Created per Out (which is equivalent to RC/27, which is found in Nelson Lu's reports), which can serve as a metric for overall offensive production for our purposes.
AGE: The age we are comparing to the previous year's production
(i.e. the age 27 line compares how age 27 players did versus age 26)
COUNT: number of players in the sample (400+ PA at AGE & AGE-1).
30 player minimum sample to be included in the table below
AB1: # of at bats at age N-1
AB2: # of at bats at age N
AVG, OBP, SLG: ratio of change from N-1 to N in this stat HR_2B, HR_H: same as above, but for HR/2B ratio and HR_H ratio
RC_OUT: ratio of change in Runs Created per Out from age N-1 to N
AGE COUNT AB1 AB2 AVG OBP SLG HR_2B HR_H RC_OUT 21 54 27411 28600 1040 1042 1051 903 1109 1112 22 170 86271 89080 1012 1018 1035 891 1136 1059 23 324 166333 169966 1008 1017 1010 997 1029 1031 24 501 257619 262265 1000 1010 1005 997 1031 1015 25 678 349074 352995 1006 1010 1011 976 1040 1024 26 835 428589 435722 995 1002 999 975 1029 999 27 929 483867 478894 991 999 990 986 1009 985 28 923 476541 469018 993 1001 989 1043 982 987 29 867 445564 441217 993 995 991 997 1001 983 30 784 402334 401527 989 995 985 990 1001 976 31 652 338692 332899 987 993 982 1056 965 970 32 536 278558 272739 989 993 984 961 1007 973 33 432 223242 217934 986 989 986 1025 993 970 34 324 167554 161813 980 986 967 1048 937 946 35 228 116587 114137 981 985 969 1070 943 947 36 167 86422 82234 984 990 972 1058 944 956 37 97 49521 47469 974 974 960 1007 972 925 38 64 32125 30436 974 983 964 1058 940 938 39 38 18543 18100 960 958 963 1014 1033 907
As you can see, there's a steady increase in production until about age 26-27, followed by a accelerating decrease. In fact, this suggests that players peak slightly earlier than age 27, but still close enough to support the overall conclusion that players do *not* peak, on the average, anywhere near age 30 or more.
[ Home ] [ Library ] [ Bookstore ] [ Contact ] [ Search ]
Last Updated: Contact webmaster@stathead.com for corrections or problems
Copyright 1997-2001 by Keith Woolner. All included authors retain the copyrights to their original works.