Closer of the Year: Part 2

Last week, we profiled Francisco Rodriguez' 2008 season in the first installment of our search for the closer of the year. This week, we'll turn to the obvious NL choice, Brad Lidge. Lidge was K-Rod's NL counterpart on TSN's All Star team and won MLB's award for relief pitcher of the year. He was only the second modern closer to go perfect for an entire season (Eric Gagne was the first, in 2003), going 41 for 41 in the regular season. With one Albert Pujols safely at home in St. Louis, he continued his perfection through the postseason and saved the final game of the World Series against the Rays. We can probably safely say it was the best season anyone has ever had after being traded for the worst hitter in baseball, but again we ask, was it the best of 2008?

This pattern, if you haven't noticed, is an example of foreshadowing. No, Lidge wasn't the best closer in 2008. But I'll be honest with you, this whole Closer of the Year thing was really just an excuse to look at K-Rod's and Lidge's seasons. I don't care who the best closer in 2008 was (it was Mo Rivera, by the way), I care about looking more closely at these seasons and seeing how to better assess them than just accepting the general perception of their supremacy among relievers, and finding out where the save totals go wrong in player evaluation.

I'll admit here, Lidge's season is in a much more grey area than K-Rod's. Lidge didn't pitch as well as some of his peers, but his shortcomings didn't really cost his team any games. He didn't blow a save or lose a game all year, and he only contributed a negative WPA in 4 games all year. It helped that his worst games came when his team was already trailing or tied, such as July 25, against the Braves, when he entered down 1-0 in the top of the 9th and gave up 5 runs on 18 pitches (only 8 of them strikes) without recording an out, or August 27, against the Mets, when he came in with a runner on first (his only inherited baserunner all year) with 2 outs in the 8th inning of a tie game and let that runner, plus 2 more, score before getting the third out and leaving the game. But really, while he didn't always do it spectacularly, he pretty much just came in and advanced his team's chances of winning every time he went out there.

Lidge led all relievers in WPA. He was at least in the top 10 in FIP, WPA/LI, and REW (I'm looking at you, K-Rod). He was legitimately really, really good. It is, after all, a lot harder to get by on smoke and mirrors without blowing a single save than when those devices afford you the leeway to blow 7. Still, while Lidge isn't going to be gracing the cover of Poof Magazine anytime soon, he wasn't without his more modest illusionary aids. Think resin dust and shine-balls.

Lidge surrendered a career high BB/9 and his worst K:BB ratio since his rookie year. When he settled into the Majors as a closer, he not only had absolutely filthy stuff (14.93 K/9 in his second season), he had good control (2.85 BB/9 that year). He regressed a bit in 2005, but it was another great year. With his ability to miss bats (52% contact rate in '04 and 60% in '05), get hitters to chase pitches out of the zone (led baseball in O-Swing% in '05), and throw tough pitches in the zone when he had to, he was pretty much everything you'd ever want in a closer. Then, all of a sudden, in 2006, something changed. His walk rate that had been in the high 2s jumped to 4.32. His K-rate was down as well, though still spectacular at 12.48, but the walks jumped from very good to bad. As we did with K-Rod, let's look at Lidge's career path in exR/G (1 is career average for Lidge, above that is worse, below is better):



As you no doubt already know, it is significant that this happened in 2006 because something happened at the end of 2005 (circa Game 5, NLCS) that seemed to shatter Lidge. From that point, he seemed afraid to give up home runs and trust in his stuff in the zone. He had never allowed enough balls in play to worry much about his G/F ratio (it bounced from .88 in '03 and .69 in '05 to 1.52 in '05), but in general, he had been a slight fly ball pitcher to that point, with a career ratio of .95. Since then, Lidge has consistently given up ground balls over fly balls at a ratio of 1.23. It seems this shift probably came in 2005 rather than 2006 (in which case the shift is even more extreme), but the new approach seemed to come with a more cautious tinge in 2006, if we are to take the accompanying drop in strikes he threw as any indication. In a cruel twist of fate, Lidge's HR/FB rate jumped to a ridiculously high 16.1% (he had been at around 10% even pitching in Houston) in 2006, which no doubt compounded the mental issue as well as directly hurt his numbers.

The walks have never recovered. It is probably safe to say that Lidge will never return to his days as an elite K:BB pitcher. As he showed this year, though, that doesn't mean Lidge is done. He still is, and always has been, an elite strikeout pitcher. Even as fans in Houston and the team itself gave up on Lidge, many analysts projected a return from Lidge. His numbers had been distorted by unsustainably high HR rates, and whether his confidence was shattered or not, he could still dominate hitters. Still, nobody foresaw this kind of season from Lidge, and if the walks actually got worse, where did the improvement come from? We can see in the graph above that Lidge was not severely improved in 2008, and he was nowhere near his pre-Pujols dominance.

One minor improvement was in hit batsmen. He only hit 1 in 2008, and when you factor those into the walk rate, it was actually worse in 2006. That's not particularly significant, though. Rather, the most glaring improvement came in his HR/FB ratio, which was all the way down to 3.9%. As the graph above normalizes Lidge's HR/FB rate to his career norm, this created quite a discrepancy in his actual ERA and and his exR/G. In fact, let's take a look at the same type of graph, but using Lidge's actual ERAs as our data instead of his exR/G.



Notice how the points are similar to the first graph, with the exceptions of 2006 and 2008 where his HR/FB ratios were unreasonable extremes. On the surface, it would appear that Lidge has been on a steep rate of improvement since 2006 and is all the way back to his old form. That's really not the case, though, when we look at the first graph. Did Brad Lidge pitch better in 2008? Yes. Did he pitch leaps and bounds better? No. He didn't pitch nearly as poorly as his results showed in 2006, and he didn't pitch nearly as well as his results showed in 2008.

There's more to Lidge's success than just the home run rate, though. One of Lidge's major perceived flaws over the years has been that he is prone to blow up once in a while. He'll have some bad games. In his 5 years in houston, Lidge had 8 games where he failed to get more than 1 out and allowed at least 2 runs. Seven times did he fail to record an out and still allowed a run. Compare that to Mo, who has 5 of the first type of game in his 14 year career and only 1 of the second type, which came his rookie year when he was still a set-up guy and came in to pitch to one batter, gave up a single, was pulled, and then the guy scored later. The following table compares how often Lidge has these games (expressed as a percentage of his appearances that fit in one of these categories) to how often a few other notable closers have them:

	.1 IP, 2+ R	0 IP, 1+ R
Lidge	2.2%	1.8%
Rivera	0.6%	0.1%
Papelbon	0.0%	0.0%
Nathan	1.3%	0.4%
(in MN)	0.9%	0.3%
Hoffman	1.3%	0.6%

*Nathan has two lines because he wasn't a closer or an established pitcher until he got to Minnesota, and was a starter most of his time in San Fran, so one is his career, and the other is when he was in a comparable role to the rest of the pitchers on this list

Part of Lidge's success this year wasn't so much that he completely avoided these bad games (note the two games detailed above fit into these categories), but that they came at the right times. One of them may have cost his team the game (had he shut down the opposition and closed out the game, Philly would have won because they scored what would have been the tying and winning runs in the 9th), and the other cost them extra innings, but it wasn't reflected in the box score, so Lidge remained perfect.

Lidge also narrowly avoided a couple other blow-ups that glossed over some of his weaker peripherals. A prime example was on September 16 in Atlanta, when he notched his 37th save and actually contributed his highest WPA of the year thanks to the high leverage index of the appearance despite pitching poorly. He entered with an 8-7 lead to start the 9th and walked 3 batters (including Jeff Francoeur, which is almost like striking out Joe Sewell), but stranded them all when the Braves mercifully sent Gregor Blanco to the plate with the bases loaded and 2 out. Overall, Lidge's strand rate of 82.9% was the second highest of his career and well above his career rate of 77.4%, but in high leverage situations, it reached extreme levels over 90%. Baserunners he allowed scored 2.6 times less frequently in his 19 appearances with a leverage index above 2.5 than they did in the rest of his games. The result was a gaudy stat line inflated over the overall quality of Lidge's pitching.

Part of this may just be Lidge's ability to turn it up in tight spots. Contrary to what a lot of people thought about him over the previous few years, he's always been a bit better in tougher situations. His career WPA is well above his career WPA/LI * LI. His Clutch ratings (a measure of how he has performed in high leverage situations compared to how he normally performs) have pretty much always been positive by a noticeable amount. He's shown that he's better in those situations for long enough that chances are, its legit. But the discrepancy was significantly greater this year than its ever been, to the point that most of it is almost certainly just the luck of distribution.

Lidge had an undeniably great year in 2008, judging from his results. He didn't pitch as well as Rivera, or a few other relievers in the game, but with his good and his bad coming at all the right times, it's hard to imagine anyone else repeating his success if placed in the back of the Phillies' bullpen, and that at least separates his 2008 from K-Rod's. In that way, there really is a legitimate case for Lidge as the best closer in baseball last year, or at least for his being the best year by a closer. I prefer to try to separate a pitcher's abilities and the quality of his pitching from the factors outside of his skill (like unusual distribution patterns of his performance or rates well out of line with career norms that are historically unsustainable), so when I said earlier that Lidge was not the closer of the year, it was in line with my philosophy of pitcher analysis. I can see the other side on Lidge, though. Maybe for you, Brad Lidge was the best of 2008.

The one thing that I think we can all agree on, though, is that nobody in Philly misses Michael Bourn.