April In The Minnesota Twins Bullpen

Not sure how you are counting here.  Tonkin had 10 games.  

 

In 4 of those he had inherited runs (6 total.)  In all 4 games he let all 6 runs score.  So he did not do his job.

 

In other 2, he was scored upon, one of them a game loser.  So he did not do his job.

 

So for sure in 6 games he did not do what he was called to do.

 

and that becomes 4-6

 

In other game he gave a hit, a walk, and got two outs.  Did he do his job?  Maybe.

 

There is no way he did his job in 7 games though

You're double-counting one game where Tonkin allowed an inherited runner and allowed a run of his own (the 16 inning one).

Also, the other game where Tonkin allowed a run of his own, it was in a 3 inning long relief appearance.  I think 1 run in 3 innings counts as doing his job, particularly under those circumstances.

It appears Seth is dinging Tonkin for his first 3 inherited runner games, but not the most recent one vs Cleveland.  He inherited a runner at second and only one out and let him score, but he otherwise got 5 outs that day without allowing another run (or leaving any inherited runners for others).

Same with Abad.  4/27 against Cleveland.  A game the Twins lost 5-6.  He inherited 2 runs and both of them scored, allowed an additional unearned run and got just one out.  He did not do his job by any means.

 

So why that is a good game for him, I do not know.

Abad inherited second and third with nobody out.  He allowed an RBI groundout, and then a sac fly to score both runners (then Rosario dropped the sac fly, which put another runner on base that Tonkin allowed to score).  Abad had to face two reasonably good, hard to K hitters too in Lindor and Brantley.

Might be a spot where a simple +/- system doesn't work -- maybe this should be, at best, a neutral performance from Abad?

Here are the overall success rates of the relievers in April:
 

Pitcher -- Yes-No (Success Rate)

Casey Fien -- 8-4 (66.7%)

Trevor May -- 7-5 (58.3%)

Fernando Abad -- 11-0 (100.0%)

Ryan Pressly -- 8-3 (72.7%)

Michael Tonkin -- 7-3 (70.0%)

Kevin Jepsen -- 5-5 (50.0%)

Ryan O’Rourke -- 5-2 (71.4%)

Glen Perkins - 0-2 (0.0%)

Taylor Rogers -- 1-0 (100.0%)

Alex Meyer -- 0-1 (0.0%)

So, in the season’s first month, Paul Molitor has gone to his bullpen 76 times. 52 of those times, or 68.4%, have been successful.

 

Seth, since you yourself have questioned the meaning and usefulness of your measurement, may I suggest switching to something more quantitative?  So it is more easily tracked, understood, and compared across teams/years?

For example, I just scanned the Fangraphs game logs for WPA/LI and noted who had positive/negative marks for each appearance, and came up with the following results for April:

Pitcher, Positive-Negative

Fien 8-4

May 7-5

Abad 10-1

Pressly 8-3

Tonkin 5-5

Jepsen 5-5

O’Rourke 5-2

Perkins 1-1

Rogers 1-0

Meyer 0-1

As you can see, it's a pretty darn close match to your system.  The only differences are Perkins first outing (a scoreless mop-up inning for which you gave him a demerit), the one Abad appearance mentioned in my previous post, and two of Tonkin's appearances.  One of those he pitched 3 innings in relief of a starter's early exit -- maybe we have a special rule for such games?  And the other was allowing an inherited runner to score from second with one out, which could probably go either way.

In any case, it would be pretty easy to pull this WPA/LI info for every reliever/team, if you are interested. Might make this analysis more interesting going forward.

Strict WPA would be a pretty close match too, it would "forgive" Tonkin's 3 inning appearance but it would shift one appearance each on the ledgers of O'Rourke and Pressly, it appears.  Maybe an average of the two (WPA and WPA/LI) could effectively allow for half-credit (or half-demerit)?

I'm looking at Fan Graphs SD / MD ratings for the Twins.

You can even look at it by month, for an even comparison to Seth's numbers:

http://www.fangraphs.com/leaders.aspx?pos=all&stats=rel&lg=all&qual=0&type=3&season=2016&month=4&season1=2016&ind=0&team=8&rost=0&age=0&filter=&players=0&sort=15,d

But as you can see, only 37 of our 77 relief appearances even met the threshold for SD/MD designation.  Maybe there should be a "neutral" number in between?

Looks like pretty much the same thing... 

Let me know if you want to switch, I could run these numbers for every player/team/year/month/etc. which would give this analysis a lot more context.

Going off on my WPA/LI correlation noted upthread...

For 2016 so far, the Twins have 64.5% of relief appearances resulting in positive (or more accurately, non-negative) WPA/LI. MLB as a whole is at 67.1%. The Twins rank 14th out of 15 AL teams (ahead of only Toronto), and 23rd out of 30 MLB teams.  All four of our division mates are in the top 11.

Also, here are the results for 2015, where the Twins ranked 19th, and 6 of the top 7 teams made the playoffs:

I don't know if it's worth diving in any deeper, but at least this gives Seth's numbers some context.

Also, how does this measurement compare to, say, reliever WAR?  Here are the Twins recent MLB ranks in "doing their job" by WPA/LI, followed by their ranks in reliever RA9-WAR at Fangraphs:

2010: 15th in "doing their job" / 7th in RA9-WAR
2011: 27th / 28th
2012: 9th / 17th
2013: 13th / 8th
2014: 27th / 22nd
2015: 19th / 13th
2016: 21st / 24th

Actually, it appears RA9-WAR should be combined with LOB-Wins to account for runners stranded...