A typical trade analyzer gives you a single number: "Team A wins by 7.2." That number is almost always generated by summing per-player season-long value, which is useful exactly when nothing else on your roster matters — i.e., never. A trade is a delta on a roster, in a league, with a schedule. Our analyzer treats it that way.
The impact tree, not the impact number
Trade evaluation should consider category impact, roster fit, and the schedule rather than only a summed player value. Pick Pilot's beta uses rolling form and available schedule context, not an active licensed rest-of-season projection feed. Longer-horizon estimates depend on those inputs and their stated assumptions; they do not establish either manager's future record.
League-native, not global
Most public tools weight categories uniformly. Yours are weighted by how close the category is in your standings. If you're tied in rebounds and blocks but 40 clear in points, a trade that improves your points by 12% is worth less than a trade that nudges rebounds by 3% — and the analyzer knows it because it looks at category_standings, not an abstract replacement baseline.
The 30/60-day post-hoc grade
The analysis doesn't stop at accept. For every trade in the league, Pick Pilot records the projection delta and then, 30 and 60 days later, the actual delta. The spread between them is the "grade." Over time the grade becomes your own tune-able dial: if you systematically over-value rookies in trades, the grade tells you before you do it a fourth time.
Refusals are a feature
Sometimes the available inputs cannot support a single verdict. A rolling-form improvement can conflict with playoff-week schedule needs. Review the result's limitations and confidence rather than treating either estimate as a guaranteed outcome or assuming a universal refusal threshold.