How FactorForge Scores Your Portfolio
April 2026
Ask most investors what kind of investor they are and they'll have an answer. Value, growth, somewhere in between. What they rarely have is data to back it up.
Their holdings are spread across two or three brokerages, maybe a 401k on top of that. Each platform shows your own slice — returns, allocations, a few charts. None of them tell you what the whole picture says about you.
The tools that could answer that question are either too simple — pie charts, performance returns, basic allocations — or require a Bloomberg terminal and a finance degree to operate.
FactorForge is built for the space in between. Upload your holdings from any brokerage and the model evaluates every position across six independent investing frameworks. The result isn't just a score — it's a portfolio-level view of your actual factor exposure: where you're concentrated, what style your holdings collectively reflect, and where different lenses agree or disagree.
Six lenses, six independent scores
Every holding is evaluated through six frameworks that run independently — each lens doesn't know what the others concluded, and they can disagree.
| Lens | What it measures | Representative metrics |
|---|---|---|
| Value | Is the stock trading below its fundamental worth? | P/E, P/FCF (price to free cash flow), P/B (price to book), P/S |
| Quality | Is the business financially strong and cash-generative? | FCF margin, gross margin, accruals ratio (cash earnings quality), net debt/EBITDA (debt relative to earnings) |
| Growth | Is the company expanding revenue and earnings? | Revenue growth, revenue acceleration (is growth speeding up or slowing down?), margin trend, ROE (return on equity) trend |
| Momentum | Are recent price and earnings trends working in its favor? | 12-month price return, analyst rating, short interest |
| GARP | Does growth justify the price? | PEG ratio, revenue growth, ROE |
| Balanced | Equal weight across all five frameworks | — |
Each metric is scored on a continuous scale and weighted by how reliably it predicts outcomes in that lens — a metric with stronger historical signal carries more weight.
The result for each lens is a single score and a verdict: Strong, Moderate, or Weak. Those verdicts appear on a dashboard — one per lens, per holding — so you can see at a glance where lenses agree and where they diverge.
Sector calibration
The same metric means different things in different industries. A P/E ratio that signals undervaluation in technology would look stretched in utilities. A high price-to-sales ratio is expected for a SaaS company with 80% gross margins — alarming for a grocery chain running on 2%.
Every threshold in the model is sector-calibrated. The model knows it's comparing a semiconductor company to other semiconductor companies, not to the market at large.
Fundamental data — earnings, book value, debt, cash flow, gross profit — is sourced directly from SEC EDGAR filings: the official regulatory filings companies submit to the U.S. government. These are the actual numbers, not estimates from a third-party data vendor.
How a verdict is reached
Individual metric scores sum to a lens score. That score maps to a verdict — Strong, Moderate, or Weak — with a confidence level.
Confidence is HIGH when most of a lens's weighted metrics point the same direction. When five quality metrics all read positively, the model is relatively certain.
Confidence is MEDIUM when signals conflict, data is sparse, or an edge case limits what the model can reliably compute.
Confidence is LOW when the model has too few scoreable metrics to reach a reliable conclusion — common for companies with limited filing history or unusual financial structures.
You can see both effects in recent analyses on this blog. GOOGL ran 5 Strong and 1 Moderate across its six lenses — HIGH confidence throughout, with the value lens as the lone dissenter on price multiples. SBUX showed a split verdict with MEDIUM confidence on the value lens, partly because negative book equity (from years of aggressive buybacks) makes certain value metrics mathematically undefined rather than genuinely favorable.
When the model isn't sure, it says so.
Those per-holding verdicts roll up into a portfolio-level picture — your overall style tilt, sector concentration, dividend income, and ETF overlap across all your holdings at once.
ETF look-through
If you hold ETFs, the model doesn't treat them as black boxes. For major U.S. ETFs — including iShares, Vanguard, SPDR, and Invesco factor funds — the model expands each fund to its underlying holdings and scores each position individually, weighted by its share of the fund.
A portfolio of three ETFs still gets per-stock factor scoring. You can see which underlying positions are driving a fund's quality score, which ones are dragging on momentum, and where your funds overlap in ways you might not have expected.
What the model doesn't capture
Factor models have known edges and known blind spots.
Momentum can invert. The momentum lens is calibrated for fundamentals-driven markets. When markets are driven by narrative or liquidity rather than earnings, price returns can run ahead of the fundamentals the model tracks — and stocks in that gap may score lower than their recent returns suggest. This is a known limitation of factor-based momentum signals, not a bug.
EDGAR data lags. Fundamental data comes from annual 10-K filings. A company that deteriorated in the middle of a fiscal year won't show it in the model until the next filing. The model captures the business as of the last reported period, not in real time.
International ETF sector weights are approximated. For funds holding international equities, the sector allocation data is estimated rather than scraped from primary sources. Scores for these funds carry more uncertainty than domestic ETF scores.
The model scores what it can measure. It has no view on management quality, competitive moats, regulatory risk, or macro. A company with excellent financials and a strategy headed in the wrong direction will score well until the financials catch up.
The limitations don't make the model less useful; they make it more honest. Knowing what a tool doesn't see is as important as knowing what it does.
The best way to understand what the model sees in your holdings is to run them through it. If you'd like to follow along as the methodology evolves — new metrics, new lenses, improvements to the model — the list below is the best way to stay informed.
FactorForge analysis is for informational purposes only and does not constitute investment advice. Factor scores reflect model output based on historical data and may not predict future performance. See our Terms of Service for full disclaimers.