from Amazon
Saturday, September 16, 2006
Willie Randolph and Billy Wagner
Willie Randolph said this:
I’ve used him so many times to nail down an important win for us in non-save situations. He’s given our bullpen a sense of completeness that we didn’t have last year.
He’s talking about Wagner. Thanks to Fangraphs, we can look at this assertion. Breaking up his games into whether he got a save or blown save, and into games where he got neither:
SV+BS: 43 games, 177 batters faced. In games where he recorded an overall positive WPA (which are all of his saves, natch), he had +4.49 wins added. In games where he had a negative contribution (his blown saves not coincidentally), he was -2.29 wins. So, when he comes into a save situation, he’ll add twice as much win probability as he will remove. Overall, he’s +2.20 wins in these situations.
Others: 22 games, 97 batters faced. +1.76 wins in the positive games, and -0.58 wins in the negative games for a total of +1.18 wins.
Think about that. He’s +2.20 wins in the save-type games, over 43 games, for an average of +.051 wins per game. And in the non-save games, he’s +1.18 wins over 22 games, for +.054 wins.
If we count the number of games where he’s had a big impact (+/- .10 wins), I count 33 positive impacts, of which he’s only gotten 24 saves. He’s had 7 negative impact games, with 5 blown saves.
If we also look at the mild-impact games (+/- .04 to .10 wins), he’s had 14 mild-positives, and no mild-negatives.
It certainly looks like Randolph has used Wagner in non-save situatiions, and did so with Wagner having great impact.
***
How about Rivera?
He’s had 36 save-type games, facing 150 batters. His WPA was +2.74, with +4.24 in save games and -1.50 in blown saves. Overall, he’s +.076 WPA per game. Great stuff!
How about non-save games? 23 of those, facing 126 batters. Rivera certainly is not used in many save situations! And how’d he do here? +0.53 wins, with 10 high-impact games. However, 6 of those were high-positive, and 4 were high-negative. He was +1.68 in the positive games, and -1.15 in the negative games. Certainly Rivera didn’t fare so well in these non-save games. But those 4 high-negative games were just as bad, or worse, than his blown saves. He’s +.023 wins per game.
One more with Papelbon, this year’s version of “who will knock off Rivera?”.
41 save-type games, facing 173 batters. WPA of +3.62, with +4.72 in save games and -1.10 in blown saves. Overall, +.088 per game.
In non-save games, 18 games, 84 batters, for +1.6, giving him +.089 wins per game! Wow. Included in here are 5 games where he pitched at least 2 innings, and gained +.30 to +.40 wins in each game: 21-Apr, 13-Jun, 24-Jun, 29-Jul. 12-Aug.
If we count his high-impact games, he’s had 28 high-positive and 21 saves to show for them, and 4 high-negative (with 3 blown saves).
For Rivera, he’s had 24 high-positives, with 18 saves on them, and seven high-negative (with 3 blown saves). This is not the Rivera of old, but it is the old Rivera.
***
(Note: Fangraphs does not record Holds, and therefore, there may be games where Wagner came into a save situation, and still left it in a save situation for someone else. Since we are talking about Wagner, it’s likely he has no holds.)
Of the top 50 relievers in terms of WPA, only 3 of them have +WPA > ABS(-WPA). And 2 of them (Cla Meredith and Mark Lowe) aren’t even closers. The only one who is a closer is Joe Nathan. When you think about it, that is a rather remarkable achievement for a closer. Since they always come into situations where their team already has a high probability of winning, their downside risk is much greater than their upside potential.