The source intends to compare an exponential frontier with a linear consensus. With exponent point thirty five and initial compute exponent twenty six, frontier is twelve times ten to point thirty five times two plus point eighty five times years, capped at one hundred. Counterfactual baseline is about sixty point fourteen; year three is capped one hundred. Consensus starts forty and adds six per year, reaching fifty eight. Bars encode one hundred and fifty eight at four pixels per unit. These are authored synthetic formulas rather than empirical capability units. The original script contains a malformed parenthesis in a milestone expression, causing a parse error before any of its code or bindings execute. The local draft preserves that original failure and does not pretend these formulas run. The intended loop creates forty one samples at quarter-year intervals spanning twenty twenty five to twenty thirty five. Bars encode forty one sample positions and three milestones at ten pixels per count. The scrubber instead moves by tenths of a year; a tolerance of point fifteen can match two nearby samples and the later matching sample overwrites the earlier ratio. The break rule requires frontier greater than one point two five times consensus after twenty twenty five point five, and floors the first sampled year. If no break occurs it still falls back to twenty twenty seven. That fallback is not evidence a break was detected. Native slider movement cannot recompute the original frozen captions because the script never parses. Counterfactual year-three frontier one hundred divided by consensus fifty eight is about one point seventy two, not the static four point two displayed in the original markup. Both bars use one hundred pixels per ratio unit, so their widths consistently compare the intended calculation and frozen caption. The horizon caption is separately forced by twenty twenty seven minus first milestone consensus year twenty twenty nine times twelve, minus twenty four months. It does not solve the scaling curve for a milestone arrival. Native export is unbound due to the parse error and produces no file. Preserving that failure, static SVG allocation and all-footer placement is local subset evidence; no forecast, scientific validation, public video or full functional completion is claimed.