In case you need reminding of what Koji did last year: from June until September Koji appeared in 53 games, pitching 53.1 innings, allowing only 17 hits, 5 runs (only 4 earned - 0.68 ERA), and 5 walks, while striking out 72 batters. During that time period, batters posted a .097/.126/.154 line. His BABIP was a ridiculous 0.155. He allowed only 1 HR during those 53 innings. He continued that dominance into the playoffs and we all know what happened in October.
Expecting Koji to post those numbers again is both unfair and unrealistic. For one, those 53.1 innings were possibly the greatest stretch by a reliever in baseball history. And for two, Koji is 39 years old and coming off of the most innings pitched in his MLB career by far. He used to be a starter in Japan, but hasn't' started since 2009 when he first joined Baltimore. Even then he only pitched 66 innings over 12 starts. However, measuring this year compared to most of last year can be instructive. For one, its what Koji aspires to replicate this year. Plus, we can tell what he isn't as successful doing this year compared to last, since the distinction will be brighter. And who knows, maybe he'll come close to it and we can all be happy. Even a slightly worse Koji will still be a very valuable asset.
On the surface in 2014 we have seen the tale of 2 Koji's. For the first 5 games of the year he was just as automatic as before, pitching 5 innings, allowing 3 hits, no runs, no walks, and striking out 7. Hitters were only hitting .158/.158/.211 with a .250 BABIP. Then he was shut down for a week with shoulder tightness. Since returning from his week off, Koji has been a different pitcher: 10 games (9.2 IP), with 10 hits allowed, 2 earned runs (1.86 ERA), 3 walks, and 15 strikeouts. What has changed is that batters are hitting .278/.333/.444, with a BABIP of .421, off of him in that 10 game stretch. The BABIP number really stands out to me as it is way above his career norm:
Outside of 2010, his BABIP has hovered around .200. This is interesting for a couple of reasons. The first is that pitchers generally have little control over their BABIP. They rely on defense to make plays, and a bad defense can increase the BABIP. The second is that BABIP is usually not as consistent from year to year, given the just mentioned defensive fluctuations. Koji has kept a consistent BABIP over 4 seasons, playing for 3 different teams. Even if we expect some regression, its fair to say part of Koji's struggles since coming back have been bad luck related. Koji said last week (link) that this is an April/May thing and that last year he didn't pitch that well during the same period. I took a quick look at his stats from last April/May and found that he was partly correct. He still pitched very well for relief pitching standards, but nowhere near what he did from June to September: 22 games (21 IP) allowing 16 hits, 5 earned runs (2.14 ERA), 4 walks, and 29 strikeouts. Batters hit .205/.244/.423 with a .267 BABIP. It is certainly possible that he is a better warm weather pitcher than cool weather pitcher. We know that his second half stats are vastly better than his first half stats for his career. His K/BB ratio is six times better in the second half than first half, and his ERA drops by a run.
However, its still important to look at that .421 BABIP and see whether this is just typical early struggles, or a result of age/injury-related decline in stuff. I took a look at his batted ball profiles for 4 different periods: April/May of 2013, June-Sept of 2013, 2014 pre-shoulder injury, and 2014 post shoulder injury. As always, the 2014 stuff is subject to SSS bias, but if any trend sticks out we should be able to see it. The batted ball data is from Baseball-Reference, which was the only place to aggregate intra-month data. Fangraphs has his GB% as lower, and LD% about 10% lower. I cross-checked with pitch F/X data from BrooksBaseball and both his fastball and splitter LD/BIP percentages were in the mid 20s, so I kept the B-Ref data.
[tablegrid= Koji Batted Ball Statistics ]Stat April/May 2013 June-Sept 2013 2014 Pre-Injury 2014 Post-Injury Strike% 70% 76% 75% 68% Strike Looking % 19% 17% 19% 13% Strike Swing % 15% 20% 19% 19% GB% 32% 46% 33% 36% FB% 68% 54% 66% 64% LD% 12% 14% 25% 23% PU% 14% 17% 33% 9% K% 35% 40% 37% 38% BB% 5% 3% 0% 8% BABIP 0.267 0.155 0.250 0.421 Pitches/Batter Faced 4.03 3.92 3.58 3.90 [/tablegrid]
A couple of things stand out. The first is that the main difference between April/May 2013 and June-Sept 2013 was the reduction in fly balls and the increase in ground balls. His line drive and pop-up rate stayed consistent between the two period, but he traded fly outs for ground outs. Since ground balls are more likely to go for hits than fly balls (ex line drives), we can assume he had weak contact on the ground balls, and a solid infield defense behind him, or both.
What is driving that high BABIP this season is likely the large increase in line drive percentage, since line drives fall for hits about 70% of the time. Since coming back from his week off for the sore shoulder, his strike percentage has decreased, especially the strike looking percentage. Before his injury he was getting lucky that the spike in LD% wasn't resulting in more hits. His walks have increased as well, which is expected when strike percentage decreases, but he still isn't walking the park.
This probably raises the question of why the LD% has spiked all season (before and after injury). Koji's main issue this year, compared to June-Sept of last year is his performance against RHH. Righties have feasted on both his fastball and splitter this year:
[tablegrid= Koji Batted Ball vs RHH ] 2013 June - Sept 2014 2013 June-Sept 2014 Pitch Outcome Fastball vs RHH Splitter vs. RHH Strike 35.5% 31.0% 41.0% 31.3% Ball 22.8% 29.6% 20.8% 21.9% Whiffs 12.7% 12.7% 29.5% 25.0% BIP 14.8% 15.5% 16.2% 21.9% GB/BIP 32.1% 18.8% 57.1% 71.4% LD/BIP 17.9% 27.3% 10.7% 14.3% FB/BIP 25.0% 36.4% 17.7% 14.3% PU/BIP 25.0% 18.2% 14.3% 0.0% AVG 0.143 0.429 0.074 0.308 SLG 0.262 0.714 0.130 0.539 BABIP 0.214 0.455 0.103 0.500 [/tablegrid]
Although he hasn't succeeded with both pitches, his fastball seems to stand out as the worse pitch against RHH. It is an easier pitch to command than the split, and we can see that hasn't thrown it for as many strikes this year. The whiff rate is the same, and the amount of times it is put in play has slightly up ticked, but the issue is what happens when it is put in play. Compared to most of last year, the ground ball rate has plummeted and the line drives have spiked, leading to right-handed batters slugging .714 against his fastball. The usage of the splitter has completely changed this year. As you'll see in the charts below, he has switched the side of the plate he is working RHH with the splitter.
I used Brooks Baseball to look at his fastball zone charts against RHH for this year, and June-Sept of last year.
As would be expected, the issue is one of location. He loves to feed RHH fastballs on the outside part of the plate. However, this year he has missed outside 7% more than last June-Sept, especially in the middle-outside spot. He is pitching up and away more than last year too. But the main takeaway is that he has not located his fastball in the bottom part of the zone at all. Last year, we saw 13.24% of his fastballs hit the bottom third of the zone, most of them middle-away. This year, just 1 fastball has hit the lower outside portion of the zone. When you throw 89mph fastballs up in the zone, you are going to hit hard.
The same goes for the splitter. His troubles in keeping it down this year are really surprising. Again, when splitters elevate, they flatten and are easier to hit. Here are this splitter zone charts against RHH from last year June-Sept and all of this year:
Last year he fed RHH a steady diet of splitters down and away, sparingly coming inside. This year, he has only throw 4 splitters down and away to RHH and has come inside a ton. Over half his splitters thrown have been inside. Even more troubling is the most are inside and up in the zone. Its unclear whether he is just missing inside, or if this is an approach change by Koji and staff. What we do know is that it is not working.
I don't think its time to worry about Koji at all. He is still pitching well, but the biggest reason he isn't replicating last year dominance is that he isn't locating his pitches as well. If he can start placing it in the lower part of the zone, and using his splitter outside more, then I expect the ground balls to increase and the line drives to decrease.
Expecting Koji to post those numbers again is both unfair and unrealistic. For one, those 53.1 innings were possibly the greatest stretch by a reliever in baseball history. And for two, Koji is 39 years old and coming off of the most innings pitched in his MLB career by far. He used to be a starter in Japan, but hasn't' started since 2009 when he first joined Baltimore. Even then he only pitched 66 innings over 12 starts. However, measuring this year compared to most of last year can be instructive. For one, its what Koji aspires to replicate this year. Plus, we can tell what he isn't as successful doing this year compared to last, since the distinction will be brighter. And who knows, maybe he'll come close to it and we can all be happy. Even a slightly worse Koji will still be a very valuable asset.
On the surface in 2014 we have seen the tale of 2 Koji's. For the first 5 games of the year he was just as automatic as before, pitching 5 innings, allowing 3 hits, no runs, no walks, and striking out 7. Hitters were only hitting .158/.158/.211 with a .250 BABIP. Then he was shut down for a week with shoulder tightness. Since returning from his week off, Koji has been a different pitcher: 10 games (9.2 IP), with 10 hits allowed, 2 earned runs (1.86 ERA), 3 walks, and 15 strikeouts. What has changed is that batters are hitting .278/.333/.444, with a BABIP of .421, off of him in that 10 game stretch. The BABIP number really stands out to me as it is way above his career norm:
Year | BABIP |
2010 | 0.296 |
2011 | 0.197 |
2012 | 0.203 |
2013 | 0.189 |
2014 | 0.355 |
Outside of 2010, his BABIP has hovered around .200. This is interesting for a couple of reasons. The first is that pitchers generally have little control over their BABIP. They rely on defense to make plays, and a bad defense can increase the BABIP. The second is that BABIP is usually not as consistent from year to year, given the just mentioned defensive fluctuations. Koji has kept a consistent BABIP over 4 seasons, playing for 3 different teams. Even if we expect some regression, its fair to say part of Koji's struggles since coming back have been bad luck related. Koji said last week (link) that this is an April/May thing and that last year he didn't pitch that well during the same period. I took a quick look at his stats from last April/May and found that he was partly correct. He still pitched very well for relief pitching standards, but nowhere near what he did from June to September: 22 games (21 IP) allowing 16 hits, 5 earned runs (2.14 ERA), 4 walks, and 29 strikeouts. Batters hit .205/.244/.423 with a .267 BABIP. It is certainly possible that he is a better warm weather pitcher than cool weather pitcher. We know that his second half stats are vastly better than his first half stats for his career. His K/BB ratio is six times better in the second half than first half, and his ERA drops by a run.
However, its still important to look at that .421 BABIP and see whether this is just typical early struggles, or a result of age/injury-related decline in stuff. I took a look at his batted ball profiles for 4 different periods: April/May of 2013, June-Sept of 2013, 2014 pre-shoulder injury, and 2014 post shoulder injury. As always, the 2014 stuff is subject to SSS bias, but if any trend sticks out we should be able to see it. The batted ball data is from Baseball-Reference, which was the only place to aggregate intra-month data. Fangraphs has his GB% as lower, and LD% about 10% lower. I cross-checked with pitch F/X data from BrooksBaseball and both his fastball and splitter LD/BIP percentages were in the mid 20s, so I kept the B-Ref data.
[tablegrid= Koji Batted Ball Statistics ]
A couple of things stand out. The first is that the main difference between April/May 2013 and June-Sept 2013 was the reduction in fly balls and the increase in ground balls. His line drive and pop-up rate stayed consistent between the two period, but he traded fly outs for ground outs. Since ground balls are more likely to go for hits than fly balls (ex line drives), we can assume he had weak contact on the ground balls, and a solid infield defense behind him, or both.
What is driving that high BABIP this season is likely the large increase in line drive percentage, since line drives fall for hits about 70% of the time. Since coming back from his week off for the sore shoulder, his strike percentage has decreased, especially the strike looking percentage. Before his injury he was getting lucky that the spike in LD% wasn't resulting in more hits. His walks have increased as well, which is expected when strike percentage decreases, but he still isn't walking the park.
This probably raises the question of why the LD% has spiked all season (before and after injury). Koji's main issue this year, compared to June-Sept of last year is his performance against RHH. Righties have feasted on both his fastball and splitter this year:
[tablegrid= Koji Batted Ball vs RHH ]
Although he hasn't succeeded with both pitches, his fastball seems to stand out as the worse pitch against RHH. It is an easier pitch to command than the split, and we can see that hasn't thrown it for as many strikes this year. The whiff rate is the same, and the amount of times it is put in play has slightly up ticked, but the issue is what happens when it is put in play. Compared to most of last year, the ground ball rate has plummeted and the line drives have spiked, leading to right-handed batters slugging .714 against his fastball. The usage of the splitter has completely changed this year. As you'll see in the charts below, he has switched the side of the plate he is working RHH with the splitter.
I used Brooks Baseball to look at his fastball zone charts against RHH for this year, and June-Sept of last year.
As would be expected, the issue is one of location. He loves to feed RHH fastballs on the outside part of the plate. However, this year he has missed outside 7% more than last June-Sept, especially in the middle-outside spot. He is pitching up and away more than last year too. But the main takeaway is that he has not located his fastball in the bottom part of the zone at all. Last year, we saw 13.24% of his fastballs hit the bottom third of the zone, most of them middle-away. This year, just 1 fastball has hit the lower outside portion of the zone. When you throw 89mph fastballs up in the zone, you are going to hit hard.
The same goes for the splitter. His troubles in keeping it down this year are really surprising. Again, when splitters elevate, they flatten and are easier to hit. Here are this splitter zone charts against RHH from last year June-Sept and all of this year:
Last year he fed RHH a steady diet of splitters down and away, sparingly coming inside. This year, he has only throw 4 splitters down and away to RHH and has come inside a ton. Over half his splitters thrown have been inside. Even more troubling is the most are inside and up in the zone. Its unclear whether he is just missing inside, or if this is an approach change by Koji and staff. What we do know is that it is not working.
I don't think its time to worry about Koji at all. He is still pitching well, but the biggest reason he isn't replicating last year dominance is that he isn't locating his pitches as well. If he can start placing it in the lower part of the zone, and using his splitter outside more, then I expect the ground balls to increase and the line drives to decrease.