Five stocks that returned 100%+ in 2025 but graded C or D under the 7-framework consensus screen at entry. The honest catalog of what the framework missed and why. Every methodology has structural blind spots; the cheapest insurance is naming them publicly.
All seven frameworks scored Palantir poorly at the start of 2025 because the gov-software revenue base was lumpy, the FCF/share growth was negative, and the valuation was already stretched. The rubric was right on the fundamentals and wrong on the multiple expansion.
SMCI looked like a thin-margin commodity-server reseller in early 2024, which it was. The AI capex tailwind compressed years of demand into months. The framework correctly priced a low-margin business; it had no mechanism to price the buildout cycle.
Carvana was a near-bankruptcy turnaround. The frameworks all flagged the debt and the operating losses (correctly). The framework has no special-situation deal-breaker override; turnaround compounding is structurally invisible to a quality-and-value rubric.
Mobile ad-tech with thin disclosure. The frameworks scored conservatively on management transparency (correctly). The ad-platform pivot worked spectacularly; the rubric had no way to price the operational improvement until quarterly results confirmed it.
Utility holding with merchant-power exposure. The frameworks priced a steady regulated utility; the AI-data-center power demand surge transformed the merchant component into a high-margin growth business. Framework can't anticipate exogenous demand shocks.
All five misses share a common structural feature: the forward returns came from exogenous events the framework has no mechanism to price. Multiple expansion driven by AI narrative (PLTR, SMCI), turnaround dynamics accelerated by macro tailwinds (CVNA), operational pivots invisible until quarterly results confirmed them (APP), and capacity shocks from outside the company's control (VST). The framework prices business quality and current valuation. It cannot price imminent narrative shifts, special situations, or exogenous demand shocks.
The 5+/7 consensus cohort returned a median 24.3% in 2025; if you had owned the five misses above in equal weight, you would have returned 187%. The framework missed these wins systematically. That is the methodology's honest cost of doing business: by requiring stocks to satisfy multiple value-and-quality criteria simultaneously, the screen necessarily filters out the narrative-driven multi-baggers.
The right response to documented misses is not to widen the rubric until it includes the misses. Adding deal- breaker overrides for "exception cases" would dilute the structural discipline that produces the base-rate alpha. Every miss case has its own different cause; chasing each individually would generalise to a framework with no falsifiable criteria.
The right response is to acknowledge the misses and accept that the framework targets the median outcome, not the multi-bagger tail. An investor who wants to capture the multi-bagger tail should pair the consensus screen with separate momentum or special-situation tooling and treat the two strategies as orthogonal. The consensus screen and a small momentum sleeve are not in conflict; they target different return distributions.
For investors using the consensus screen, the lesson is calibration: the 73.8 percentage-point alpha is real but comes with a clear cost. The five biggest 2025 winners were not in the cohort. The five biggest 2025 losers were also not in the cohort. That is the trade.
Educational only. Past performance does not guarantee future results. Documented misses are part of any methodology's honest record.