Few things drive me more insane than watching a tight game get turned over to a bullpen that issues walks by the dozen (I'm looking at you Kevin Gregg). So because I've needed something to constitute a study break recently, I started compiling walk rates for 2010, 2011, and 2012 for the Blue Jays vs League Averages.

This is borne out of seeing the bullpen essentially rebuilt with more controllable, higher velocity, higher strikeout rate pitchers. I like the idea; intuitively, high-velocity, high-strikeout pitchers make it harder on batters to make quality contact, because presumably their 'stuff' is much better. The problem with electric stuff is often an inability to harness it effectively, leaving these pitchers open to a loss of control, and therefore more walks.

So while I was thrilled with the idea of electric relievers in the bullpen, I was also very concerned it could spiral into a year watching a bullpen implode by issuing walk after walk.

Relief stats can be deceiving, particularly when compiling numbers from EVERY pitcher who threw an inning of relief for a team. That would involve including numbers from somebody like Jeff Mathis as part of the Blue Jays bullpen, and I just can't justify that as a reflection of the team. With that in mind, I set a limit (an arbitrary limit) of 10 relief innings pitched as a minimum for inclusion into the sample. Data was collected from 2010, 2011, and 2012 bullpens for the Blue Jays, as well as separate samples for any pitchers in the MLB throwing 10 relief innings in 2010, 2011, and 2012.

*10 Relief IP Minimum

Next I flipped through Toronto's current 40-man roster, and created a custom table of prospective candidates for the 2013 bullpen, to be used as a projection. Again, the sample included any pitchers who threw at least 10 innings of relief in 2010, 2011, or 2012. The list included Casey Janssen, Darren Oliver, Brad Lincoln, Chad Jenkins, Brett Cecil, Aaron Loup, Steve Delabar, Sergio Santos, Esmil Rogers, Jeremy Jeffress, Evan Crawford*, and J.A. Happ*. This data was includes any season pitched by these pitchers with at least 10 innings of relief in that year. It is not a projection of the totals for the season.

*Data does not include seasons from Crawford or Happ

The first thing that struck me was how much higher the League Average numbers were than the BB/9 rate of the Blue Jays. Now I remember that 2010 bullpen, and I'm pretty sure that was the year I started losing some of my hair. I wondered if there were a number of lower tiered pitchers who only pitched a few innings beyond 10 innings, and that was the reason the walk rates were so high. Regardless, I wanted a comparison of the Blue Jays to quality relievers in the MLB, so I upped the minimum inclusion value to 30 relief IP. I left Toronto's values as they were because otherwise I risked creating a very small pool of data. Results were as so:

*30 Relief IP Minimum for League Values. Blue Jays remain at 10 Relief IP.

My original plan was to stop here, create a chart to give visual representation to the data, and call it a post. After all, I have exams to study for. But the more I looked at these numbers, the more I thought although it reflects the bullpen as a whole fairly well regarding walk rates, it really doesn't do anything to reflect on the recent shift towards power arms.

So I took one final sample, reproduced the projected bullpen sample and generated a table for relief pitchers throughout the MLB with the same 30 relief inning minimum as before, but included fastball velocity as well. Years included were again 2010, 2011, 2012. To determine what constitutes a ‘power reliever', I took the lowest value in the last three years for fastball velocity of the ‘power arms' in the Toronto's bullpen (of Jeffress, Santos, Rogers, Delabar, and Lincoln), which was Lincoln in 2012 at 93.8mph, and used that as the lower limit. Then I applied that minimum to the MLB table, and determined average walk rates for relievers who threw >/= 93.8mph with at least 30 Relief IP in one of the last three years.

*30 Relief IP Minimum for League Values. Blue Jays remain at 10 Relief IP. Minimum FBv 93.8mph.

Now that is a startling difference in value, as the power arm relievers who are projected to possibly contribute in Toronto's bullpen this year appear to walk batters at an alarmingly high rate. It's a little confusing given the projected BB/9 of the 2013 bullpen is only 3.08, and yet here are the power arms at 4.02. Given the sample isn't that large, I think these are the numbers throwing the data off:

In sum, I'd call this a quasi-reliable sample of the bullpen for next year regarding walk rates, and for the most part it makes me feel good their prospects, particularly if Oliver returns (his walk rate over the last 3 years is a measly 2.19 BB/9). It certainly is difficult to shake the feeling that the newer, harder throwing relievers may have some serious issues throwing strikes - but given the clear difference between hard throwing MLB relievers vs baseball relievers as a whole, that just seems to be part of the equation for anybody throwing 93.8mph and above in general.

(Having said that I'll draw your attention to the 3.59 BB/9 rate for pitchers 93.8mph and above and compare that with the league average walk rates for MLB relievers throwing at least 30 innings. Harder throwers walk more batters. I'd have liked to include a line graph here to illustrate the difference, but I can't figure out how.)

Concerns

It's obviously not a perfect sample. Setting the minimum relief IP at 10 innings and 30 innings is hardly a perfect measure of quality. Both are still reasonably small enough to create some SSS issues - both the hope was that with a bullpen sample of 10 prospective players over 3 years there would be enough data to clear (at least a little) those issues. Particularly so since values were calculated by TOTAL BB divided by TOTAL IP in the entire sample, and not simply BB/9 divided by the number of players (why I would even consider that, I have no idea).

Likewise, setting the ‘power reliever' level at 93.8mph was just to give a relative comparison to the arms the Blue Jays have available. Hardly statistically sturdy.

Finally, some players ended up excluded from the data sets. For example, JA Happ isn't included in the prospective bullpen. Since I was using relief IP as the measure, and excluding any innings recorded from starts, he didn't meet the 10 relief inning minimum in any year since 2009, and the sample only extends 3 years back, not 4. Same thing goes for Evan Crawford, who only threw 8 IP last year.

I think this gives me an idea of what to expect from the bullpen next year as far as walking batters is concerned - hopefully it does for everyone else too. I'm very clearly not a statistician, but the trends here suggest as a whole the bullpen should fare very well in not excessively walking batters compared to the rest of the league.