I wanted to look at the performance of the bullpen when it really matters - when the game is in the balance. I define that as any time from the 6th inning on and the Cardinals have a lead of one, two, or three runs, are tied, or are trailing by 1 run, or qualified for a standard save (I made no allowances for a three-inning save). When I collected the data, I gathered up stats for both starters and relievers, so I am publishing both sets of numbers. This is not meant to be anything about true talent level or trying to be predictive, just reporting what happened and giving it a little narrative.
Going through the stats: WPA is Win Probability Added (Fangraphs version), a measure of the impact of any given play based on game situation; Out is the total outs generated by the pitcher; Rch is the number of batters that reach against the pitcher; Run is the runs scored against the pitcher while he is on the mound (including other pitchers' inherited runners); ROW is runs on watch average, similar to earned run average but using the Runs mentioned before; and WHIP is similar to its standard counterpart, but uses Rch instead of walks and hits.
For the chart with Saves, Holds, and Runs, the definition for a save is no different than normal. I applied the standard rules for a Hold for relievers for leads (L Held), tie ballgames (T Held), and one-run deficits (I held) (enter the game in a certain situation, get at least one out, and leave the game/end the inning no worse than how you entered); for the tie and deficit situations, I didn’t give credit for someone that caused more harm than good (i.e., got one out, but put two on) and was pulled mid-inning. The starters are handled basically the same except that they lose credit when they are in the lead as well. The Runs are those runs scored against a pitcher when he is on the mound that either blow the lead/save (BSR) or tie (BTR), or allow the insurance run (BIR) to score (put the deficit at two runs); a walk-off, two-run homer in a tie ballgame only counts as a BTR, but a three-run double with a one-run lead counts for one each (BSR, BTR, BIR).
|Player||Save||L Held||T Held||I Held||BSR||BTR||BIR|
For each player:
Motte – Although he wasn’t the best performer in terms of game impact, Sauce was certainly the stopper, allowing few baserunners and striking out a ton. He gave up more runs and more big runs than anyone else, but he was a net positive regardless of what inning he worked (8th +.221 WPA, 9th +.678, 10th +.537).
Boggs – Boggs had the best WPA among the relievers, the biggest reason why the 8th inning was two full wins better than any other (6th thru Extras: -.868, -.465, 2.348, .238, -1.590).
Salas – Probably the biggest whipping boy among the relievers, but he rarely had the chance to succeed (justified in that he seemed to bring an extra can of gas to the mound with a lead, registering 16 outs against nine baserunners and seven runs scored).
Rzepczynski – As long as Matheny brought in Scrabble in the 8th or later he was fine (WPA in 6th/7th -1.497, in the 8th 0.666).
Mujica – Would have been nice to have him all year long – Mujica shut down the 7th inning (WPA 1.129) in the two months he was here. The few times he ventured outside that comfort zone were less than optimal.
Marte – Victor was actually pitching well for us, posting a WPA of +.592 thru 6/17, but he coughed up over a full win’s worth of runs from then until the All-Star break and didn’t see quality time again until October.
Freeman – Sam had an excellent ROW (which is sometimes created by LOOGY usage patterns) and passable WPA and WHIP numbers.
Browning – Barrett’s numbers were good in the 6th/7th or with a lead, but sucked from the 8th inning on or when tasked with a tie ballgame or worse.
Rosenthal – Rosie’s only blemish was an extra-inning walk-off by Anthony Rizzo. Otherwise, you are looking at seven innings, no runs, and ten strikeouts.
Lohse – Kyle was the only starter with good numbers in both the sixth and seventh innings. Nobody with any significant innings had a worse strikeout rate, though.
Westbrook – Westy posted a WPA of -.652 in the sixth inning, +.385 afterward. That seems to be a fairly common pattern consistent with the eyeball test: the manager will let a starter who is doing well with a lead keep going until he starts to get into trouble – any damage gets done against the bullpen.
Lynn – Lance actually had better numbers than any other starter (ROW, WHIP, K’s) when the game was on the line. His struggles as a starter occurred on the other end of the spectrum and only became visible here as a reliever.
Wainwright – Within these situations, Wagonmaker performed poorly in every case: 6th or 7th inning, leading, tied, or trailing. Tip to the Cardinals – get him a big lead.
Garcia – As good as the relievers were in the eighth inning, the entire staff sucked that bad in the sixth (-2.084). Jaime was the worst culprit, collecting 33 outs while putting on 24 baserunners and surrendering eleven runs, never mind the inherited runners he handed to the bullpen (overall bullpen in the 6th, 17.2 innings, 17 runs scored).
Kelly – Joe had the best numbers of any pitcher overall, with a combined ROW under two runs. He did allow a tie-breaking run in extras, but he was otherwise excellent, his 5-2/3-inning relief appearance his major highlight.
Sanchez – I was impressed with his stuff in 2011, but last year’s numbers were awful. A WHIP of 3.00! and a WPA of -.808 in only five innings pitched.
McClellan – He was bad.
Miller – We’re obviously dealing with SSS, but one strikeout among eight outs recorded? Tsk-tsk.
Fuentes/Cleto/Romero/Carpenter – Not enough here to work with.
Another item to look at: the initial downward spiral started with the walk-off single by Joe Mather on April 23rd. Through the eighth inning of that game, the team WPA in these situations stood at 1.273; by the end of May, it had plummeted to -3.513. The team numbers in the other five months were basically at or below their season averages, but the WPA in May was -3.268 with a WHIP of 1.78 and a ROW of 5.86.
One more while I'm thinking about it: I did something like this when I looked at the 2008 team. For the first 100 games that year, our bullpen was trying very hard to give away nearly every game and the offense tried to make up for their mistakes. (The final 62 games saw the bullpen revert to neutral, but the offense stopped caring, undoing all their good work from before.) Anywho, the offense was tested all the time and managed to score 54 decisive runs (tying, go ahead, or insurance, 6th inning or later) over their first 100 games, an admirable total. In comparison, last year's pitching staff allowed 127 decisive runs over 162 games, including 22 blown saves and 14 blown leads by the starters.