The Sports Page

The Diamond in the Rough Is Mostly Rough.

The Mets are fifteen games under .500 and selling before the deadline, so which prospects are actually worth dreaming on? Only one has enough at-bats to judge — and a late-season audition tells you almost nothing. Here is the math.

The Sports Page · Prospect Math at the Deadline · Model: Beta-Binomial with a league-average prior

The question every fan of a losing team starts asking in July: who in the farm system is worth getting excited about? For the Mets — roughly fifteen games under .500 and selling before the July 31 deadline — there are three answers, and none is the one the romance of a prospect list wants.

One: their best prospect is already gone. On January 21 the Mets traded their No. 1 prospect, shortstop Jett Williams, and pitching prospect Brandon Sproat to Milwaukee for Freddy Peralta and Tobias Myers — a win-now swap made by a team that then lost early and often. The prospect you were going to wait on is now in another organization's Triple-A.

Two: the one bat left to dream on is hitting .191. That is Ryan Clifford, a 22-year-old first baseman/outfielder headed to the All-Star Futures Game in Philadelphia on the strength of raw power — a 90th-percentile exit velocity around 108 mph and a hard-hit rate north of 52 percent, some of the best contact-quality numbers in Triple-A. He also leads the International League with 117 strikeouts. Elite raw material, unfinished contact.

Three, and this is the part that matters most: for everyone below Clifford, the numbers so far are too small a sample to mean much. That is not a knock on the players. It is arithmetic, and it is the reason a losing team's summer feels more promising than it usually is.

The one statistical question

The instinct that keeps hope alive on a bad team is that a small burst of good numbers means something. A September call-up hits .350 over his first three weeks and the fan base pencils him into next year's lineup. The math says: not so fast. A cup of coffee is one of the least informative samples in all of sports.

Use the newsletter's standard tool for a rate like batting average: the Beta-Binomial. Start with a league-average prior — a Beta distribution centered on a .250 hitter, with a strength of about 200 plate appearances of prior belief (alpha = 50, beta = 150). Now let a rookie go 14-for-40, a scorching .350 debut. Update the prior with those 14 hits and 26 outs and the posterior barely twitches:

A .350 explosion over 40 at-bats moves our estimate of the player's true ability from .250 all the way to .267 — seventeen points. A dreadful 6-for-40 pulls it down to .233. The cup of coffee, hot or cold, is almost inert. It cannot make a star and it cannot bury one, because forty at-bats simply do not contain enough information to overwrite what we already believed. This is not a quirk of the chosen prior; it is the arithmetic of small samples. The dream survives the audition precisely because the audition proves nothing.

Why Clifford is the interesting case, not the safe one

Clifford is the exception that clarifies the rule, because his sample is no longer a cup of coffee. Through roughly 320 plate appearances in Triple-A this year, he has enough of a track record that the numbers are starting to speak — and what they say is genuinely ambiguous rather than empty. The power is real and it is elite; contact quality that good is a skill, not a fluke, and it stabilizes faster than batting average. The strikeouts are also real: a rate near 36 percent of plate appearances against Triple-A pitching, well north of the roughly 22 percent a big-league hitter runs, and strikeout rate is one of the fastest-stabilizing stats there is.

So the honest scouting line is not "wait and see." It is a specific, testable bet: the raw power is a finished asset and the contact is the open question. If the whiffs come down even to merely bad, the exit velocity plays and there is a middle-of-the-order bat here. If they do not, he is a Quadruple-A slugger — the kind of hitter whose Triple-A power never survives contact with a major-league breaking ball. The Futures Game invitation is a bet on the first outcome. The .191 is the market pricing in the second.

Who else is in the rough

Below Clifford, the promise is real but comes with the same warning label: small samples, loud tools, quiet results so far.

Every one of them fits the pattern the model warns about. A tool loud enough to dream on, a result column that has not yet cooperated, and a sample too small — or a season too young — to settle the argument. That is what "prospect" means. It is not a promise; it is a probability with a wide error bar.

What it means at the deadline

A seller in late July faces a choice the math frames cleanly. The Mets can cash their veterans for other teams' Cliffords — buying wide error bars in bulk and hoping a few resolve upward — or they can hand their own kids the at-bats a lost season makes cheap. Both are defensible. Neither should be confused with certainty. The great trap of a losing summer is that a two-week hot streak from a call-up feels like a discovery. The Beta-Binomial is there to remind you it is mostly noise, and that the only prospect whose numbers have earned the right to move your priors is the one who has been at it long enough to strike out 117 times.

The romance of the diamond in the rough is durable because it is unfalsifiable in small samples. Enjoy the dream — that is what a lost July is for — but price it honestly. Most rough stays rough. The rare stone that turns into a franchise does not announce itself in forty at-bats. It announces itself over the long, boring, statistically significant grind that separates the two.