Hamels’ consequences
by Harry PavlidisJuly 07, 2010
This is part two of a look at Cole Hamels. Part one took a look at Hamels' pitch mix and is overall performance. In short, we compared Hamels using only three pitches to Hamels using four (or more) and found his best performance came during the three-pitch periods. We also wondered if his newest pitch—a cut fastball—was already on the way out.
In his last game, which happened before the first article went to press (sic), Hamels threw 14 percent cutters, no sinkers, 16 percent change-ups, 8 percent curveballs and 62 percent fastballs. So the cutter is back two games in a row after taking nearly two full games off. The two-seam fastball remained on the shelf. Looks like we can stick with the same four stretches of play for Hamels:
Pre-Sinker: 4/2/2008 - 9/7/2008
Sinker's Peak: 9/13/2008 - 8/15/2009
Fewer Sinkers: 8/21/2009 - 10/31/2009
With Cutters: 2010
Let's try and answer the questions that came up last time.
1. Was the new pitch effective?
This one is pretty easy. Hamels' sinker is about one run per nine innings worse than the average sinker. So "no", it was not effective. His cutter has been hit hard, but if we throw away Hamels' actual batted-ball outcomes and use league average outcomes for each type (line drive etc.), the cutter appears to be above average. It's too early to tell, but there is cause for both optimism and concern. We'll get more into the nitty gritty of effectiveness in the fourth and final question.
2. Were the original three less effective with or without it?
Here the results are mixed, and really vary based on how you measure effectiveness. If you take batted-ball control out of the equation (rv100e) you get one answer, leave them in (rv100a) and you get another. As with the cutter above in question one, it's hard to draw a firm conclusion.
Hamels' bread-and-butter pitch is certainly his four-seam fastball. It's actually done a bit better when paired with the cutter, but rv100e improved when paired with the sinker while rv100a ballooned. His best pitch is his change-up, and it was basically flat in rv100e when Hamels added a sinker or cutter to his mix. Unfortunately, expected outcomes did not mirror reality, as change-up rv100a spiked during both four-pitch phases. The curveball's relationship with the sinker mirrored the fastballs, an improved rv100e but an increase in rv100a. The curve, paired with the cutter, is as bad as it has been.
If you're thirsty for numbers, question four will bring a quencher of an answer. Three's no slouch.
3. Did he stop or reduce use of an original-three pitch in certain situations (counts)?
Things certainly changed by situation when Hamels changed his pitch mix. Our baseline is the pre-sinker era of 2008. Each section going down the left side is compared to the baseline, as shown in the right column.
| pre-sinkers | # | Fastball | Change-up | Curveball | |||||||
| first | 816 | 55% | 27% | 18% | |||||||
| ahead | 951 | 56% | 27% | 17% | |||||||
| even | 516 | 49% | 36% | 15% | |||||||
| behind | 619 | 59% | 37% | 4% | |||||||
| full | 115 | 47% | 50% | 3% | |||||||
| 3017 | 55% | 32% | 14% |
There's your baseline. The empty columns to the right will be filled below. Red numbers indicate a decline from baseline, and the value is relative to the baseline (so a .091 is a 9.1 percent increase relative to baseline; you can see the raw values on the left).
| sinker's peak | # | Fastball | Change-up | Curveball | Sinker | sinker's peak | Fastball | Change-up | Curveball | ||
| first | 646 | 54% | 27% | 11% | 8% | first | (0.012) | (0.010) | (0.373) | ||
| ahead | 751 | 46% | 26% | 14% | 14% | ahead | (0.179) | (0.032) | (0.184) | ||
| even | 450 | 47% | 30% | 15% | 8% | even | (0.036) | (0.168) | (0.027) | ||
| behind | 539 | 53% | 33% | 5% | 9% | behind | (0.103) | (0.101) | 0.106 | ||
| full | 125 | 44% | 35% | 6% | 14% | full | (0.063) | (0.302) | 1.452 | ||
| 2511 | 50% | 29% | 11% | 10% | (0.092) | (0.080) | (0.198) |
All pitches, in all situations, were used less during this period with the notable exception of the curveball. Overall, the curveball took the biggest hit by sharing time with the sinker. Full-count change-ups were also dramatically impacted.
Now Hamels all but drops the sinker, but things don't all go back to where they were before.
| fewer sinkers | # | Fastball | Change-up | Curveball | Sinker | fewer sinkers | Fastball | Change-up | Curveball | ||
| first | 241 | 57% | 32% | 10% | 1% | first | 0.038 | 0.190 | (0.455) | ||
| ahead | 288 | 58% | 30% | 13% | 0% | ahead | 0.036 | 0.084 | (0.271) | ||
| even | 166 | 48% | 39% | 11% | 1% | even | (0.009) | 0.086 | (0.252) | ||
| behind | 174 | 51% | 43% | 5% | 2% | behind | (0.135) | 0.165 | 0.055 | ||
| full | 30 | 40% | 60% | 0% | 0% | full | (0.148) | 0.190 | (1.000) | ||
| 899 | 54% | 36% | 10% | 1% | (0.014) | 0.129 | (0.304) |
The change-up came back with a vengeance, and the curveball's decline accelerated.
Now look what happens with the cutter in the mix.
| with cutter | # | Fastball | Change-up | Curveball | Sinker | Cutter | with cutter | Fastball | Change-up | Curveball | |
| first | 415 | 59% | 19% | 9% | 2% | 11% | first | 0.071 | (0.291) | (0.477) | |
| ahead | 458 | 48% | 21% | 12% | 1% | 17% | ahead | (0.140) | (0.222) | (0.274) | |
| even | 315 | 52% | 25% | 7% | 2% | 15% | even | 0.070 | (0.304) | (0.564) | |
| behind | 346 | 64% | 25% | 1% | 1% | 8% | behind | 0.090 | (0.319) | (0.667) | |
| full | 77 | 56% | 31% | 0% | 1% | 12% | full | 0.189 | (0.382) | (1.000) | |
| 1611 | 55% | 23% | 8% | 1% | 13% | 0.015 | (0.279) | (0.454) |
Some return of fastballs, but the change-up and curveball have been sent to the back of the queue to a greater degree than we saw with the sinker's full-time presence.
4. Of the original three, has the fastball been impacted less than the curveball and change-up by the absence or presence of a new pitch?
Yes. Here are some key metrics grouped by season and by "three" or "more", in reference to the "peak sinker and "with cutter" periods.
| Type | Segment | # | rvERAa | rvERAe | Swing | Whiff | B:CS | IWZ | Chase | GB% | HR/FL% |
| Change-up | three | 1271 | 0.31 | 0.92 | 0.603 | 0.416 | 1.7 | 0.597 | 0.441 | 49% | 5.5% |
| Change-up | more | 1092 | 3.14 | 1.12 | 0.580 | 0.414 | 1.5 | 0.584 | 0.432 | 48% | 14.9% |
The change-up is outstanding, but seems to fly over the wall more when there's a fourth pitch in the mix. That's not something pitchers have control over, or is that too much of a coincidence?
| Type | Year | # | rvERAa | rvERAe | Swing | Whiff | B:CS | IWZ | Chase | GB% | HR/FL% |
| Change-up | 2008 | 1065 | 0.68 | 1.22 | 0.593 | 0.388 | 1.8 | 0.601 | 0.416 | 48% | 6.1% |
| Change-up | 2009 | 934 | 2.00 | 0.79 | 0.595 | 0.426 | 1.3 | 0.608 | 0.451 | 49% | 11.0% |
| Change-up | 2010 | 364 | 3.40 | 0.99 | 0.582 | 0.467 | 2.4 | 0.519 | 0.457 | 50% | 20.0% |
Over time, the change-up is seeing less of the zone and taking more trips over the fence.
Let's look at the other victim of change, the curveball.
| Type | Segment | # | rvERAa | rvERAe | Swing | Whiff | B:CS | IWZ | Chase | GB% | HR/FL% |
| Curveball | three | 501 | 4.34 | 6.15 | 0.329 | 0.194 | 2.6 | 0.419 | 0.203 | 37% | 2.6% |
| Curveball | more | 398 | 6.59 | 5.88 | 0.276 | 0.200 | 3.2 | 0.387 | 0.139 | 54% | 10.5% |
Now that's a strange one. More ground balls (yea!) and more gopher balls (boo!). Not really a great pitch however you cut it.
| Type | Year | # | rvERAa | rvERAe | Swing | Whiff | B:CS | IWZ | Chase | GB% | HR/FL% |
| Curveball | 2008 | 452 | 4.52 | 6.07 | 0.336 | 0.211 | 2.6 | 0.427 | 0.205 | 35% | 5.4% |
| Curveball | 2009 | 326 | 5.88 | 5.66 | 0.273 | 0.191 | 3.4 | 0.374 | 0.147 | 57% | 0.0% |
| Curveball | 2010 | 121 | 6.90 | 6.91 | 0.281 | 0.147 | 2.8 | 0.405 | 0.139 | 46% | 14.3% |
A bit of a yo-you ride for the ground ball rate. As you can see, the curveball is a bit of a mixed bag. A bag of not-so-good-stuff, but too mixed to say what happens when it has a new friend.
Last, but not least, the fastball. For whatever reason, changes in repertoire aren't associated with big swings in fastball effectiveness. There's still reason to believe the fastball is better in 2010, whether or not the cutter is to thank.
| Type | Segment | # | rvERAa | rvERAe | Swing | Whiff | B:CS | IWZ | Chase | GB% | HR/FL% |
| Fastball | three | 2135 | 3.34 | 4.14 | 0.478 | 0.148 | 1.7 | 0.538 | 0.318 | 32% | 9.0% |
| Fastball | more | 2130 | 4.19 | 3.87 | 0.454 | 0.161 | 1.7 | 0.570 | 0.269 | 36% | 9.0% |
If anything, as noted above, the fastball doesn't like the sinker but likes the cutter. The latter being more clear, or less muddy as it were.
| Type | Year | # | rvERAa | rvERAe | Swing | Whiff | B:CS | IWZ | Chase | GB% | HR/FL% |
| Fastball | 2008 | 1805 | 3.27 | 3.94 | 0.484 | 0.141 | 1.8 | 0.537 | 0.326 | 35% | 10.4% |
| Fastball | 2009 | 1577 | 4.64 | 4.04 | 0.465 | 0.161 | 1.8 | 0.544 | 0.291 | 32% | 8.9% |
| Fastball | 2010 | 883 | 3.22 | 4.09 | 0.430 | 0.174 | 1.4 | 0.607 | 0.228 | 35% | 6.2% |
What have we learned?
That things are complicated. Other than that, we can say a few things about Hamels.
- An increased rate of flies and liners turning into home runs is observed when Hamels adds a fourth pitch heavily to the mix. Coincidence?
- While the fastball and sinker may or may not have been compatible, the cutter and the fastball may be
- The jury is still out on the cutter, but the sinker is welcome to stay away
- His change-up is very good, it's a shame it's used less and less
- His curveball I can take or leave. As a commentor noted in part one, it's more of a show-me pitch than anything
My vote: keep working on the cutter, but have some more faith in the change-up.
References and Resources
PITCHf/x data from MLBAM and Sportvision. Batted ball data from MLBAM. Pitch classifications by the author.
Harry Pavlidis admits he has a baseball problem. He is the founder of Pitch Info LLC, His pitch classifications power the player cards at Brooksbaseball.net. Feedback, questions and comments are appreciated - Email harrypav@gmail.com and Twitter @harrypav