Putnam Proof Trainer

Calibrated proof practice for serious Putnam prep.

YC energy, proof discipline

Train proofs like it is a real operating system.

Putnam Proof Trainer turns the historical archive into a focused practice engine. It tracks work, estimates your level, and routes you toward confidence wins, on-level reps, and stretch problems using archived Top N score distributions as calibration data.

Static historical datasetServer-side practice stateAdaptive difficulty ladder
1Single-user system
0Runtime scraping jobs
12Problem slots tracked
Top NCalibration population

Built for training, not browsing

The archive is still there, but the product is organized around practice tracking, difficulty calibration, and next-problem selection rather than passive chart consumption.

Practice memory

Track attempted, partially solved, solved, hinted, abandoned, and archived work with notes, time, and attempt count.

Difficulty calibration

Use perfect solve rate, nonzero rate, attempt rate, and average score to estimate how demanding each archived problem was.

Adaptive queue

Get a ladder of recommendations around the current level estimate while keeping the archive available for manual selection.

Simple loop, strong signal

The product stays narrow on purpose: pick a problem, work it seriously, record the result, and let the queue adjust.

01

Open the next problem

Use the recommended queue or jump directly to any year and slot in the archive.

02

Log the attempt cleanly

Record status, time, attempts, notes, and whether the result came with hints or solution reading.

03

Recalibrate

Use the historical distribution data to decide whether practice should move down, hold, or push harder.

Private training infrastructure

Historical data stays static. Practice state persists server-side. The app is password-protected and optimized for one user who wants a disciplined Putnam workflow instead of another dashboard.