About a week ago I looked at how John Gibbons has been using Casey Janssen in both save situations, and non-save but critical situations. Based on the data I concluded that it was almost certain that Gibbons had a bias towards saving Janssen for save situations, at the expense of using him critical non-save situations that in many cases are more important than save situations.
In the comments, WPGJFan suggested looking at Gibbon's previous managerial stint to see if there was a similar pattern. A really good idea, so that's what I'm going to do here, with one caveat. In my view, this analysis really only applies when a tea hasa closer who is head and shoulders the best pitcher in the bullpen. In 2005, Miguel Batista was the Jays closer, but an ace reliever he was not. In 2006, we have B.J. Ryan, a big name free agent established closer who had an excellent season, as Casey Janssen has had thus far. So it's a great basis for comparison. In 2007, Ryan missed most of the year, and Jeremy Accardo was the main closer. Since he was not an established closer, and not really clearly better than Jason Frasor or Scott Downs, I don't think it's a good comparison. By 2008, Ryan was back but Gibbons got canned in June and bounced the closer role around, so we'll stick with 2006.
I was going to go into some details before building into a conclusion, but after reading this great Slate piece, I'm going to instead put my conclusions here. In examining B.J. Ryan's usage in 2006, I don't find any conclusive evidence of Gibbons managing to the save - but with a huge caveat: the nature of that season gave him few opportunities. Therefore, I don't think there's much to be said either way, but I think looking at the season at a whole, Gibbons had a mild tendency to manage to the save in using B.J. Ryan (which I will discuss below).
The 2006 Blue Jays Season
2006 was a weird year for the Jays, as they played a lot of games that weren't really close one way or the other, which seemed to often happen in streaks. This makes it more difficult to maximize value from the best relievers in the bullpen. Of B.J.'s 65 games, 14 appearances were in garbage time, with leverage below 0.5. However, on average these came 3.5 days after his last appearance, with only two of them less than 3 days after - so it was just Gibby getting him work, not egregiously misusing him.
Nonetheless, those 2006 Jays did have a reasonable number of games which produced likely save situations, racking up 43 FSS. By contrast, the 2013 Blue Jays (through June 8) have accumulated 13 FSS in 61 games, which prorates to about 35 in total, well shy of the 2006 total. If they start winning more, that should pick up. What the 2006 Jays did not have was a lot of games that were tied late or went to extras, which tend to produce non-save, critical situations. In 2013, the Jays have already played in 12 games featuring FCNS. The 2006 Jays had a total of 6. Simply put, the lack of those situations in 2006 makes it difficult to come to a conclusion about whether Gibbons would tend to refuse to use his closer as the sample is really too small.
What Can Be Said About 2006?
In total, of the 43 FSS, B.J. Ryan was used in 38 of them (88%). Of the 5 times he was not called on, 4 were situations where a 1 to 3 run Blue Jay lead after their opponent batted in the 8th was extended to render the 9th a non-save (very low leverage) situation. The other one was the last day of the season, a completely meaningless game. So essentially without fail, every time that there was a FSS and the save opportunity materialized, Ryan was used.
In the 6 FNCS, Ryan pitched in 5 of them (83%). At a high level, those are similar rates of usage, so one can't say he was definitively managing to the save based on the data. Given the limited sample of the latter, one can't reject the possibility either - three of them came with 4 or more days of rest, so it would have been frankly stupid not to use a well-rested relief ace. In both FCNS on 2 days rest, Ryan was used. The only time he wasn't used, he had been used the day before, but was pretty well rested having only pitched once in week before that.
But That's Not Cased Closed
As for my contention that Gibby had a mild tendency to manage to the save, I submit the following points. Keep in mind that these are observations from 7 years after the facts, looking at boxscores from a high level. I do my best to contextualize as best I can, but it will be imperfect.
1) As noted above, with the exception of game 162, in every FSS in which the save opportunity did not disappear due to a lead being extended, Ryan was used. Not once was there occasion to leave the same reliever in the game. However, in the 4 times a 1 to 3 run lead was extended to more than 3 in the half inning immediately before the opponent's lack hacks, only once did Gibbon send Ryan out even though it is likely he would have been warming for the imminent expected save opportunity.
Moreover, in 2006 there were 7 games in which the Jays had a 3 run lead after their opponent batted in the 8th and Ryan had not entered the game (other times he came to a 3 run game in the 8th, usually with runners on). Four times they did not score and the 3 run lead held going to opposing team's 9th inning. All four times, B.J. pitched. Three times the lead was extended to 4, but Ryan only came in once. This suggests managing to the save, since the change in leverage from a 3 run lead to a 4 run lead is that that significant.
2) In addition to the 38 FSS appearances for Ryan (recording 34 of his 38 saves), there were 4 other games where he got a save because a save situation developed. Two of these were at least reasonable usages, but two in my view indicate bias towards racking up saves:
A : 5/23/06 - Halladay took a 4-1 lead to the 9th against TB (who were still terrible at this point). After getting the first two outs, he allowed a single and was removed at 110 pitches. In came B.J., in a low leverage situation to get the one out save. But what makes this egregious? It was the third straight day he was used, having thrown 50 pitches in the two preceding games (32 in a stressful 6 batter, 4 out, 1 run lead save the day before).
B: 6/3/06: Jays took a 4 run lead to the bottom of the 9th. With two on and two outs, Ryan was brought to get the last runner. This was a low leverage situation in which the tying run was not even at the plate. Would he have been brought in if not for it being designated a save situation?
3) In 2006, Gibbons frequently inserted B.J. into the 8th inning of a game where the Jays lead by 1 to 3 when he came in. It happened 14 times. In two of those, the Jays extended the lead so that Ryan was still eligible for a save if he came back out for the 9th, but the situation was now very low leverage. In the first, on April 25th with a five run lead against NYY, he pitched the 9th having faced 3 batters in the 8th. I suppose this can be defended, but this was the 25th day of the season, Ryan's 11th appearance, and 12th inning pitched. In the interest of keeping him as fresh as possible for the remaining 5 months, was sending him out the best decision?
In the second such situation, on June 2, Ryan faced one batter and got the last out of the 8th to preserve a 4-3 lead. The Jays then scored 9 runs in what I imagine was a very long top of the 9th. I doubt any reliever would come back out in that circumstance
4) Twice Ryan pitched three straight days, something that is generally considered prudent to avoid wherever possible. One sequence is discussed above, but in the other the second appearance was a one inning save up 3 runs. So, in both circumstances, it would have been possible to avoid using him three days consecutively with decision making based on leverage and ignoring easy save situations.
Considering all of the above, I think there is a case to be made that there was some managing to the save by Gibbons in 2006.
A Couple Points on Defining FSS and FCNS
In the course of looking through 2006, as well as observing the bullpen management in last weekend's Padres series, I found it necessary to modify how I defined FSS and FCNS in Part I.
First, is the situation where there is a FCNS occurs, and no pitching change is made. This occurred in the Baltimore game I discussed in part 1: Oliver had an easy 8th, and came back out in a tie game for the 9th. When this occurs, it's hard to criticize the manager. Leaving aside the merit of sticking with a guy who is pitching well, just the uncertainty in how long tied games in the 9th demands conserving relievers where possible. So, when there would otherwise be a FCNS and the reliever preceding the FCNS inning stays in, the FCNS is deemed not to occur.
Likewise, with FSS, if the starting pitcher is still in the game, and comes out for the 9th with a 1 to 3 run lead, that doesn't really tell us anything about bullpen usage, and including these in total number of situations could bias the data. Therefore, when a starting pitcher remains in the game in what would otherwise be a FSS, a FSS is not deemed to hav occurred.
Finally, when originally doing this with Janssen, it was really easy because Janssen is being exclusively used in the 9th. However, B.J. Ryan was frequently brought in the 8th, and kept in for the 9th. We definitely want to count these in FSS. So, FSS include situations where the closer enters in the 8th in a situation where by finishing the game he is eligible for a save.