Between August 1 and August 15, six prospects we track reached the major leagues. Five of them carried a probability we had locked in public weeks earlier. With Sub-Batch B twenty days from grading, this is the first real look at whether the recalibrated model is doing what it claimed it would do.
Mid-August has become one of the busiest debut windows on the calendar, and the reason is written into the collective bargaining agreement rather than into any scout’s report. Rookie eligibility cutoffs and the Prospect Promotion Incentive both reward teams for getting a top prospect onto the roster while there is still enough season left to matter, and September roster expansion sits right behind it. When our model weights service time and organizational incentive as heavily as it does, this is the stretch where those features either earn their place or do not.
Joshua Baez produced the loudest of the six. He homered on the first pitch he saw in the majors, homered again in his second at bat, and homered a third time in the sixth inning, becoming the first player in the history of the sport to go deep in each of his first three career at bats. He finished 3 for 4 with five runs driven in as St. Louis beat the Cubs at Wrigley. We had him at 22 percent.
Angel Genao was the most predictable and, for our purposes, the most useful. He collected four straight hits in his debut on August 5, the first player in Guardians history to get four in a first game and the twenty-second since 1900. He was our highest featured number in Sub-Batch C at 45 percent, and we had written three weeks earlier that his 40-man roster spot and Triple-A assignment made him the most likely name in the cohort. Kaelen Culpepper homered in his second plate appearance for Minnesota. Hagen Smith threw two hitless innings for the White Sox. George Lombard Jr. homered for his first career hit in the Bronx.
Here is the part that matters more than any individual debut. A probability model is not judged on whether its favorites come in. It is judged on whether the group it called 45 percent shows up more often than the group it called 25 percent, and whether the numbers land near the rates they promised.
Read down the chart. Players scored at 45 percent or higher have debuted half the time. The 30 to 44 percent group is at 22 percent. The 20 to 29 percent group is at 4 percent. Everyone the model put under 20 percent has not debuted at all, zero for thirty-nine.
That ordering is perfect, and it is worth being clear about why it is the meaningful result. The June 18 checkpoint showed the old model ranked players correctly while printing probabilities that ran low, and the v0.5 patch rebuilt the mapping table to fix exactly that. If the fix worked, we would expect not just correct order but rates that track the labels. Twenty days from the first grading date, the top band is running at 50 percent against a 45 percent prediction and the bottom band is running at zero against a prediction near 5 percent. That is what calibration looks like when it is working.
Two honest qualifications. These windows are still open, so every one of these rates will rise before it is final, and the top band is four players, which is not a sample anybody should lean on. The real number arrives on September 9 for Sub-Batch B and October 8 for Sub-Batch C, and it arrives as a Brier score whether it flatters the model or not.
The other thing that happened this month is the one we have been telling you about since July. Miami has decided to shut Thomas White down for the rest of the season and prepare him for 2027 rather than push the left shoulder capsular sprain back onto a mound. He was our largest published number at 62 percent, and he will not pitch in the majors this year.
So Sub-Batch B grades on September 9 with its single biggest call already lost. There is no version of that which is good news, and there is a version of it that is useful: we said in the July injury watch that this number was probably going to miss, and we said it while it could still have gone either way. The miss goes on the log with a named reason, which is that the model scored a pitcher at 62 percent while he was already on the minor league injured list. An injury-status feature at run date is the leading candidate for the v1.0 build, and the reasoning is written up in Inside v0.5.
The same article was right about Culpepper for the opposite reason. He returned from the thumb injury and debuted on August 7, and his 33 percent resolved yes. One piece, two injured players, two different guesses, both landing where they were argued.
Five graduations reshaped the top of the live board. Charlie Condon leads at 48 after a Triple-A season that has him second in the Pacific Coast League in OPS with 20 home runs. White sits beside him at 48 and falling, since the shutdown registered as the organization signaling patience. Lazaro Montes holds at 44 in Tacoma. Carlos Lagrange dropped 12 points to 26 after landing on the Triple-A injured list with a shoulder problem of his own, which is a real blow to a converted reliever who was being built up for a big league bullpen role this year.
Eighty-seven players remain active on the board. Twelve of the fourteen Sub-Batch B predictions are still open with twenty days to run, and sixty-two of the sixty-five in Sub-Batch C have until October 8.
All five of these receipts were filed today, in one sitting, covering debuts that happened between August 1 and August 15. That is late, and it is the same failure we disclosed with Luis Lara in July and Aiva Arquette earlier this month. The daily scan is not catching debuts reliably enough, and the fix is the audit we described: a weekly pass that flags any tracked player with no ledger activity in thirty days. Seventy-one active players currently carry no logged event at all, which is exactly the blind spot that lets a debut sit unrecorded for two weeks.
The locked predictions were never affected. Those numbers were set in June and July and cannot move. What lagged was the live commentary on top of them, which is the less important half of the system and still the half we said we would run properly.
Sources: MLB.com on Baez’s three-homer debut · MLB.com on Genao’s call-up · MLB.com on Hagen Smith · MLB Trade Rumors on Culpepper · MLB.com on Lombard · MLB.com on White’s shutdown · SI on Lagrange’s IL stint. Model values from the Sub-Batch C run record and the public event ledger.