This may be an important question, considering the talk of steroids and other performance enhancing drugs. Bonds certainly has done something unusual, perform at a very high, even peak level, in his late thirties.
I wrote an article on this for my own site last year (the link is at the end of this article). Bonds not only did better at an older age than he did at a younger age, but he ranks very high compared to how other players hit at an old age, often first in things like relative slugging percentage and relative HR%. Here I will look at things from a slightly different angle. First, I look at what percent of his career value came at an "old" age and compare that to other players. Then I compare the performance trajectory of his career to other players who played full-time (or close to it) over the same ages that Bonds did.
For the first case, the stat I used was RCAA or runs created above average. It comes from the Lee Sinins Complete Baseball Encyclopedia. Here is how he defines it: "It's the difference between a player's RC total and the total for an average player who used the same amount of his team's outs. A negative RCAA indicates a below average player in this category." It is also park adjusted. Runs created was initially a Bill James creation which is an estimate of how many runs a player creates for his team based on his stats.
To compare "old" performance to career performance, I found all the players with 4000 or more plate appearances (PAs) up to age 35 and 1000 or more PAs from 36-39. I found their RCAA for their career and their old age. Some players had negative RCAA in one or both cases, so they were dropped (I try to adjust for this below with a "replacement level" version). Then I ranked these players by their "old" age RCAA as a percentage of their career RCAA (up to age 39). The top 15 are listed below in Table1.
Bonds ranks high at 6th. Although he is not first, no other player is close in "old" RCAA.
A player who had a negative RCAA up to age 35 but went positive after surely improved quite a bit. But percentages don't make any sense with negative numbers. So I used a crude "replacement" level. I assumed that the replacement level hitter would hav -20 RCAA for each full season (I used 700 PAs). So I divided each player's PAs by 700 to get the number of seasons. Then that was multiplied by 20. That total got added to their RCAA to get "replacement" RCAA. Table 2 has the leaders. Bonds ranks very high again.
I could have chosen some other age to be the "old" age. So I checked 35-39 vs. under 35 and 37-40 vs. under 37. Using 35-39 as "old" age, Bonds ranked 8th in RCAA and 10th in replacement RCAA. Using 37-40, he was 6th and 12th. In both of those cases, as in the 36-39 old" case, no other player is even close to Bonds in "old" age RCAA. In the 35-39 case, the next highest RCAA in the top 15 (like the ranks in the tables here, by %) is 266 belonging to Edgar Martinez. In the 37-40 case, the next highest RCAA in the top 15 is Martinez at 177. So Bonds did not get the highest percentage of his value at an old age, but among the leaders, he dwarfs them in RCAA.
My second way of looking at how Bonds has aged involves looking at his performance over time. I used offensive winning percentage (OWP). This is another Bill James stat which says what a team's winning percentage would be if they had 9 identical hitters and gave up an average number of runs. Since I got my data from Lee Sinins Complete Baseball Encyclopedia again, this is park adjusted.
Bonds' OWP from age 36-39 was about 21% higher than prior to that. The graph below shows how his performance changed over time. The trend line is a third order polynomial.
Here is the graph again without the trend line so you can look for the trend without help. I think it's there.
What does a normal trend look like? For this, I found players who, like Bonds, played full-time or close to it from age 21-39 (that is up to 2004 for Bonds). Bonds had 400+ PAs in each season at these ages. I found eleven players who had at least 300 PAs for each of these ages (with the exception that 5 of them each had one season between 200 and 299 but no one went below 232). I found them by looking at all players who had played at least 20 seasons up through 2002. They were
Harold Baines
Hank Aaron
Dwight Evans
Eddie Collins
Al Kaline
Carl Yastrzemski
Tris Speaker
Ty Cobb
Rickey Henderson
Brooks Robinson
Bill Dahlen
I found the weighted average of their OWPs at each age. The graph below shows the typical trend as these guys aged.
As you can see, the trend is for OWP to peak and then fall. This is much different than the Bonds graphs. It appears that around 32-34, Bonds is starting to trend down, but then it shoots up dramatically. This is unusual. None of the 11 players in the group above have a trend like this. In fact, only 3 of them had a higher OWP from 36-39 than before 36. Baines had the biggest improvement, with an OWP 12% higher. The chart below shows his change in OWP as he aged.
I don't see much of a trend. It is also hard to spot a trend for the other two who did better from 36-39, Aaron and Evans (so I did not see a need to put them here).
It appears that Bonds' increased performance level as he aged is, if not unique, highly unusual.
Links to other research on Bonds and aging patterns in general
"Has Anyone Aged as Well As Barry Bonds?"
http://www.geocities.com/cyrilmorong@sbcglobal.net/Bonds.htm
"Smoothing Career Trajectories" by Jim Albert "By the Numbers" August 2002
http://philbirnbaum.com/btn2002-08.pdf
JC Bradbury has these studies posted at his site:
http://www.sabernomics.com/sabernomics/index.php/2004/03/age-and-batting-performance-in-mlb/
http://www.sabernomics.com/sabernomics/index.php/2004/03/age-and-batting-performance-continued
http://www.sabernomics.com/sabernomics/index.php/2004/03/more-on-age-and-batting-performance
http://www.sabernomics.com/sabernomics/index.php/2004/03/age-and-pitching-performance
"ESTIMATED AGE EFFECTS IN BASEBALL" By Ray C. Fair
http://cowles.econ.yale.edu/P/cd/d15a/d1536.pdf
Dr. Fair is a well known and respected economist.