Visit Website →
← Back to blog

Five Minute Graham Number Screen in Oracle Investments

August 31, 2026
Five Minute Graham Number Screen in Oracle Investments

The Graham number is the maximum price Benjamin Graham's formula assigns to a stock, calculated from earnings per share and book value per share using the constant 22.5. It works as a conservative screening cap, not a full valuation, and Oracle Investments users typically pair it with profitability and financial health scores before drawing any conclusions about a stock's price.

TL;DR:

  • The Graham number is most reliable when applied to stable, profitable, asset-heavy companies like banks and industrials, not high-growth or asset-light firms.
  • Negative earnings or book value make the formula unusable, and a stock trading above the Graham number generally indicates overvaluation based on Graham's criteria.
  • The number should be viewed as a ceiling for valuation, not a price target, and requires additional analysis of debt, earnings quality, and industry factors before investing.
  • The original formula is based on historical conservative market standards from Graham's era, which may result in lower limits than modern valuation multiples.
  • Using the Graham number alongside profitability and financial health scores helps focus research on the most promising, undervalued candidates.

Table of Contents

How Do You Calculate the Graham Number?

The formula is short enough to memorize: Graham number = √(22.5 × EPS × BVPS). Multiply earnings per share by book value per share, multiply that product by 22.5, then take the square root. The result is a dollar figure you compare directly against the current share price.

EPS is net income divided by shares outstanding. Graham recommended using a multi-year average EPS rather than the most recent quarter or year, since a single period can be skewed by a one-time gain, a write-off, or a temporary slump. A three-year average smooths that noise and gives a steadier number to work with.

BVPS is shareholders' equity divided by shares outstanding, sometimes called book value per share. Use the most recent balance sheet figure, and if a company carries a lot of goodwill or intangible assets, note that the raw book value may overstate what shareholders would actually recover in a liquidation.

The 22.5 constant is not arbitrary. It comes directly from 15 multiplied by 1.5, the two ceilings Graham built into his defensive investor criteria: a price to earnings ratio no higher than 15, and a price to book ratio no higher than 1.5. Multiply those two limits together and you get 22.5, the combined cap that keeps a stock from being expensive on both counts at once.

That structure creates real math limits:

  • The formula cannot run on negative EPS. A company with a net loss produces a negative number under the square root, which has no real solution.
  • The same applies to negative BVPS, which shows up when liabilities exceed assets.
  • If EPS and BVPS are both positive but their product, times 22.5, still falls below the current price, the stock simply fails the screen. Nothing gets computed wrong. It just tells you the price is too rich by Graham's original yardstick.

When Does the Graham Number Actually Apply?

The formula was built with a specific kind of company in mind: profitable, asset-heavy, and stable. Industrial companies, utilities, banks, and mature manufacturers with steady earnings and a real balance sheet full of tangible equity tend to produce Graham numbers worth trusting.

It tends to fail, or at least mislead, in a few predictable situations:

  • Asset-light businesses. Software and services companies often carry little book value relative to their earnings power, so BVPS understates what the company is actually worth.
  • High-growth stocks. A company priced for 30% annual growth will almost never clear a P/E of 15, even when the growth case is sound.
  • Cyclical earners. A single strong or weak year in commodities, autos, or homebuilding can send EPS, and the Graham number with it, far off its normal range.
  • Financial engineering. Heavy buybacks, one-time asset sales, or aggressive accounting can inflate EPS or BVPS without reflecting a real change in the business.

Once a stock clears the screen, treat that as the start of the work, not the end. Check the margin of safety between price and Graham number, look at debt levels, confirm earnings quality wasn't propped up by a one-off, and check return on equity to see if the profitability is sustainable.

Pro Tip: Never buy on a Graham number alone. It is a filter that narrows a universe of hundreds of stocks down to a shortlist worth researching, not a green light to click "buy."

What Does a Real Graham Number Calculation Look Like?

Here's the full calculation with sample numbers, the kind you'd plug in after pulling a company's financials.

  1. Gather the inputs. Say a company has a three-year average EPS of $1.50 and a current BVPS of $10.00.
  2. Multiply. 22.5 × 1.50 × 10.00 = 337.5.
  3. Take the square root. √337.5 ≈ $18.37. That's the Graham number.
  4. Compare to price. If the stock trades at $15.00, it sits below its Graham number, which flags it as a candidate for further research. If it trades at $22.00, it has already exceeded the cap and fails the screen.
  5. Double-check the underlying ratios. At a $15.00 price, P/E works out to 10 (well under the 15 limit) and P/B works out to 1.5 (right at the limit), so both individual constraints hold along with the combined product. Online calculators can confirm this arithmetic in seconds if you'd rather not do it by hand.

A price below the Graham number doesn't mean undervalued in any final sense — you can organize your personal finances first using this free net worth calculator to see how investing fits your overall plan. It means the stock earns a second look, the kind where you dig into debt, cash flow, and whether that $1.50 EPS is likely to hold up next year.

How Do You Use the Graham Number Inside Oracle Investments?

Running this screen by hand works fine for one stock. It gets tedious fast across a watchlist of twenty. Here's the workflow that actually saves time inside Oracle Investments:

  • Step 1: Pull the inputs. Open a stock's detail page in Oracle and note the EPS and BVPS figures, using averaged data where the app provides it to smooth out one-off distortions.
  • Step 2: Run the math. Compute 22.5 × EPS × BVPS and take the square root, either by hand or with a calculator, and flag any data caveats like negative earnings that make the formula unusable.
  • Step 3: Cross-check with Oracle's scores. A favorable Graham number carries more weight when it's backed by strong profitability, valuation, and financial health scores rather than sitting alone.
  • Step 4: Build an action plan. If a stock clears the screen, dig into recent filings for one-off items, confirm the margin of safety you want before buying, and size the position accordingly.

The point isn't to replace judgment with a formula. It's to spend your limited research time on the handful of stocks that already clear a basic bar, instead of every ticker on the market.

Quick Checklist for a Five-Minute Graham Number Screen

Run this on any stock before deciding whether it deserves a deeper look:

  • Pull the three-year average EPS, the latest BVPS, and the current market price.
  • Confirm P/E is at or under 15 and P/B is at or under 1.5, or that their product doesn't exceed 22.5.
  • Calculate the Graham number and compare it directly against the market price.
  • If the price comes in lower, move to follow-ups: check debt levels, review earnings quality for one-offs, read recent management commentary, and decide on position size before acting.

Where Did the Graham Number Come From?

Benjamin Graham built this formula for a specific reader: the "defensive investor" he described in The Intelligent Investor, someone who wanted rules simple enough to apply without deep financial training. Graham taught at Columbia Business School and mentored Warren Buffett, and his defensive-investor criteria, low P/E, low P/B, steady earnings, became the backbone of what value investing means to most people today.

The formula reflects the market Graham was working in during the mid-20th century, an era of higher interest rates and, generally, lower valuation multiples than markets have carried in recent decades. A P/E of 15 and a P/B of 1.5 were not unusual ceilings then. Today, stocks that clear both bars simultaneously are considerably rarer, particularly outside of financials, utilities, and other asset-heavy sectors.

That history matters for how you should treat the number now. It was never meant as a universal rule for every market environment. Graham himself built in the margin of safety concept precisely because he expected his own numbers to be approximations, not laws. Using the Graham number today means applying a mid-century conservative benchmark to a market that has, on average, priced growth and intangible value far more richly than it did when Graham was writing. That gap is worth remembering every time a modern stock fails the screen by a wide margin: sometimes that's a real warning, and sometimes it's just a sign the market has moved.

How Does the Graham Number Compare to Other Valuation Metrics?

The Graham number isn't competing with the P/E ratio or the P/B ratio so much as combining them. A standalone P/E ratio tells you how much you're paying per dollar of earnings, but says nothing about the balance sheet backing those earnings. A standalone P/B ratio tells you how the market prices net assets, but ignores whether the company is actually profitable. The Graham number forces both conditions to hold at once, which is exactly why it's stricter than either ratio alone.

Discounted cash flow analysis sits at the other end of the complexity spectrum. A DCF model projects years of future free cash flow, discounts them back to the present using a chosen rate, and adds a terminal value, producing an intrinsic value estimate with far more nuance than the Graham number offers. The tradeoff is real: a DCF model needs assumptions about growth rates, margins, and discount rates that are genuinely hard to get right, and small changes in those assumptions can swing the output by 30% or more. The Graham number needs only two numbers you can pull straight from a balance sheet and income statement.

The practical answer isn't to pick one. Use the Graham number to cut a long list of stocks down to a shortlist in minutes, then bring in P/E and P/B individually to understand which constraint is doing the work, and save a full DCF model for the handful of names that survive the first pass. Treating a quick formula and a detailed model as competitors misses the point of why each one exists.

How Does the Graham Number Compare to Other Valuation Metrics? — overview diagram

Does the Graham Number Mean Something Different for Growth vs. Value Stocks?

Applying the Graham number to a value stock and a growth stock produces two very different kinds of information, and confusing the two is where a lot of new investors go wrong.

For a mature, asset-heavy value stock, a Graham number close to or above the current price is meaningful. These are businesses where earnings tend to be stable and book value reflects real plant, equipment, or financial assets, so the formula is working with inputs it was designed for. A bank trading below its Graham number, for instance, is genuinely worth a closer look.

For a growth stock, the formula tends to fail almost automatically, and that failure doesn't mean the stock is overpriced. Running a Graham number screen against a company like this mostly tells you it's a growth stock, information you probably already had.

The better approach for growth names is a framework built for them, like growth at a reasonable price investing, which adjusts the P/E ceiling for the growth rate instead of holding it fixed at 15. The Graham number stays most useful as a value-stock filter, and treating it as a universal grade for every stock in the market is the fastest way to misread it.

What Are the Biggest Mistakes People Make With the Graham Number?

The most common error is treating the output as a price target rather than a ceiling. A Graham number of $18.37 doesn't mean the stock is worth exactly $18.37. It means Graham's conservative criteria stop endorsing the stock above that price. The gap between the current price and the Graham number is a starting point for research, not a promise of upside.

A second frequent mistake is feeding the formula a single quarter's EPS instead of an average. One strong or weak quarter can swing EPS by a wide margin, and since EPS sits directly inside the square root, that swing distorts the entire result. Averaging several years of earnings protects against exactly this kind of noise.

A third mistake is ignoring what happens when the formula breaks. Some investors, faced with negative EPS or negative BVPS, force a calculation anyway or substitute a rough guess. The formula simply isn't built for a business in that position, whether it's a young company burning cash or one that's carrying more liabilities than assets. That's a signal to use a different valuation approach entirely, not a bug to work around.

Finally, plenty of investors run the screen once and stop, treating a favorable Graham number as the finish line rather than the start. It says nothing about debt levels, industry headwinds, management quality, or whether the earnings behind the EPS figure are likely to repeat. The formula narrows a list. It doesn't replace the analysis that comes after.

What Do Real Cases Show About Using the Graham Number?

The Graham number tends to work best exactly where Graham designed it to work: mature, profitable companies during periods when the broader market isn't pricing in extraordinary growth. Bank stocks and industrial names trading at depressed multiples during market downturns have historically been the clearest matches, since their earnings and book values are grounded in tangible, recurring business.

The failures are just as instructive. Investors who ran the Graham number against fast-growing technology companies through the 2010s and 2020s consistently found those stocks trading multiples above their calculated cap, sometimes 3 or 4 times higher. Treating that gap as proof of overvaluation would have meant sitting out some of the market's strongest performers, because the formula was never built to price in the earnings growth those companies were expected to deliver.

The other recurring failure mode shows up with companies carrying inflated book values from goodwill and intangibles left over from acquisitions. A Graham number calculated on that kind of BVPS can look reassuringly low relative to price, right up until an impairment write-down erases a chunk of that book value overnight and the "cheap" stock turns out not to have been cheap at all.

The pattern across both cases points the same direction: the Graham number rewards investors who apply it to the businesses it fits, and punishes anyone who treats a clean pass or fail as the whole analysis. It's a filter with a specific design, not an oracle for every stock on the market.

What Do Real Cases Show About Using the Graham Number? — overview diagram

A Practical Note on Using the Graham Number Well

The Graham number earns its keep as a fast filter, and Oracle's scoring approach leans on the same principle: narrow a huge universe of stocks down to a shortlist worth real research, rather than pretending one formula can rank every company correctly. That's the honest use case, and it's a good one.

The pitfall I see most with new investors is treating a passed screen as a finished decision. A stock trading under its Graham number still needs a look at debt, earnings quality, and whether the business itself is durable. Skipping that step turns a useful filter into a false sense of certainty.

Position sizing and diversification matter more than getting any single Graham number exactly right. Even a well-researched value pick can go wrong, and the formula won't warn you about industry disruption or bad management. Spread the risk, and let the number do the narrow job it was built for.

— Matt

Run Your Own Graham Number Screen in Oracle Investments

Oracle Investments turns a formula that normally takes a spreadsheet and a few minutes per stock into something you can check in seconds, because EPS, BVPS, and Graham's own constraints sit right on the stock detail page alongside Oracle's profitability, valuation, and financial health scores.

Oracleinvestments

That combination is the real advantage over doing this manually: instead of pulling numbers from separate filings and hoping you didn't average the wrong years, you get the inputs and the context for judging them side by side, across a database of over 260 stocks. Scanning a watchlist for names trading below their Graham number, then immediately checking whether the underlying business scores well on profitability and debt, takes minutes instead of an evening with a calculator. If you want to see how quickly a shortlist comes together, try Oracle Investments and run a screen on a few stocks you're already watching.

Sources

Investopedia's Graham number guide covers the practical interpretation and averaging guidance referenced throughout this piece. The Wikipedia entry on the Graham number lays out the formula and its mathematical constraints in more detail. Omni Calculator's Graham number tool lets you run the arithmetic instantly. For related concepts, Oracle's guides on intrinsic value calculation and reading a value rating fill in the surrounding context.