Multi-Week Trajectory Curve
Understanding saved exceptions in the two-track model
Read the explanation
The two track simulator has exact input exceptions. Its saved Nicole fixture, eight point nine million streams, five thousand sales and two point two million airplay, returns two hundred fourteen point eight points directly. The general equation would multiply those displayed units by fourteen point two, ten point four and twelve point four, sum them and round to two hundred five point seven. The bars compare these values at one point five pixels per point. A slight input change can leave the exact fixture branch and change the result unexpectedly. These authored calibrations are not independently verified chart scoring or track performance. The source peak function accepts a debut rank argument, but its generic rank calculation does not use that argument. It starts at twenty seven and subtracts point one four three times points minus three hundred twelve point four, then rounds and clips to one through one hundred. Seven extra points near that baseline change rank twenty seven to twenty six. The two bars show twenty seven and twenty six at ten pixels per rank; better rank is shorter. Special near point matches can return fixed twenty seven or forty one. This is an authored mapping, not a confidence interval or observed next chart rank. The generated trajectory uses debut rank for its first point and peak rank for the second. Later ranks then worsen by saved growth factors and additive offsets: peak twenty seven becomes thirty one, then thirty eight, then forty nine. The bars compare twenty seven and forty nine at seven pixels per rank. The combined points label simply sums the two tracks and does not enter either track peak calculation. None of these weeks are fetched observations. In an offline test without D three the source numeric update and chart renderer differ in availability, so local checks must report native export and preserved geometry separately from healthy chart drawing.