Framework

The Graham-Dodd
framework, codified

How we translate 90 years of value investing doctrine into a quantitative screening system — and extend it with live merger arbitrage signals.

Origins

Standing on the
shoulders of giants

Benjamin Graham and David Dodd published Security Analysis in 1934, in the aftermath of the 1929 crash. Their core insight: the market is a voting machine in the short run and a weighing machine in the long run. Price and value diverge — and patient investors can exploit that divergence.

"The margin of safety is the central concept of investment. It is the secret of sound investment in a single phrase."
— BENJAMIN GRAHAM · THE INTELLIGENT INVESTOR, 1949

QuantumCompass operationalizes these principles into a real-time screening engine — not as a black box, but as a transparent, inspectable system where you can see every number and understand every calculation.

1934
Security Analysis published
Graham and Dodd establish the intellectual framework for value investing, including NCAV, earnings capitalization, and the concept of intrinsic value.
1949
The Intelligent Investor
Graham refines the framework for individual investors, introduces Mr. Market, and formalizes the 33% margin of safety threshold. Warren Buffett calls it "the best book about investing ever written."
1956
Graham-Newman dissolves
Graham's fund, which returned ~20%/year from 1936–1956, closes. His students — Buffett, Schloss, Ruane — go on to prove the framework at scale.
1984
Superinvestors of Graham-and-Doddsville
Buffett's Columbia speech demolishes the efficient market hypothesis by documenting 9 students of Graham who all beat the market using the same value framework.
2026
QuantumCompass
The complete Graham-Dodd toolkit — four IV methods, five scoring dimensions, live merger arb — available in a single instrument panel.
Valuation

Four methods.
One blended view.

No single valuation method is reliable in all conditions. QuantumCompass blends four approaches — weighting them by conservatism — so that disagreement between methods signals uncertainty, and agreement signals conviction.

25%
01 / 04
Moderate
Graham Number
√( 22.5 × EPS × BVPS )
Graham's own formula for the maximum price a defensive investor should pay. Combines earnings yield (P/E ≤ 15) and price-to-book (P/B ≤ 1.5) into a single threshold. Returns zero for companies with negative earnings or book value.
25%
02 / 04
Moderate
Earnings Power Value
EBIT × (1 − tax) / WACC
Bruce Greenwald's extension of Graham — values a company on its normalized earnings power, assuming zero growth. Strips out capital expenditures required for growth, leaving only the value of the current business in steady state.
30%
03 / 04
Most Conservative
Net Current Asset Value
( Current Assets − Total Liabilities ) / Shares
Graham's most conservative method — the true "cigar butt" screen. If a stock trades below NCAV, you're buying the liquidation value of current assets and getting all long-term assets for free. Receives the highest weight because it is the hardest test to pass.
20%
04 / 04
Most Liberal
Liquidation Value
( Tangible Assets × haircuts ) − Liabilities
Estimates what a business would fetch in an orderly liquidation. Applies haircuts to asset classes: receivables at 85%, inventory at 50-75%, fixed assets at 15-50%. More generous than NCAV because it values long-term assets. Receives the lowest weight accordingly.
Scoring

Five dimensions.
One composite score.

The composite score ranks securities within their universe across five weighted dimensions. It is a relative ranking tool — not an absolute buy signal. A score of 80 means this security ranks better than ~80% of its peers across all five dimensions.

30%
Value
Based on Margin of Safety. Tier A (≥50%) scores 80-100. Tier B (≥33%) scores 50-79. Tier C (≥15%) scores 20-49.
25%
Safety
Altman Z-Score (40%), Debt/Equity ratio (30%), Interest Coverage EBIT/Interest (30%). Higher = safer balance sheet.
20%
Catalyst
Confirmed EDGAR deal = 85. Special Situations universe = 35 baseline. Merger Arb without confirmed filing = 50.
15%
Arb
Deal-specific: Expected value of the arb position accounting for P(close), spread size, and annualized return. 0 for non-arb names.
10%
Risk
Penalty modifier. Higher score = lower risk. Heavily indebted or distressed companies are downranked even if they appear cheap.
Merger Arbitrage

The P(close) model

Every deal in the live EDGAR feed is assigned a probability of closing. Unlike simple spread analysis, our model accounts for multiple deal-specific factors that meaningfully predict success or failure.

Deal Type Base Rate
Tender offers (SC TO-T) close more reliably than mergers. Base rates: Tender 90%, Cash Merger 86%, Stock Merger 75%.
Base
Spread Size
Wide spreads (>30%) signal market skepticism. Tight spreads (<2%) near deal close get a boost. The market is often right.
Negative if wide
Announcement Age
Fresh announcements (<14 days) carry uncertainty — the market is still digesting. Seasoned deals (>90 days) that haven't broken are more likely to close.
Positive over time
Time to Close
Deals close to their expected close date (<30 days) get a significant boost. Long-horizon deals (>270 days) carry more uncertainty.
Positive if imminent
Dutch Auction Flag
Dutch auction tender offers are structurally more complex and carry an additional penalty in the model.
−0.4 logit
Unknown Acquirer
When the acquirer cannot be identified from the filing, counterparty risk is elevated. The model applies a conservatism penalty.
−0.5 logit
Expected Value
EV = P(close) × spread − (1 − P(close)) × downside. Downside assumed −15% (normal) or −25% (speculative >20% spread).
Output
Kelly Fraction
Optimal position size per Kelly criterion: f = (p×b − q) / b where b is the reward-to-risk ratio. Displayed as a portfolio sizing guide only.
Sizing guide
⚠ Model limitations: The P(close) model is a quantitative heuristic trained on deal type base rates and spread mechanics. It does not account for qualitative factors such as regulatory environment, management character, geopolitical conditions, or specific antitrust risk. All probability outputs should be treated as rough guidance, not predictions. See our Risk Disclosure for full details.

Ready to put the framework
to work?

14-day free trial. No credit card. Full access.

Start Free Trial →