Secondhand machinery,
priced and explained.

SHM buys and sells used heavy machinery. Pricing lived in one expert's head, and he is leaving. The job is not just to predict a price - it is to earn the trust he had: explain each valuation, admit uncertainty, and flag a mispriced lot. Because SHM both buys and sells, a calibrated, explainable price beats a slightly sharper point estimate.

The model prices machines sold tomorrow, so it is trained on the past and tested on the most recent sales. A random shuffle would let it peek at neighbouring-week prices and look far better than it is. We ran that shuffle too, to measure exactly how much it would have flattered us.

From a Ridge baseline to gradient boosting and an entity-embedding neural net, with the boosters early-stopped on a validation slice. Random Forest is the production model: the interpretable workhorse, it pairs cleanly with exact explanations and calibrates without bias.

Age leads, then the machine's size class parsed out of its product description - the levers an appraiser pulls first. Every prediction breaks down per machine, so a buyer sees why a number came out the way it did.

Conformal prediction wraps each estimate in a calibrated band - the band is what turns a guess into a buy or sell decision. A flat band in log space is wider in dollars for expensive machines, which is the right shape.

Measured only on data it never trained on, the error climbs from about 0.25 to 0.34 across 2009-2012 as the market moves. That is the signal to retrain - tracked, not hidden.