Fall 2026

Consider an election model forecasting a candidate having a 70% chance of winning. Does that mean the candidate will receive 70% of the vote? No. It means that, across a range of plausible scenarios, the model expects that candidate to win about seven times out of ten. If the other candidate wins, the model was not necessarily wrong; a 30% event can happen anytime.

College admissions works much the same way. When we label an application a “reach,” “target,” or “likely,” we are not predicting an individual outcome with certainty. We are describing a range of possible outcomes based on the years of application information available to us. We are actually pretty good at it: over the past three years, the College Counseling Office has correctly predicted over 90% of our students’ application outcomes. That gives us confidence in the data, experience, and judgment behind our counseling. But does it make us clairvoyant? Unfortunately for all of us (and for all of you!), no.

Every year a student is admitted somewhere the historical data suggested was unlikely. Another receives disappointing news from a college where nearly every indicator pointed in the more favorable direction. A waitlist appears where we expected an admit and an admit appears where we expected a waitlist. Truthfully, our sample sizes are small. The electorate changes every year and so do the candidates. There are variables we can observe but cannot fully measure: the strength of an essay (and the person who reads it!), the composition of an overall applicant pool, the particular institutional priorities of a college that year, and some variables we cannot observe at all.

So of course, we ask you to use the data available to you and to ask us what we are seeing in return. Build a college list focused on both probabilities and possibilities. Our experienced team has seen enough cycles to recognize patterns but please know: we will always leave room for the margin of error. If three years at 90% has taught us anything, it’s to never call the race before the polls close.