Impact of League Change on Stolen Bases Revised

After my last post, I had a quite a few people comment on some other very possible variables that could potentially explain why there was a difference in stolen base totals after an individual player changed leagues.  Most people suggested other variables that might account for any differences in league changes.  Some of these variables included age of player, OBP, and what league each player was switching to.  Obviously in a simple paired sample t-test, there are absolutely going to be other possible (and probable) explanations for the difference in stolen base totals when players change leagues.  In my last study we were just simply testing to see if there was a difference, and we didn’t really go into why.

In this post, I’ve made some changes to my study to improve its accuracy and I’ve also extended my study to include data from the second year after the league change as well.  This might help to address some of the grey areas that I had in my first study.  Here’s the steps I took to perform my study this time around:

• Find players in the last 25 years that have switched leagues and spent at least two consecutive years in a row in the second league
• Each player had to have at least 20 SB in their year before the league switch to be considered for this study
• Each player had to have at least 300 AB in all three seasons observed to be considered eligible for the study
• Calculate the statistic AB/SB, which gives us the ratio of AB per Stolen Base for each year
• Calculate descriptive statistics for each year
• Perform a paired sample t-test to see if there is any significant difference between years

In my last study, I simply used percent change of stolen bases from year to year without taking into account the amount of playing time each player received.  The AB/SB statistic should fix that problem.  The data I collected is too large and messy to provide in this case, so I guess you guys will just have to trust me that I’m not making up random numbers.  I had a sample size of 30 in case anyone is interested.  Here are some of the descriptive statistics of our results for AB/SB.  The year indicated is the first year after changing leagues (so the second year for each player in the study).

As we eyeball our results, we see that the data in this case tells us a slightly different story than the first study I did.  We actually see the AB/SB ratio improve from Year 1 to Year 2 (a lower AB/SB is better).  We see that the medians between Year 1 and Year 2 are much closer than the means.  This means there were likely a couple outliers that influenced the mean.  As we turn to Year 3, we see that this yields us our worst results.  This is most likely due to a natural decline in speed and stolen base ability as players age.  Remember, the majority of these players are likely in their late 20s or early 30s.  Younger players are less likely to immediately change teams after they have already established themselves as quality players.

We obtain a high p-value of 0.19 when comparing the first and second years.  We also obtain a high p-value of 0.28 when comparing the second and third years.  This means that we cannot conclude that there is a significant difference between our Year 1 & Year 2 results and also our Year 2 & Year 3 results.  However we do obtain a significant p-value of .01 when comparing our Year 1 and Year 3 results.  In this case, we do have enough evidence to conclude a difference in AB/SB between Year 1 and Year 3.  Again, this is likely due to the natural decline of speed.

So it looks like this revised study of the impact of league change on stolen base totals gives us some different results than before.  It appears that the league change should NOT actually have a significant impact on stolen base total.  Most players that were baseburners in their younger days usually do see somewhat of a decline starting in their late 20s or early 30s.  Any decline in stolen base total should be attributed to this much more than it should be attributed to the league change.

What does this mean for the Indians?  Well it means some bad news unfortunately in terms of Michael Bourn.  Bourn will turn 31 in December of this month, so he’s at the age where a lot of speedy players have already begun their decline or start to decline in terms of stolen bases.  The drop off in SB totals we saw last year from Bourn can likely be attributed to this decline.  This doesn’t mean that Bourn can’t put up good stolen base totals for the next couple of seasons.  I don’t think the 30 SB plateau is out of question.  However, for a guy that has stolen 40+ bases in his previous 5 years before 2013, that doesn’t sound as impressive. I could be wrong, but I would guess 40+ SB days are likely long gone for Bourn, and that the Indians likely won’t ever be seeing the player they were expecting to get in terms of stolen bases.