What Travels

Every July the same scene repeats. The trade breaks, the graphic goes up — 3.18 ERA, or 5.02 — and an entire fanbase decides in four seconds whether their team just got a steal or a salary dump. It is the most natural thing in the world to grade the trade on the line. It is also, by our own published numbers, grading the acquisition on the least reliable information available about him. This piece is the field guide: when your team trades for a pitcher this month, here is what he actually brings in the moving truck — and what stays behind.

Nothing here is new analysis. It’s the deadline-shaped assembly of four studies we’ve already published and stress-tested — The Coin-Flip Line, Same Stuff, Different Brain, What’s Real and What’s Noise, and The Firemen — each verified by two independent methods before it ran.

1. The arm travels. The line doesn’t.

Start with the one chart to keep open all deadline week. Take every pitcher stat we measured in the Coin-Flip Line study — same test, same players, same window for all of them — and sort by how well each one repeats from season to season. That repeatability is a good proxy for the question a trade actually asks: how much of this number is the pitcher, and how much is the situation he’s leaving?

Three tiers, one ruler. The arm — release point, arm angle, extension, spin, velocity — repeats at 0.90–0.97: your new pitcher’s physical identity arrives intact on day one. The skills — what he throws, the ground balls, the whiffs and strikeouts — repeat at 0.62–0.80: real traits, visible within a few starts. The line — wOBA against, BABIP, barrel rate, win probability, even xwOBA against — repeats at 0.10–0.23. The half-season numbers a trade is graded on carry mostly bounce, and they are the only tier the broadcast graphic shows you.

This is the same lesson the Coin-Flip Line put a mechanism under: a stat’s reliability tracks how much signal each event carries. A radar gun reading is nearly pure signal — one fastball settles it. A half-season of contact outcomes is a few hundred coin-weighted events wearing a decimal point. Neither trading team controls that; it’s just what the numbers are made of.

2. The brain travels too

The subtler cargo is the one we spent two research rounds on this month. Some starters reliably beat what their raw stuff says they should allow — an edge worth about 26 runs a season from best to worst, repeating year over year at 0.59 (a different test than the chart above — a year-over-year correlation of each starter’s modeled edge — so read it beside the ladder, not on it). The deadline question is obvious: if a crafty starter’s edge comes from his catcher’s game-calling or a friendly park, a trade destroys it.

It doesn’t. When we hunted for what carries that edge, park and catcher combined were bounded to a small slice — roughly 85% of the edge belongs to the pitcher himself. We couldn’t name the mechanism (nobody can yet — that was the honest headline of Same Stuff, Different Brain), but we could show it isn’t the things a trade takes away. The brain rides in the truck. If your team traded for a pitcher who beats his stuff, expect him to keep beating it in the new uniform.

3. The reliever special — because most deadline trades are relievers

Deadline week is mostly bullpen week, and bullpen trades are graded on the noisiest ledger in the sport. From our bullpen studies, in one breath: a reliever’s velocity carries year to year at 0.95 and his whiff rate at 0.67, while the win-probability ledger he was just traded on — the saves, the blown leads, the “he’s been shaky lately” — repeats at 0.17, and his blown-lead rate at essentially zero. Twenty leveraged innings is a tiny sample wearing a narrative.

So when the new setup man blows a lead in his first week — someone will — the overwhelming likelihood is that nothing is wrong. The stuff your front office bought is the stable part; the first month’s ledger is the weather. And remember what the save column can’t see: 84% of the win probability relievers create comes outside save situations, so a trade for a “non-closer” can quietly be a trade for one of the most valuable arms on the market. That’s not a hypothetical — it’s where the hidden firemen live.

4. How to read your new pitcher

The field guide, in the order the information becomes trustworthy:

• Day one: velocity, movement, release — believe them immediately. They are him.
• Within a few starts: whiffs, strikeouts, ground balls, what he throws. Real traits, fast to stabilize.
• Trust the track record, not the season: if he’s a starter with a multi-year history of beating his stuff, that edge is his and it travels. Don’t let a sour half-season line talk you out of a durable trait.
• Discount almost entirely: the half-season ERA, wOBA against, BABIP, HR/FB, and every win-probability stat — in either direction. The bargain ace and the broken bust implied by those numbers are, more often than not, the same pitcher having different months.
• If it’s a reliever: all of the above, doubled. His ledger is the least informative thing on his card.

None of this says trades don’t matter — roles change, usage changes, and a new pitching lab can genuinely rebuild an arsenal. What it says is narrower and more useful: the number on the trade graphic is the one part of the pitcher that isn’t coming.