Skill Changes, August 6

16 pitchers with a real velocity change, 8 more whose stuff or command has moved, 8 hitters with a measurable skill shift.

The inputs, not the results. A pitcher who has lost velocity is a different pitcher before his ERA catches up, and a hitter whose contact quality has shifted is a different hitter before his batting average agrees.

Pitchers losing velocity

Year-over-year on the primary fastball, for pitchers with at least 100 of them in both seasons. Velocity is the earliest and least ambiguous signal an arm gives you: it moves for physical reasons and it stabilises faster than any result stat. A full mile per hour is roughly the tenth percentile of the league, real, not drift.

Pitchers gaining velocity

The same measure the other way. Added velocity on a starter often means a mechanical change or a role change; on a reliever it is usually the shorter outings doing the work.

Pitchers changing stuff and command

Stuff+, Location+ and Pitching+ measure the pitch itself rather than what the hitter did with it, so they move before the ERA and before the strikeout rate. This is a slower read than velocity and a broader one: it catches a new slider or a command gain that the radar gun never sees. Ordered by the size of the change, then by where the arm sits on our board.

Hitters

Contact quality and swing decisions, the parts of a hitting line that settle before the results do.

What this means

Bailey Ober is the one I would act on. He was already living at 90.3 on the four-seamer, which is the bottom of what a starter can carry, and the reading now says 88.2. Everyone else on the velocity-loss list is giving up speed from a cushion: Estrada is still throwing 95.8, Abreu 95.3, García 95.0. Ober has no cushion. His whole act is a steep downhill plane and a fastball that plays above its speed at the top of the zone, and the margin between "plays above its speed" and "gets hit a long way" is about two miles an hour. A forward rank of 298 says the projection has already marked him down some. I do not think it has marked him down enough.

The case against me is decent, and it is about timing. Velocity in these samples is not season-long truth. Starters run cold in April, run cold coming off a layoff, and a big man who has had shoulder trouble may simply be building up. If the 88.2 is a March-and-April number that becomes 89.8 in June, then rank 298 is roughly right and I have talked myself out of a mid-rotation arm for nothing. The falsifier is clean: check the fastball reading again after five or six starts of warm weather. If it is still under 89, the strikeout rate goes with it and 298 becomes optimistic. If it climbs back above 90, buy him from whoever read this table and panicked.

Contrast that with Camilo Doval, who added two full ticks to the sinker and now sits at 98.6. This will amount to nothing, or close to it. Doval's problem was never how hard he threw. Nothing in this data says his location improved, and he does not appear anywhere on the stuff-and-command list, which is where you would expect a genuine change to show up. Rank 678 for a reliever with that arm is a bet on role, not on the radar gun, and more velocity from a pitcher who already threw 96.6 does not by itself buy back a ninth inning. Adding him because of this row is chasing the least informative version of a good number.

The hitter section is where I would spend my waiver claim, with one caution the site is explicit about: the skill signal is display-only and is not baked into the rankings, so nothing here has been priced. Jordan Walker is the sharpest case. An xwOBA gain of 77 points with 119 points of xSLG attached is not a batting-average blip, it is contact quality moving, and Walker's entire prospect case was always the raw power arriving. He is also the player whose real risk is not skill at all: he needs the lineup card. If he is still starting five days a week a month from now, the underlying numbers say he outperforms whatever you paid. If he is platooning, the xSLG is a nice fact about a bench bat.

Cal Raleigh's row cuts the other way and I would not read it as a sell signal. A 175-point xSLG drop is enormous, and for a catcher who carries his value on home runs it is exactly the shape you fear. But catchers grind down, and expected numbers built on batted balls swing hard when a guy spends six weeks fouling pitches off with a beaten-up hand. Hold. The cost of being wrong on holding a top-two catcher for another month is small. The cost of dumping one into a market that has not seen this number yet is a season.

How to read this

Velocity is the most trustworthy line here. It stabilises quickly, it moves for physical reasons, and a starter who has genuinely lost a mile per hour is usually telling you something before the ERA does. It is measured on the primary fastball only, and only for arms with a real sample in both seasons.

The stuff and command section is a descriptive layer and is not part of our projection, and neither is the hitter section. Neither has been validated into the rankings and we do not pretend otherwise: where these and the board disagree, that is something to watch rather than a correction to either.

Low-confidence signals are excluded from both sections, a little under half the feed. A small reading off a handful of batted balls or a few dozen pitches is noise wearing a decimal point.

All articles · Redraft rankings · Dynasty rankings