The bullpen in save situations

by chuckb on Nov 12, 2008 6:00 PM EST

This is a comment from Solanus that I pulled over from today's daily thread. I thought it would be better served as a fanpost where we could all access it and comment on it. -- chuckb

Link to the original comment

Our biggest problem last year was probably our inability to hold on to leads. Knowing this, I thought it might be interesting to get a statistical picture of just how effective each our bullpen guys were during save-type situations.

I went through the game logs on Fangraphs.com and parsed out any outcome that was: produced by a reliever; occurred in the 6th inning or later; & occurred with a 3-run-or-less lead and/or qualified as a save situation (4-run lead & 2 runners on, or 5-run lead & bases loaded). (Sometimes you can have a 3-run lead and it is not a save situation, but it wasn’t worth quibbling over.) Anyway, here they are:

WPA	Name	Out	Reach	Run	BS	Sv	IP	ROW	WHIP	Holds	K

-2.670	Team	416	219	88	31	42	138.2	5.71	1.58	106	112
-.179	Franklin	109	50	23	8	17	36.1	5.70	1.38	13	26
.951	McClellan	108	57	17	5	1	36.0	4.25	1.58	30	27
-2.921	Izzy	47	40	21	7	12	15.2	12.06	2.55	2	14
-.722	Perez	39	15	7	4	7	13.0	4.85	1.15	6	15
.511	Springer	31	13	4	2	0	10.1	3.48	1.26	15	11
-1.057	Flores	22	22	9	2	1	7.1	11.05	3.00	14	4
.059	Villone	20	11	3	1	1	6.2	4.05	1.65	16	6
.717	Motte	14	2	0	0	1	4.2	0.00	0.43	4	5
.295	Garcia	12	3	0	0	0	4.0	0.00	0.75	3	1
-.058	Thompson	9	2	1	1	0	3.0	3.00	0.67	1	0
.040	Reyes	5	1	2	0	1	1.2	10.80	0.60	2	3
-.306	Mulder	0	2	1	1	0	0.0	INF	INF	0	0

Going through the columns: WPA is Win Probability added, a measure of game impact (see Fangraphs); Out is the number of outs recorded during "save situations"; Reach is the number of batters that reached base; Run is the number of runs scored while a reliever was pitching (not to be confused with earned runs); BS is blown saves; Sv; IP; ROW is Runs On Watch Average, basically the Runs listed before, converted over like ERA; WHIP, similar to the normal acronym, except I included all batters who reached; Holds; & K

(I have no numbers to compare any of this to; I did all of this by hand, a few minutes at a time at work, so I haven’t gotten around to doing all of the teams. Unless someone knows how to easily pull up this specific data online, I probably won’t complete the "study".)

Picking through the numbers, Izzy shines like a beacon, or whatever the "I suck at this" equivalent is. He recorded 47 outs while putting 40 batters on base, allowing 21 baserunners to score. Plus his WPA would indicate that he cost the Cardinals almost three full games in the standings.

Flores was nearly as bad, matching up his 22 outs with his 22 batters reaching & plating 9 guys as well. His Hold total looks impressive, given how much time he missed, but it is important to remember – Tony would not leave Randy (nor Villone, really) out long enough in a tight spot to really screw things up on his watch. Going through each particular at bat with a 1-run lead, Randy faced 14 batters; he recorded 5 outs and put 9 guys on base, throwing a WP, allowing 3 stolen bases, & blowing two save opportunities.

Speaking of Villone, his Hold total (16) & one lone save are amazing considering he only recorded 20 eligible outs.

Franklin & McClellan have rather similar numbers (IP & K), so let’s compare the two. Franklin was just a touch below average, according to WPA, where Kyle was over a full win better. This can be greatly attributed to when Franklin blew his save opportunities (1 in the 8th inning, 5 in the 9th, & 1 in the 10th) vs. when McClellan did (1 in the 6th & 4 in the 8th). Also noteworthy is that Kyle was actually 31 for 36 in save opportunities, although he was only credited with one save (30 holds).

Perez wasn’t really that bad, especially compared to the team averages (better ROW, K-rate, much better WHIP). His blown opportunities rate was higher (27% v 17%), but he didn’t have the luxury of the LOOGY one-out-hold.

Springer was under-utilized despite terrific numbers. One reason that I can see for this is that Russ has exhibited problems in the past coming out for a second inning or consecutive days. Tony probably factors that into his role-selection and has to plan out Springer’s usage patterns a little more; someone like Franklin he can "abuse" a little more with fewer ill effects. (NOTE: Russ’ 8th inning numbers were rather poor, including both of his blown saves.)

The rest of the group was quite proficient (not including Mulder). Motte & Garcia combined allowed no runs & 5 baserunners while recording 26 outs. Thompson did fairly well, although he had one blown save and most of the outs didn’t technically qualify as save situations. Reyes only allowed one hit against 5 outs, but it was a 2-run HR (no real harm – it was with a 3-run lead).

Tweet Comment 35 comments | 5 recs |

Do you like this story?

Comments

Great work, Solanus.

by chuckb on Nov 12, 2008 6:01 PM EST reply actions

+1

by spants on Nov 13, 2008 2:30 AM EST up reply actions

That Team WHIP...

Is really depressing. No wonder we let so many games slip away.
Nice analysis, too.

by JWO on Nov 12, 2008 6:21 PM EST reply actions

My impressions

This screams one thing we already know…that we desperately need to find a real loogy. That’s a role that, if performed somewhat competently, lends itself to good results on WPA. You are there to get that one or two big lefthanded outs…you’re supposed to have the platoon splits that give you the big advantage over time, you are usually placed in relatively tight spots, and the team will often hand the ball off and let a righthander finish the clean up in the inning.

You shouldn’t have a negative WPA in a specialty role like a loogy, and that Flores did that poorly in a specialized situation (for a few years now) tells you what you need to know. I hope the team can address this effectively.

Other than that, it is interesting that as a whole (Springer excluded), the “reliable” veteran hands weren’t that reliable at all. I understand they probably want a steady hand as a closer. I can appreciate that, and am actually kinda on the fence about it myself. But for the love of God, if we are going to go with a veteran closer can we keep ourselves in a short term contract somehow? The kids show a little ability here, and the bullpen isn’t a good place to make long term committments.

by Merry CRasmus on Nov 12, 2008 6:27 PM EST reply actions

Very nice analysis

I wondered about this during the regular season while Franklin was sputtering and I’m curious to know if anyone else thinks K-Mac would have a realistic chance at being our closer next year? I think he’s probably better suited for the longer relief roles and using him in just one inning at a time is more than likely a waste of his ability…but has anyone else here considered this or am I just crazy?

Patiently awaiting the day Colby Rasmus does this: .275/.381/.551/.932, 29HR, in St. Louis...

by RunninRedbird on Nov 12, 2008 6:53 PM EST reply actions

Our strikeouts were so low out of the bullpen

A strikeout type bullpen can go a long way even if they aren’t great. The next year we will definitely improve even without any additions because we will have strikeout guys pitching full time like kinney, motte and perez.

by vivaelpujols on Nov 12, 2008 7:29 PM EST reply actions

Question

Is there anyway to weight the ROW for inherited runners? I think WPA addresses that somewhat, but I just wonder if it might be (even) more insightful if we could take that into account. For example, I remember one of Perez’ blown saves was in a one-run game where he entered with a man on third and nobody out. That run that scored on a SF shouldn’t carry the same weight as the ridiculous number of times Flores came in to start and inning and walked the leadoff man.

I love what you did – thanks for the hard work.

Those Pilgrims ain't lookin' so proud now...

by giveml on Nov 12, 2008 9:36 PM EST reply actions

I thought that runs on inherited runners were always charged to the pitcher that put them on

even if it was someone who just got a single.

They say that it's never too late, but you don't get any younger...

by Valatan on Nov 12, 2008 9:56 PM EST up reply actions

That is true

but I interpreted the comments to mean that Runs on Watch is calculated by who is pitching when the runs actually scored, not who is charged them for ERA purposes.

Those Pilgrims ain't lookin' so proud now...

by giveml on Nov 12, 2008 10:40 PM EST up reply actions

Hand-crafted statistics

I just read each game log off Fangraphs, typed the save-situation-worthy numbers into Excel, and gleaned info from the data. I didn’t really worry about which runners were inherited & from whom.

That said, the Cubs do deserve my pity, but never my support.

by Solanus on Nov 13, 2008 10:27 AM EST up reply actions

How about...

Josh Kinney as a closer???

by Elpot on Nov 13, 2008 2:21 AM EST reply actions

he would definitely be a candidate, but i think he falls in with k-mac in my thinking.

both have the potential but no track record. so, i would definitely put him in the mix ahead of franklin if our plan A closer flops. but i’d hate to pencil kinney in as our Plan A closer, without any prior closing experience and without more than a September’s showing that he is even healthy. I would rather start with Perez as Plan A and have Kinney be my third or fourth choice in the role.

but I agree that Kinney has the potential.

by tom s. on Nov 13, 2008 2:29 AM EST up reply actions

Kinney may not have the track record

But unlike McClellan, he has been dominant in his brief time in the bigs. In 32 innings he has allowed just 20 hits and 9 walks while striking out 30 for a 2.53 ERA. Those are closer-like numbers. In the postseason he has a 0.00 ERA and has given up just 3 hits in 6 1/3 innings while striking out six. I think that the closer position is overated, and if Kinney has a good spring then we should consider him the top candidate. I think that the difference b/t a guy like K-rod and a guy like Kinney isn’t that much in terms of wins and losses to the team. If Kinney can put up numbers like he has so far in the bigs during the course of a full season, then he could be just as good as some of the dominant closers in the game. The only difference is that Kinney costs next to nothing which would allow us to improve in other areas.

by vivaelpujols on Nov 13, 2008 3:00 AM EST up reply actions

Well

When you use the “small sample size” rationale, go back and look at McClellan’s first 52 innings in 2008 (his first half split):

2.94 ERA; 41/12 K/BB ratio; 1.21 WHIP, and a .680 OPS against.

He was nearly as dominant his first 38 innings as Kinney was before Kinney got hurt. KMac just had a terrible second half, which would tend to happen to a guy in his first year in the big leagues, his first year in a middle relief role, and having pitched 52 innings in the first half of the season. Considering the number of games the Cardinals had played, that extrapolates to a little over 90 innings for the year. Very few good middle relievers throw that many innings in a season.

Suffice to say, he was leaned on very heavily in the first half of this year and I think his second half split is an indication of that.

"I just wish that the late Harry Caray were still around so I could hear him mispronounce 'Kosuke Fukudome' every fukun' night" -- Dennis Miller

by fourstick on Nov 13, 2008 11:13 AM EST up reply actions

totally agree

i love josh kinney, and think he’ll make a huge difference in 2009. In fact, I wouldn’t complain if we went into 2009 with him as our declared “closer”.

Just wanted to comment on KMC’s usage in the first half of the season- 52 innings as a reliever- that makes my stomach hurt.

the enemy's gate is down.

by SleepyCA on Nov 14, 2008 12:41 AM EST up reply actions

It's K-mac fault for being so darn good

by vivaelpujols on Nov 14, 2008 1:00 PM EST up reply actions

I think as highly of Kinney as most,

but he’s only pitched seven innings in the last two years.

To say that the “only difference” between him and K-Rod or other dominant closers is that Kinney is cheap is what my friend Mr. Greenspan calls “irrational exuberance.”

I wish him well and I hope he’s healthy. He can be a knock-out pitcher if he’s healed. But I can’t believe that following the Cardinals for any period of time hasn’t given you a decent amount of skepticism about pitchers coming back from serious injury.

by tom s. on Nov 13, 2008 3:26 AM EST reply actions

I am saying that if Kinney is healthy and can put up close to his career numbers over a full season

than he would be as good as Krod.

by vivaelpujols on Nov 13, 2008 3:46 AM EST up reply actions

Actually

If he did that over a full season he’d be better than K-Rod.

I just don’t think a guy coming off of two years of rehab is going to throw 70+ innings for the Cards next year, which is what we would need from him if he’s given the closer’s role.

"I just wish that the late Harry Caray were still around so I could hear him mispronounce 'Kosuke Fukudome' every fukun' night" -- Dennis Miller

by fourstick on Nov 13, 2008 11:14 AM EST up reply actions

not neccesarily

we have excellent middle relief depth with Perez and Motte and Franklin and McClellan and hopefully two good loogy’s. THerefore would really only have to use Kinney in save situations.

by vivaelpujols on Nov 13, 2008 2:24 PM EST up reply actions

You can't just use your closer in save situations

He needs regular work to stay sharp, just like every other pitcher out there, and that is especially true for bullpen guys. As a closer he’d pitch in most save situations and probably every 3rd day otherwise, 1 inning or less each time out, which relates to about 60 or 70 innings.

"I just wish that the late Harry Caray were still around so I could hear him mispronounce 'Kosuke Fukudome' every fukun' night" -- Dennis Miller

by fourstick on Nov 13, 2008 5:05 PM EST up reply actions

Further notes

Here are the team numbers for the 9th inning save opportunities, 8th inning, & other opportunities.

Inning	WPA	Out	Reach	Run	BS	Sv	IP	ROW	WHIP	Holds	K
9th	-2.698	128	74	39	13	-	43.1	8.23	1.73	n/a	39
8th	-.932	131	72	24	13	-	43.2	4.95	1.65	n/a	29
Other	.960	155	73	25	5	-	51.2	4.30	1.39	n/a	44

Without spitting out all of the individual numbers, here are the highlights:

9th inning – Franklin was surprisingly effective, not hurting the team (neutral WPA) and having good peripherals (low WHIP, 1K/IP), but yielded plenty of runs; Perez was the exact opposite, in fewer IP; Izzy was unsurprisingly ineffective

8th inning – Franklin was outstanding as the setup guy; Perez was possibly better; McClellan struggled in the 8th & 9th; Izzy should never be brought out before the 9th, if at all; Villone was OK, Flores was horrid

Other opportunities – McClellan’s best role was the 7th; Springer was even better, but usually relegated to lower leverage situations; Franklin & Perez seemed rushed into games, as both struggled when brought in early; Flores & Villone combined for 20 outs & 19 batters reaching base

One other option I might take a look at is the Bill James model of what a closer should be (more focused on tying-run, go-ahead-run, & insurance-run situations) and see how we fared / who was used.

That said, the Cubs do deserve my pity, but never my support.

by Solanus on Nov 13, 2008 4:25 PM EST reply actions

All that tells me

is keep franklin, don’t resign Izzy, villone or flores.

by vivaelpujols on Nov 13, 2008 7:39 PM EST reply actions

Addition by subtraction

We improve dramatically merely by subtracting Izzy and Flores. Mix in a full season of Kinney and I feel very comfortable with our bullpen (except for the loogie slot). I just think it will be a waste if we go and blow salary on a closer. Turn the job over the young guns, resign Springer, and be done with it. Why is that such a hard decision for Tony and Dunc?

by jjray on Nov 13, 2008 11:56 PM EST reply actions

I agree

Although, I think that we should make an attempt at signing Hoff to a 1 or 2 year deal. He would be cheap and he would allow Tony not to have to rush Perez and Motte. Also the effect that Hoff would have on those 2 young pitchers would be great. If he can teach Motte a change up than he would pay for himself.

by vivaelpujols on Nov 14, 2008 12:48 AM EST up reply actions

refresh my memory

is 1.0 WPA a win, or is it 0.5?

the enemy's gate is down.

by SleepyCA on Nov 14, 2008 12:44 AM EST reply actions

IIRC

I think +1.000 WPA means that the team’s record improves by one game in the standings (add one win + subtract one loss). You are limited to a finite number of games to work with; a player is not going to “allow his team to play one more game and win that one”, he’s going to “turn a loss into a win”.

That said, the Cubs do deserve my pity, but never my support.

by Solanus on Nov 14, 2008 9:57 AM EST up reply actions

I thought .5 counted as a win

since each game started out at 50-50 (even though 50-50 is a somewhat flawed premise).

by azruavatar on Nov 15, 2008 8:21 AM EST up reply actions

You're correct, sort of

The .5 counting as a win is correct, but it is like taking the current record and adding a win, not replacing a loss with a win.

It would be nice to see some sort of home-away adjustment, but that starts adding arbitrary values into the mix (do you use historical home field advantage? for that particular season only?). Then again, Fangraphs adjusts there values to reflect yearly changes in offensive numbers, so it probably wouldn’t be that hard to implement.