Every Pitcher Should Throw a Cutter

Well, at least it seems that way, given all the anecdotal evidence I've seen over the past year.  Brian Bannister went from ineffective stat-head starter to league-average pitcher by adding a cutter (which he's throwing over 50% of the time).  Ryan Franklin ditched his slider in favor of a cutter, and is posting the lowest FIP of his career (thanks to a super-sized ability to keep the ball in the park.)  And Mariano Rivera, well, he's been the best relief pitcher of all time with "only one pitch".

I asked Harry Pavlidis to help move us beyond the anecdotal stage of cutter analysis and he threw back a ton of data my way.  Short story: it's a great pitch, but thrown only 3% of the time in the majors.  The longer story follows below.  To start here are some descriptive metrics for five popular pitches across all counts.

Pitch classifications are Harry's when available, the default Gameday classifications when not.
% - Percentage of all pitches thrown.
Zone - Percentage of pitches in called (not rulebook) strike zone.
Swing - Percentage of all pitches swung at.
Whiff - Percentage of whiffs per swing.
BABIP - Batting average on balls in play (not including HRs.)
HR CON - Percentage of contacted balls (including fouls) that are home runs.

Fastballs are thrown a lot.  One reason is that they're thrown for strikes more often than any other pitch, leading to fewer walks.  So why throw a non-fastball?  Because curve balls, sliders, and off-speed pitches generate more whiffs and aren't hit as hard.  This data backs up the assumption that fastball should be thrown when balls are nearly as hurtful as a hard-hit ball, while other pitches are more useful when giving up a hard-hit ball is a big loss compared to simply wasting a pitch.

The cutter, however, comes with the best of both worlds.  It's thrown in the zone as often as pure fastballs, but generates more whiffs (although not as many as the other three pitches), and isn't hit as hard.  It appears, then, that it's a jack of all trades pitch, able to be pumped over the plate repeatedly without strong repercussions.

0-0 Counts

Instead of looking at pitches overall, we can break down results by count.  Here are the same metrics, but only for pitches thrown in 0-0 counts:

On 0-0 counts, fastballs are thrown a bit more often than overall, while the other pitches are thrown a bit less often.  Curve balls are a bad choice to begin an at-bat, ignoring what happens after the first pitch, as they're chased only one-third as often as overall and yield home runs 71% more often.  They do get thrown a bit (10%) less often, but that number should probably be even lower, given the results of a 0-0 curve ball.

Most other pitches are a bit less effective on 0-0 counts than they are overall (while being swung at about 60% as often), mostly due to higher home run rates.  The exception is the cutter, which shows no significant changes in a 0-0 count.  It and the slider appear to be the best pitches to start an at-bat, with a high percentage thrown in the zone and the best BABIP and HRCON rates.  Strikes that aren't hit hard?  Yes, please.

0-2 Counts

In general, pitches thrown in 0-2 counts are swung at 10% more often, chased out of zone 25% more often, thrown inside the strike zone 40% less often, produce BABIPS 10% lower, and turn into home runs 40% less often when contacted.  As you might guess, curve balls and sliders are thrown significantly more often in 0-2 counts.  Fastballs actually see the biggest rise in whiff rate, although it's the off-speed pitch which appears to be the most effective -- hitters love to swing at it, but don't do much with it even when they make contact.

But here's a surprise: 0-2 cutters are significantly less effective than other 0-2 pitches.  While they're still thrown nearly as often, BABIP for 0-2 cutters jumps 10% to .321 and they're more likely to be hit for a home run than any other pitch.  All that even though they're swung on and missed just as much as any other pitch inside the strike zone.  Moral of the story?  Go off-speed on 0-2 pitches and don't throw your cutter.

2-0 Counts

Let's jump straight to the cutter analysis, because it's crazy.  2-0 cutters are thrown in the zone 65% of the time, much higher than any other pitch, they're whiffed at less than any pitch besides fastballs, yet become home runs about half as often as any other pitch.  Sliders are also very effective at preventing 2-0 home runs, but get pounded all over the field for in-play hits (.343 BABIP) while cutters don't (.273 BABIP).

One stat not in the chart is extremely telling: when balls are contacted (either in play or for home runs) every non-cutter pitch sees the slugging percentage of those contacted ball go up 22% to 34% in 2-0 counts, while cutters see no increase whatsoever.  Other pitches have a SLGCON between .610 and .687, while cutters sit at .496.

In other words, in 2-0 counts, a situation which favors the hitter to a large degree, throwing a cutter appears to neutralize the hitter's advantage.  The results of cutters in 2-0 counts are similar to the results in other counts, or even a bit better.

If you're a pitcher forced to throw a 2-0 pitch, you'd be quite happy with a .270 BABIP, would you not?  Learn a cutter.

Moving Forward

There are obviously a number of issues with the numbers presented and potential conclusions drawn from them.  One, Gameday pitch classifications are in their infancy and while Harry's are better, they aren't perfect and aren't available for all pitches.  Two, since cutters are a rare pitch, they likely would lose some effectiveness if thrown more often by the same pitcher or if thrown by more pitchers.  Three, there are certainly some pitch sequencing effects going on, as Josh Kalk discovered.  In fact, it appears that throwing a cutter hurts the effectiveness of the next pitch if it's a fastball or curve ball (while helping a subsequent cutter or change-up).  And four, perhaps most of the pitchers who throw cutters are the good pitchers.  There's definitely a need for further analysis.

On top of noticing that the cutter is extremely effective, I'd love to figure out why.  Cutters are the straightest pitch thrown, similar to a slider, but don't break much at all.  Straight sounds like a bad idea, but because of a pitcher's release point and target, means the pitch is actually traveling across the hitter's field of vision instead of tailing in directly towards the catcher's glove at the end, like with a fastball.  Is a cutter easier to control than a slider for some reason?  Is it more difficult to tell the difference between a fastball and a cutter than a fastball and a slider?  Any ideas there?

And from a scouting point of view, is there any reason certain pitchers can't learn to throw a decent cutter?  They appear to be easily controllable, like a fastball, but carry the power-neutralizing effects of pitches with more movement.  And, perhaps most importantly, cutters can be thrown against hitters of the opposite handedness, negating platoon advantages.  Brian Bannister, care to share your thoughts?