do players steal less early in the season?

[bryanbutler]

based on the discussion in this thread, i thought i'd have a look at some data.

unfortunately i can't find MLB data broken down by month or week on the interwebz anywhere.

fortunately, i used to do all the stats-keeping for my main league with all of my own programs, so have weekly raw stats (from the old TQStats) from the years 1995 through 2006.

the two linked plots i think show pretty definitively that there is no change at all in the rate of steals, at any time during the season. the idea that players run more as the weather warms up is an urban myth.

first plot is the ratio of SB to PA (in %) as a function of time in the season - each season is a different color (though some get reused you get the picture). second plot is that same ratio, but divided by the average for that year (to take out year-to-year trends of more or less overall SB).

[Controller Jacobs]

You can get MLB splits in B-R, which includes per-month (link: http://www.baseball-reference.com/le...=MLB&year=2012)
Here's 2012 SB/G
April/March: 418/676, 0.62
May: 578/850, 0.68
June: 541/808, 0.67
July: 545/766, 0.71
August: 531/850, 0.62
Sept/Oct: 616/910, 0.68

overall: 3229/4860, 0.66

A jump after April/May seems to support the hypothesis, but then there's the odd drop in August.

Looking at age splits, young player were much more likely to get a SB, so maybe there is some correlation with month due to delayed early call-ups (such as with Trout) and the September call-ups.
Age 25-: 955/1214
Age 26-30: 1534/2231
Age 31-35: 611/1107
Age 36+: 129/306

[TranaGreg]

You can get MLB splits in B-R, which includes per-month (link: http://www.baseball-reference.com/le...=MLB&year=2012)
Here's 2012 SB/G
April/March: 418/676, 0.62
May: 578/850, 0.68
June: 541/808, 0.67
July: 545/766, 0.71
August: 531/850, 0.62
Sept/Oct: 616/910, 0.68

overall: 3229/4860, 0.66

A jump after April/May seems to support the hypothesis, but then there's the odd drop in August.

Looking at age splits, young player were much more likely to get a SB, so maybe there is some correlation with month due to delayed early call-ups (such as with Trout) and the September call-ups.
Age 25-: 955/1214
Age 26-30: 1534/2231
Age 31-35: 611/1107
Age 36+: 129/306

good analysis. though I'm not sure the differences from month to month are substantial enough - I still think panic is in order.

[bryanbutler]

i don't see a jump at all in those numbers. if i were home i'd do the statistics, but i'm in the car waiting to pick up the kids.

ETA: thanks for that B-R link. i found the splits page but couldn't figure out how to get the monthly ones.

[Todd Zola]

Season AB PA HR SB CS success HR/SB PA/HR PA/SB
2010 year 165353 185553 4613 2959 1129 72.38% 1.6 40.2 62.7
2011 year 165705 185245 4552 3279 1261 72.22% 1.4 40.7 56.5
2012 year 165251 184179 4934 3229 1136 73.97% 1.5 37.3 57.0
2010 April 23428 26606 648 462 163 73.92% 1.4 41.1 57.6
2011 April 27060 30384 724 544 208 72.34% 1.3 42.0 55.9
2012 April 22891 25577 638 418 181 69.78% 1.5 40.1 61.2
2013 April 12663 14166 373 187 68 73.33% 2.0 38.0 75.8

[Todd Zola]

Opportunities are 1B+BB+HBP -- it's crude but without breaking into the BIS database, it will have to do.

% Opportunity per PA % run per opportunity
2010 year 24.7% 8.9%
2011 year 24.3% 10.1%
2012 year 24.0% 9.9%
2010 April 25.0% 9.4%
2011 April 24.3% 10.2%
2012 April 23.8% 9.8%
2013 April 24.0% 7.5%

[bryanbutler]

April/March: 418/676, 0.62
May: 578/850, 0.68
June: 541/808, 0.67
July: 545/766, 0.71
August: 531/850, 0.62
Sept/Oct: 616/910, 0.68

average = 0.663
standard deviation = .036

there is only the smallest statistical difference between 0.62 and 0.68 (one is slightly more than 1-sigma from the mean, one less than half). so there's no statistically believable "jump" after april/march (which is what i assume you meant instead of april/may).

ETA: and todd's data backs this up. there is just nothing to show that players steal more when it is warmer (later in the season).