Wacha Went Eight and We Watched From the Sideline

Fourteen games on a Friday and not one of them handed us something we could genuinely stake money on. That's the whole story on the betting side — the model looked at every side, every total across the card, and nothing cleared the bar. No moneyline cleared five percentage points of edge over the market. No total cleared ten points of over/under edge. The record stays clean because there was no record to make: 0-0-0 on moneylines, 0-0-0 on totals, no units staked, no units lost.

Some people see a no-bet day and assume the analyst chickened out. That's not what happened. The whole discipline is this: when the market has priced something correctly, or close enough that the gap doesn't justify the risk, you close the laptop. Friday was one of those days. The baseball was real. The edges weren't.

Nothing made the card on Friday. No money was down. That is not a hedge or a soft landing — it is the position, stated plainly, and it is the correct one given what the model returned.

There were no graded leans on moneylines or totals to report either. The board was quiet. A no-bet slate is a real, on-brand outcome. We will take it over forcing action into a market that didn't give us room to work.

One thing worth saying out loud: the discipline to sit on your hands in a 14-game Friday environment, when there is social pressure to have a card, is exactly what protects the unit count over a long season. We are not in the business of manufacturing conviction. When the data says pass, we pass.

The DFS slate was a different conversation entirely, and Friday's pitching performances were something. Let's go through what actually happened.

Top Scorers

• Michael Wacha, Kansas City Royals — 39.6 DK points. Eight innings, ten strikeouts, zero earned runs. Our projection had him at 14.69. That gap is enormous and we will get to it, but first just appreciate the line: eight innings of shutout ball with double-digit strikeouts from a 34-year-old. That is a vintage outing from a guy who has quietly been one of the more reliable arms in the American League for years.
• Hunter Brown, Houston Astros — 33.4 DK points. 6.2 innings, ten strikeouts, one earned run. Our projection was 16.07. Brown missed bats at an elite rate and gave the Astros a genuine quality start. He was a $9,400 salary and returned more than double what we projected.
• Logan Henderson, Milwaukee Brewers — 31.1 DK points. Seven innings, seven strikeouts, one earned run. Our projection was 18.98, which was our closest call of the top three starters on the night. Henderson did not have salary data available in our system, which is a tracking note we will flag internally.
• George Springer, Toronto Blue Jays — 29 DK points. Four-for-four with a home run and an RBI. Our projection was 7.21 at a $4,200 salary. That is a massive beat — Springer went out and had one of those nights where everything he hit found grass or left the yard.
• Cam Schlittler, New York Yankees — 28.6 DK points. 5.2 innings, eight strikeouts, zero earned runs. Our projection was 18.62, which was the best-calibrated call we had among the top five scorers. At $10,800, Schlittler was the highest-priced arm on the board and he delivered.

Where We Hit

The model identified the right arms — Wacha, Brown, and Henderson were all flagged as projectable options. The direction was correct on all three. The magnitude of Wacha's and Brown's performances, however, outran what any reasonable pre-game projection was going to capture. Eight shutout innings with ten punchouts is a ceiling outcome. The model projected a solid Wacha start; he delivered an elite one. That is not a modeling failure — that is the nature of pitching variance. You identify the pitcher most likely to post a big number, and sometimes they post a historic one.

Where We Missed

Three names to own here, and we will own them without dressing it up.

• Dean Kremer, Minnesota Twins — projected 14.13, scored minus-one DK point. Four innings, three strikeouts, five earned runs. That is about as rough as a pitching line gets. When a starter allows five earned in four innings, you are negative on DraftKings before the lineup card is even dry. We had him as a projectable arm and he was not one on Friday.
• Casey Mize, San Diego Padres — projected 13.31, scored 4.5 DK points. Five innings, one strikeout, two earned runs. The strikeout miss is the part that stings here — one punchout in five frames is a strikeout rate that does not support a DFS investment regardless of the earned run total. We missed the stuff on Friday.
• Kohl Drake, Arizona Diamondbacks — projected 12.76, scored 3.4 DK points. Four innings, four strikeouts, five earned runs. Another five-earned-run outing. Friday was not kind to pitchers at the lower end of the projection range, and Drake was a clear example of why pitching variance is the most dangerous variable in DFS construction.

Of 206 projections matched to final scores across 413 players graded on Friday, the three names above represent the sharpest downside misses. The pitching floor risk was real and it bit us in three separate spots.

Two things, and they pull in opposite directions, which is what makes this game honest.

First: the no-bet discipline works. Fourteen games and nothing cleared the edge threshold. We did not manufacture a play to have something on the board. The unit count is intact. Over a 162-game season, nights like Friday are how you survive September with enough runway to act when the edges are actually there. No-bet slates are not failure. They are the model doing its job.

Second: the pitching floor in DFS is always real. Wacha and Brown were extraordinary Friday — ceiling outcomes that rewarded anyone who identified them as the top options. But Kremer at minus-one DK points is a reminder that the guy you think is a safe mid-range arm can blow up your lineup in four innings. The variance window on starting pitchers is wider than any other position in the game, and Friday demonstrated both ends of it in the same slate.

Going forward, the lesson on projection gaps — particularly on arms we had in the 12-to-15 point range — is worth reviewing in how we weight floor versus ceiling in construction. Wacha going 39.6 at a 14.69 projection is a win for target identification. Kremer going negative at a 14.13 projection is a loss for floor assessment. The methodology will look at both.

Saturday's slate is on the board. We will see what the model finds.