Do the scores predict returns?

Every week the ranked universe is frozen into a point-in-time ledger, and every month each score is tested against what the stocks did afterwards: the rank correlation with forward returns, equal-weight quintile portfolios against the S&P 500, and forward returns by discount to fair value. Results are published on this page with the number of matured dates per horizon, the sample size and the p-value, so a horizon that cannot be measured yet reads as not yet rather than as no effect. One-year evidence matures around March 2027.

What does the moat score measure?

The moat score evaluates how durable the competitive advantage is: whether the business can defend its profits over time. It weighs network effects, switching costs, cost advantages, intangible assets, efficient scale, ROIC sustainability, and signs of moat erosion. A high moat score is the precondition for almost every other dimension carrying its full weight.

How is valuation scored?

The valuation score asks whether the current price is justified by the business fundamentals and what growth rate the market is pricing in. It uses reverse-DCF, fair value versus margin-of-safety price, achievable versus implied growth, historical valuation context, catalysts, and supply-demand dynamics. A reverse-DCF is the sharpest single lens. It exposes the assumption the price is buying.

How does the methodology weight different investment styles?

The six dimensions are each scored 0 to 10 and the quality score is a fixed weighted average: moat, business model and valuation 20% each, risk and management 15% each, optionality 10%. The growth reads (growth potential and disruption) and the twelve-month read (catalysts, expected return, risk and reward) are separate components with their own weightings; they never change the six underlying scores.

Are the scores investment advice?

No. Scores are analytical signals, not investment advice. They cannot capture management character in private, regulatory surprises, macro shifts, or black swan events. Use the scores as a starting point for your own research, not as the final word. Form your own thesis, define your own kill criteria, and consult a qualified financial adviser before investing.

INVESTBOARD

Value investing · scoring methodology

How we score companies.

Six dimensions, five mental models, one transparent framework. So you can see exactly what drove every score in your portfolio.

Overview.

Six dimensions · five mental models

Every company in your portfolio is evaluated across six dimensions, each scored 0–10. The analysis draws on SEC filings, financial statements, earnings call transcripts, market data, and competitive intelligence. Five mental models are applied throughout: first principles thinking, inverse thinking, second-order effects, pattern recognition, and deferred gratification analysis.

Investment style weighting.

Style shifts the weights · same six dimensions

Weighted · moat, business model, valuation 20% each

Quality score

Risk and management 15% each, optionality 10%. The long-term composite.

Growth + disruption · max 80

Growth investing

Growth potential (25%) and disruption (25%) weighted highest. Optionality matters most.

Catalysts + reward · max 40

Active stock picking

Catalysts, expected return, and risk/reward emphasised. Twelve-month horizon.

Scoring categories.

Six dimensions · weighted by lens

Competitive Moat

How durable is the competitive advantage? We evaluate whether the business can defend its profits over time.

·Network effects: does the product become more valuable as more people use it?

·Switching costs: how painful is it for customers to leave?

·Cost advantages and economies of scale vs competitors

·Intangible assets: brands, patents, regulatory licenses

·Efficient scale — does the market naturally limit competition?

·ROIC sustainability: is return on capital backed by real structural advantage or temporary?

·Moat erosion signals: are competitors gaining ground or is the moat widening?

Management Quality

Capital allocation skill, alignment with shareholders, and long-term orientation. Actions over words.

·Capital allocation track record: ROIC trend over time, not just current level

·Insider ownership: do executives have meaningful skin in the game (5%+)?

·CEO compensation relative to company size and performance

·Candor and transparency: do they openly discuss problems and mistakes?

·FCF/Debt ratio: prudent debt management and financial conservatism

·Deferred gratification: choosing long-term value over short-term earnings beats

·Adaptiveness: proactive change leadership, not just reacting to disruption

·ROE stability — consistent returns, not deteriorating over time

Valuation

Is the current price justified by the business fundamentals? What growth rate is the market pricing in?

·Reverse DCF: what growth rate does the current stock price imply?

·Fair value vs margin of safety price using 10-year DCF model

·Achievable growth rate vs implied rate: is the market under- or over-pricing growth?

·Historical valuation context: P/E, P/S, P/FCF vs own history and peers

·Strategic catalysts that could drive multiple expansion over 5-10 years

·Supply-demand dynamics: structural demand shifts can override traditional valuation

Business Model

Economic durability and cash generation power. Can this business compound capital efficiently?

·Gross margin, operating margin, and net margin trends over time

·ROIC and ROE: does the company earn above its cost of capital?

·Free cash flow generation consistency and growth

·Revenue and earnings CAGR (3-year and 5-year)

·Capital intensity: light-asset models compound better

·Reinvestment rate and reinvestment runway for growth

·Stock-based compensation as % of revenue: real dilution cost

Risk Assessment

What could permanently impair capital? Inverse thinking: how could this investment fail?

·Net debt relative to free cash flow: can the company service its debt?

·Financial health: liquidity, interest coverage, cash position

·Competitive threats — are challengers eroding market share?

·Regulatory and geopolitical exposure

·Technology disruption risk: could the core business become obsolete?

·Customer and revenue concentration

·Second-order effects: what happens downstream if key risks materialize?

Optionality & Growth

Hidden upside beyond the core business. Future growth vectors that the market may be underpricing.

·Adjacent market opportunities and total addressable market expansion

·Platform and ecosystem leverage — can one success unlock many?

·R&D pipeline and innovation track record

·International expansion potential

·Real options value: what valuable bets does the company hold?

·Management vision for 5-10 year horizon

Do the scores work?

Evidence · weekly ledger since 8 Mar 2026 · run 2 Oct 2026

Every week the ranked universe is frozen into a ledger. Every month each score is tested against what the stocks did next: the rank correlation with forward returns, quintile portfolios against the S&P 500, and what price did after a discount to fair value. Small claims, dated, losers included. One-year evidence lands around March 2027, two-year around March 2028; until then the clock below says what can be measured.

What can be measured yet.

30 weekly snapshots · 382 names scored · 269 in this universe

1 month

snapshot dates matured

3 months

snapshot dates matured

6 months

snapshot dates matured

1 year

not yet

no window has closed

2 years

not yet

no window has closed

3 years

not yet

no window has closed

The whole ranking, not just the top five.

Value style score (value-trap adjusted) · equal-weight quintiles, held week to week

Q1 · highest score
Q2
Q3
Q4
Q5 · lowest score
S&P 500

Annualised	Q1	Q2	Q3	Q4	Q5	S&P 500	Q1 − Q5
Return	+28.3%	+25.0%	+16.7%	+44.0%	+26.0%	+25.2%	-2.4%
Alpha vs market	-6.4%	-7.6%	-7.8%	+2.1%	-15.4%	—	+9.0%

8 Mar 2026 to 30 Sept 2026 · 30 weekly periods · about 38 names per quintile · price returns, no dividends or costs

The weekly picks, the top of this ranking run as one live portfolio since June 2025, keep their own record on Discover →

Does the rank predict the return?

Spearman rank IC · mean across snapshot dates · Newey-West p

Horizon	Rank IC	p	Dates positive	Top − bottom	Dates	Names
1 month	+0.03	0.54	65%	+0.6%	26	180
3 months	+0.10	0.09	72%	+3.7%	18	161
6 months	-0.02	0.65	50%	-15.6%	4	101
1 year	not measurable yet
2 years	not measurable yet
3 years	not measurable yet

Which scores carry it.

Rank IC per dimension · ink when p < 0.10, bold when p < 0.05

Score	1 month	3 months	6 months	Trailing 52 weeks · 3-month IC
Moat	-0.00	+0.07	-0.05	+0.07
Risk	+0.02	+0.11	+0.04	+0.11
Valuation	+0.04	+0.04	-0.01	+0.04
Optionality	+0.05	+0.08	-0.03	+0.08
Management	+0.05	+0.10	+0.01	+0.10
Business model	+0.03	+0.09	+0.01	+0.09
Value-trap score	+0.04	+0.13	not yet	—

Not measurable yet: secular tailwind · expected annual return (quick dcf) · discount to fair value · margin-of-safety gap · insider net buying (3m) · super-investor holders · ai harness probability · ai resilience (10 − exposure). These entered the ledger later and get their first reading when a window closes.

Does price follow value?

Forward return by discount to fair value · one observation per company per window

Not measurable yet. Fair values entered the weekly ledger on 20 September 2026; the first three-month reading lands once those windows close, around December 2026.

Does the value-trap call hold up?

Excess return after the call · 6 months

Call	Names	vs S&P 500	Beat the market	p
Not a trap	29	+3.5%	66%	0.71
Uncertain	14	-8.2%	64%	0.37

What the data would weight.

Fitted on one-year excess returns · never applied automatically

Not yet. Needs 26 snapshot dates with 100+ names and matured 1yr returns; have 0. Until then the six dimensions carry equal hand weights.

After a kill criterion trips.

Return vs the market from the trip

No kill criterion has tripped in the ledger yet, so there is nothing to measure.

How it is measured.

methodology 2026-09-20.1

Rank IC: per-snapshot-date Spearman correlation between the signal and the forward return; mean across dates with Newey-West standard errors (lag = horizon in weeks) so overlapping windows do not inflate significance.
Quintile portfolios: equal-weight buckets formed at each weekly snapshot and held to the next one (calendar-time approach; Fama 1998, Mitchell & Stafford 2000). Q1 = highest signal. Returns are price returns in local currency; the benchmark is the S&P 500 price index, so non-US names in the ALL universe carry currency noise.
Ladders and classification stats pool at most one observation per company per non-overlapping horizon window.
Fitted weights are a Fama-MacBeth regression of forward one-year excess return on the six standardised Value-style dimensions, reported next to the hand weights and never applied automatically.
Trading costs, dividends and taxes are not modelled on either side.
Analyses are re-run when a user asks for them, so a score can stand unchanged for months. The ledger records the standing claim each week together with its analysis date; the event study (EventStudyStack) separately measures returns from each fresh analysis.

Run 2 Oct 2026 · general investment analysis · not personalized advice

Data sources.

Publicly available · numbered citations

All analysis is grounded in publicly available data: SEC filings (10-K, 10-Q, 8-K) via EDGAR, earnings call transcripts, quarterly financial statements, market pricing data, insider transaction records, and analyst estimates. Every analysis includes numbered source citations that link to original documents, so you can verify any claim.

Limitations.

What scores can't capture

Scores are analytical signals, not investment advice. They cannot capture everything: management character in private, regulatory surprises, macro shifts, or black swan events. Use scores as a starting point for your own research, not as the final word. Always form your own thesis and define your own kill criteria before investing.

From the writing

The circle of competenceBuffett's most quoted concept, made operational: the three tests that filter whether a position is inside the circle or outside.Optionality, the hidden upsideHow to spot the second business inside a great business before the segment shows up in the income statement.The value-trap checklistSix red flags every cheap-looking stock should clear before you size — and what distinguishes a real turnaround from a falling knife.