Reading a Pitcher's Arsenal: Why Pitch Mix Beats K/9

Two pitchers, same 9.4 K/9. One faces a lineup that crushes fastballs. The other faces a lineup that whiffs on sliders 35% of the time. Their K props will cash at completely different rates, even though their season numbers look identical. K/9 is a useful starting point. Pitch mix is the actual story.

K/9 Is a Lagging Indicator

K/9 tells you what a pitcher has already done. It does not tell you what tonight's lineup is built to do against him. A pitcher who racked up 11 K/9 over six starts against contact-heavy lineups can run into a free-swinging team that hunts his fastball, and suddenly the same arsenal that struck out 11 per nine looks pedestrian.

Every K is the result of a specific pitch beating a specific batter. So the right unit of analysis is not the pitcher's season rate, it is each pitch in his arsenal versus the swings he is about to face.

What "Arsenal" Actually Means

An arsenal is the menu of pitches a starter throws, weighted by how often he throws each one. Most modern starters live on three to four pitches. The mix is rarely 25/25/25/25. It is closer to 50/30/15/5, which means the top pitch is doing roughly half of all the work.

The Three Numbers That Matter Per Pitch

For each pitch in an arsenal, three stats do nearly all the work:

• Usage %: how often the pitch gets thrown. A 50% usage pitch is twice as important to the matchup as a 25% usage pitch.
• Whiff %: swing-and-miss rate when the batter offers at it. This is the K-generation engine. Anything north of 30% is plus, north of 35% is elite.
• xBA against: expected batting average given the contact quality on that pitch. Below .200 is dominant. Above .280 is a problem.

You can find league-average pitcher numbers around 23% whiff and .250 xBA. Anything substantially better than that, on a pitch thrown half the time, is a real edge in K projection.

The Match Is the Point

Here is where it stops being about the pitcher and starts being about the matchup. PropPrizm tracks each batter's xBA and K rate against every pitch type. So when a starter goes, we can ask the only question that actually matters:

For each pitch he throws, how does this lineup perform against it?

If a starter throws 58% fastballs and tonight's lineup whiffs at 28% on fastballs, that one matchup contributes 16 percentage points of total swing-and-miss before you even get to the breaking ball. If he throws 32% sliders and the lineup whiffs at 41% on sliders, that adds another 13 points. The arsenal-weighted whiff rate against this specific lineup is 29%, which is well above his season average of 24%.

That is a K Over night. The season K/9 alone would never tell you that.

The Coverage Problem

One thing to flag honestly: pitch-type splits require sample. A batter with 200 PA against fastballs has a real number. A batter with 18 PA against splitters does not. PropPrizm's matchup card shows a coverage % for each batter-pitcher pair, which is the share of the pitcher's arsenal where we have a reliable batter sample.

• Coverage 90%+: the arsenal-match number is doing all the heavy lifting it can. Trust it.
• Coverage 70-89%: solid, but be aware there is a slice of the arsenal you are guessing on.
• Coverage under 70%: the matchup is partially based on small samples or missing splits. Lean on season totals more, arsenal less.

Coverage is usually highest for veteran batters facing veteran starters. It is lowest for September call-ups, rookies on either side, or pitchers debuting a brand-new pitch.

Why This Beats K/9 Specifically

K/9 collapses everything down to one number. That is convenient, but it discards three things that matter:

1. Pitch mix has changed. A pitcher who added a sweeper in March throws a meaningfully different pitcher than the K/9 line from his first three starts.
2. Lineups are not interchangeable. A 9.4 K/9 against the Astros is a different signal than 9.4 K/9 against the White Sox. Arsenal-vs-lineup catches this. Season K/9 cannot.
3. Park and weather effects do not move K/9 evenly. Cold-weather starts cut breaking-ball bite, which shifts effective arsenal usage, which K/9 does not adjust for.

None of this means K/9 is useless. It is a fine sanity check. It just should not be the primary input.

What to Look For on PropPrizm

On every starter card, the arsenal table shows usage %, whiff %, and xBA per pitch. On every batter row, the matchup column shows that batter's xBA and K rate against the pitcher's specific top pitches. The two views are designed to be read together.

Three patterns worth watching for:

1. Top Two Pitches Both Plus

If a starter's #1 and #2 pitches both run above 30% whiff, and the lineup he faces has elevated whiff rates against those exact pitch types, the K Over is structural. This is where the largest edges live.

2. Heavy Fastball, Bad Lineup Match

A 60%+ fastball-usage starter facing a lineup with sub-20% whiff on fastballs is a classic K Under setup. The market often underprices this because the pitcher's overall numbers look fine. The lineup specifically neutralizes his primary weapon.

3. New Pitch, Small Sample

If a starter recently added a pitch (sweeper, splitter), his season usage % understates how often he is currently throwing it. Recent-form arsenal data catches this. If the new pitch is generating 40%+ whiff in a small sample and the matchup card is weighting it correctly, that is a leading indicator the market is slow to price.

The Mental Shortcut

When you are sizing up a K prop, ask three questions before anything else:

1. What are the top two pitches and how often does he throw them?
2. How does this lineup perform against those specific pitches?
3. Is coverage high enough to trust the answer to question 2?

If the answers line up (top pitches are plus, lineup struggles against them, coverage is solid), you have a real K Over edge. If they do not, no season K/9 number is going to bail you out.

Try it yourself: open any starter on the Matchup Dashboard, scroll to the arsenal table, and check the top batter rows for xBA on the pitcher's #1 and #2 pitches. Once you start reading matchups this way, the season-stat view starts to feel like driving with one eye closed.