July 7, 2010
Checking the Numbers
Weaver's Soaring Strikeout Rate
When Jered Weaver made his major-league debut on May 27, 2006, the Angels were 20-28, in last place in the American League West, and five games behind the division-leading Rangers. After the mega-prospect blanked the Orioles over seven strong innings to the tune of a 75 game score, Angels fans were more than enthused that their rotation had been vastly improved by his addition. Weaver would finish the season with a 2.56 ERA in 123 innings with an impressive 3.18 K/BB ratio and 1.03 WHIP. He walked few, proved stingy with allowing hits, and recorded his fair share of strikeouts. Though his rookie numbers were impressive, many would agree that Weaver’s lack of progress since then has been disheartening.
After seeing the lanky righty post a 1.62 ERA, 10.1 K/BB, and 13.3 K/9 in his final season at Long Beach State in 2004, the Angels drafted Weaver in the first round envisioning him becoming an ace, but he essentially stagnated over his four seasons in the majors. This is not to say he was a poor pitcher, but rather that the semi-Strasburgian hype caused his 3.73 ERA, 7.3 K/9, and 2.76 K/BB to look less effective than it had actually been. Simply put, Weaver had become a very reliable No. 2 but had shown no signs of developing into the No. 1 the Angels needed.
After ace John Lackey left for greener pastures this last winter, Weaver became the de facto No. 1 and, boy, has he lived up to that billing. Though he was somehow originally left off of the AL’s All-Star roster (before being added Tuesday night) for a game that will take place in his home park, Weaver has been magnificent in his 17 starts this season, compiling a 2.82 ERA and 4.77 K/BB ratio in 108 2/3 innings. Most impressive, however, has been his remarkable surge in the K/9 department. After averaging 7.3 punchouts per nine from 2006-09 with little deviation in each season, Weaver currently leads the major leagues with both a 10.3 rate and a raw tally of 124 strikeouts.
When surges like this occur, the first question to ask is whether the marks are stable. Well, thanks to research from the immortal Russell Carleton, we know that the K/9 rate stabilizes at around 150-200 batters faced—as in, we can safely determine that the number is not the result of a fluke or a small sample after it reaches this threshold. Weaver has now faced 437 hitters, so it is more than safe to assume that his uber-improved rate is for real. Additionally, unlike certain National League pitchers who, as I have shown in the past, may experience an uptick in their rate for striking out a higher percentage of opposing pitchers, the American League uses the designated hitter, so the improvement is free of getting some easy outs in the No. 9 spot of the batting order.
With that out of the way, what is incredibly interesting is that Weaver’s strikeout rate has not just increased a smidgeon, like say from 7.3 to 8.2 or even to 8.8; it has soared all the way to 10.3, an improvement of three strikeouts per nine innings, a delta that screams rarity. Before asking how, though, I consider it vital to see if any comparable improvements have occurred over the last half-century, and if so, were the spikes sustained? While the answer is irrelevant from the standpoint of what is actually happening now, it is of great importance for projecting Weaver from here on out—it would not be prudent to expect him to continue to strike batters out at this rate if there is a decent handful of comparable pitchers who all saw their strikeout rates plummet or revert to previously established norms after the uptick.
The first step in this line of inquiry is to see, quite simply, if other pitchers have experienced strikeout rate jumps as substantial as what we are seeing from Weaver. To that end, I pooled together all three-year spans beginning between 1954-2007—so that the third year of the final such span was last season—and restricted the list to pitchers amassing at least 100 innings in each of the seasons. I, then, computed the K/9 for these pitchers in each of the three seasons, as well as the weighted average of the three seasons combined. The next step involved comparing the K/9 in the fourth year of the span—if the span was Weaver’s 2007-09 seasons, of interest is his weighted average over those three years compared to his rate in 2010, the fourth year.
Lastly, in order to keep the results in tune with Weaver’s attributes, I further restricted the sample to include only those pitchers with three-year average K/9 rates in the 5.5-7.5 range. After applying all of these filters, Weaver’s improvement from 7.3 K/9 to 10.3 K/9 marks the fourth largest jump in a four-year span since 1954. The table below shows the top such spikes:
What first struck me upon reviewing this table was that the group represented a motley crew of pitchers, a few of which improved their mark based on a tangible characteristic. For instance, Loaiza famously improved in that 2003 season due to learning how to throw a cutter under the tutelage of White Sox pitching coach Don Cooper. Additionally, those above and on the rest of the list were fairly comparable with Weaver in terms of age. Sure, there were oddballs like Miller who somehow improved later in their careers, but for the most part, everyone improved right around their prime. Our first question can now be answered: Yes, other pitchers have experienced a similarly sharp fluctuation, so it isn’t as if Weaver’s newly discovered skill represents an unprecedented improvement. But the next question is what to do with this data?
Of the pitchers in the sample that met the aforementioned criteria, I calculated their K/9 in the hypothetical fifth, sixth, and seventh years of the span. It should be noted that this precluded several pitchers from advancing further in the sample size gauntlet as some petered out after experiencing a large spike towards the end of their careers. The point here would be to conduct a pre- and post-test comparing the three years before the spike to the three years after the spike. If the pitchers deemed comparable to Weaver in these regards managed to sustain their strikeout rates, relatively-speaking, it could be very telling of his potential success moving forward, though some regression is inevitable and declines are expected for something as simple as aging.
All told, there were a whopping 21 pitchers with three years on both sides of the spike that also met the criteria of having a pre-average between 5.5 and 7.5, with 100 or more innings in the three years leading up to the spike, and who also saw their K/9 jump by at least 1.75 strikeouts. Comparing the post-three year average to the spike year, the following tidbits emerge:
Though the sample is certainly minuscule, the results suggest that pitchers who see their K/9 spike by a substantial amount are due for a decline in the coming seasons. However, the range of these declines is so wide that it is still tough to get a grasp on what to expect for Weaver even if we trust the results of a small sample. Overall, the 21 pitchers averaged a 6.24 K/9 in the three years leading up to the spike, an 8.65 K/9 in the year of the spike, and a 7.34 K/9 in the three years following. In other words, the declines were in no way as vast as the spikes, so if we figure a worst-case scenario for Weaver he would still be expected to be in the 8.7-9.2 K/9 range over the next couple of years, assuming that his improvement is more than random.
That segues into our next question: How is he improving his rate this much? Perhaps the help of our friend PITCHf/x will help us solve this conundrum. For starters, the table below breaks down Weaver’s pitch distribution for the key ingredients of his repertoire over the last four seasons:
Now, I would normally write information like this off on account of the fact that PITCHf/x did not segregate two-seam and four-seam fastballs up until 2009, a process still being refined this season. However, the data above and below was provided by Harry Pavlidis, who has his own classification method that can be applied all the way back to the inception of the data; without the reclassification of several fastballs, Weaver would appear to have thrown no two-seamers prior to last season. Additionally, we have clinical evidence suggesting that Weaver himself did not begin throwing the two-seamer until the 2008 postseason. He worked on the pitch leading into the 2009 season and continually attempted to refine the pitch during spring training this year. The increase in usage is very much legitimate and likely a key component of his improved rate.
The increased usage has also manifested itself in his pitch distribution with two strikes, which has helped to not only increase his rate of striking out hitters but also in keeping the ball on the ground, a characteristic never associated with Weaver in the past. The table below shows his pitch breakdowns on two-strike counts over the past four years:
As expected, his reliance on the two-seamer in two-strike counts has grown significantly, but he hasn’t just pared back the four-seam heaters to make up the difference in these situations. Instead, he is throwing fewer changeups, which makes sense from the standpoint that a solid two-seam fastball provides enough of a change of pace to the hitter with its movement components compounded with decreased velocity to keep him off-guard. One of my goals with the PITCHf/x data, when we have many more years in the sample bank, is to trace changes in approach like this to see, in addition to the results-based outputs, if the inputs leading to those outcomes are sustained.
What is Weaver doing? Well, he is throwing a new pitch much more frequently than in the past that he has consistently fine-tuned to the point that it is now a very effective weapon. While a new pitch for some pitchers may result in an increase to their rate of grounders, Weaver’s toy is giving batters ample trouble making any contact. The aforementioned Carleton once found in another piece that pitchers held a seven-point advantage in the OBP department over hitters they had never faced before, which gradually lessened the more frequently the two faced one another.
Weaver has essentially become a new pitcher to many of his familiar foes due to the increased usage of a new and effective pitch. Not many pitchers have historically experienced such a large spike in their strikeout rate, but aging effects aside, the group showed an average strikeout rate greater after the spike than in the years leading up to it, even if the rate in the year of the spike was not achieved again. If Weaver can continue to rely on his two-seamer—and this is assuming it remains effective—then we should not automatically expect a substantial regression in his rate in the coming years. He might not ever strike out almost 11 batters per nine innings again, but the much higher number most likely represents a change in future outcomes due to a tangible change in the inputs.
Special thanks to Harry Pavlidis and Russell Carleton for assistance and sanity checks.