The Quantery blog
Practical notes on testing investment ideas: building screens from raw fundamentals, backtesting without fooling yourself, and reading the numbers straight from the filings. Research, not advice.
-
How to stress-test a backtest exit rule
Hold the entry signal fixed, vary maximum holds and gate-fail exits, then test whether a backtest result depends on its selling rule.
-
How to use a holdout period in a backtest
Freeze your stock-screen rules, test them once on untouched dates, and learn what a holdout can and cannot prove about a backtest.
-
How to test a thesis across market regimes
Learn how to split a backtest into predeclared market regimes, avoid hindsight labels, and tell a conditional edge from a fragile result.
-
How to check data coverage before a backtest
Use Quantery’s per-company coverage record to spot stale filings, missing periods, and short price histories before you trust a backtest.
-
How to read a backtest report
Read a Quantery backtest report from coverage to event study, portfolio return, benchmark, drawdown, and the next test worth running.
-
How to run your first backtest in Quantery
Run a Quantery backtest from an editable thesis, then read its benchmark, event study, portfolio curve, coverage, and caveats.
-
How to test whether a stock score means anything
Run progressively stricter score gates, compare identical backtests, and find out whether more points identify a stronger stock screen.
-
How to use negative controls in a backtest
Add placebo rules and impossible timing checks to a backtest so data leakage, broad exposures, and fragile stories reveal themselves.
-
Strict gates vs score gates in a stock screen
Learn when every criterion should be mandatory, when strengths may offset weaknesses, and how to test both gate designs without moving the thesis.
-
Accounts receivable turnover, explained
Learn how receivables turnover connects revenue to customer collections, where to find both lines in a filing, and how to test a collection rule.
-
How to audit a stock screen result
Audit a Quantery stock screen result from gate to matched rule, then trace each calculated value back to its filing inputs and run date.
-
How to read inventory turnover on a stock screen
Learn what inventory turnover measures, why inventory can outrun sales, and how to build a filing-based inventory discipline screen.
-
How much reporting lag should a backtest use?
Learn what reporting lag prevents, why filing deadlines are the wrong shortcut, and how to test whether a thesis survives a slower data clock.
-
How often should a backtest rebalance?
Choose a rebalance schedule from the speed of your thesis inputs, then test whether the result survives a slower or faster schedule.
-
Is your stock screen just a sector bet?
Test whether a stock screen finds the business trait you intended or merely selects an industry that naturally looks good on the chosen ratio.
-
How to screen after a Fed rate hike
Turn the Fed's latest rate increase into a test of debt service, refinancing exposure, cash generation, and the price paid for resilience.
-
How to turn a filing into a testable thesis
Use a company filing as a research prompt: separate reported facts from a claim, choose a proxy, and test the rule without knowing the outcome.
-
How to compare two screen runs
Compare two Quantery scan runs to see which companies were added, dropped, or still passing, then separate data changes from rule changes.
-
How to choose the right backtest benchmark
Choose a backtest benchmark that matches your screen's investable universe, then use factor checks to explain where excess returns came from.
-
Percentile ranks and deciles: how screens compare stocks
Learn how percentile ranks and deciles compare stocks, why the universe sets the result, and when a screen also needs an absolute threshold.
-
Piotroski F-score: eight signals, one number
Learn how Quantery's eight-signal Piotroski F-score combines filing-based checks, handles missing data, and exposes the signals behind a total.
-
Earnings yield and FCF yield: pricing stocks like bonds
Learn how earnings yield, EBITDA/EV, and FCF yield turn valuation multiples into comparable questions without mixing equity and enterprise value.
-
Enterprise value: what the whole business costs
Enterprise value adds debt to market cap and subtracts cash. Learn what that bridge means, where it breaks, and how to use it in a screen.
-
Return on capital: the quality number screens rank
Learn how return on capital connects operating profit to tangible capital employed, where each input comes from, and how to screen for it.
-
Capex and depreciation: how to read the capital cycle
Learn how capital expenditures and depreciation differ, what each says about a business's capital cycle, and how to put the gap into a thesis.
-
ROE and ROA: taking return on equity apart
Learn what return on equity and return on assets measure, how leverage changes the answer, and how DuPont analysis turns ROE into a testable thesis.
-
How to avoid counting the same signal twice
Learn how overlapping criteria can overweight one idea in an investment thesis, then separate quality, cash backing, leverage, and valuation.
-
When producer prices outrun consumer prices
What CPI and PPI say about margin pressure, where the comparison breaks, and how to test a company's margin defense from its filings.
-
Operating cash flow vs net income: the accruals gap
Why net income and operating cash flow disagree, what a persistent gap between them says about earnings quality, and how to score that gap in a rule.
-
What EBITDA measures, and what it leaves out
EBITDA is operating profit with depreciation added back. Where it comes from in the filings, what the add-back hides, and how a thesis should use it.
-
Can a screen see a dividend cut coming?
Campbell's cut its dividend 36% while free cash flow still covered it 1.4 times. What the filings did show, and how to write that into a testable thesis.
-
Revenue, net income, and margins on the income statement
Where revenue, gross profit and net income sit on a real income statement, what gross margin and net margin each measure, and where the two disagree.
-
Share count and dilution: why per-share numbers drift
Basic vs diluted share counts, where they sit in a filing, how issuance and buybacks move every per-share number, and how to test for dilution.
-
Gross debt vs net debt: how leverage reads on a screen
What total debt and net debt are, where they sit in the filings, when subtracting the cash is fair, and the two ratios a thesis uses to measure leverage.
-
Working capital and the current ratio, explained
What current assets and current liabilities are, how working capital and the current ratio come off a real filing, and where the ratio stops working.
-
When a one-time gain flatters your value screen
Dollar Tree booked $1.31 a share of tariff refunds and Dollar General $0.25. Why one-off cash gains slip past earnings-quality checks and skew trailing screens.
-
How to read a balance sheet: assets, liabilities, equity
The accounting identity behind every balance sheet, what book value really means, why a profitable company can show negative equity, and how a screen reads it.
-
Quantery 1.6.0: the universe is part of the thesis
Release notes for 1.6.0. A new Universes page brings standard index universes (S&P 500, Nasdaq-100, Dow 30) with point-in-time membership, named custom ticker lists, universe-aware backtests, and four new MCP tools. Thesis universe filters now apply on every scan.
-
Trailing vs forward numbers: what a thesis can trust
What trailing twelve months means, how it is built from four filed quarters, and why a screen scores filed numbers instead of forecasts.
-
How the three financial statements fit together
The balance sheet, income statement and cash flow statement answer three questions and lock together at two joints. Here is the map, and where a screen reads them.
-
How to tell if a backtest result is real
Three backtests of the same screen, three different verdicts. How to use perturbation and coverage checks to separate a robust result from a lucky one.
-
Quantery 1.5.0: every number can now show its work
Release notes for 1.5.0. Click any value a scan produced and see the rule that scored it, the inputs behind it, and the filing on SEC EDGAR. Plus clickable run history with two-run compare, a per-company data coverage card, and three new MCP audit tools.
-
How to choose the universe your screen runs on
Exchange lists, market cap floors, sector exclusions, and missing data decide what your stock screen can see. How to make those choices on purpose.
-
Turning an oil price spike into a testable thesis
Energy stocks surged on Middle East supply fears while the EIA forecasts lower oil next year. How to split that headline into claims you can actually test.
-
How to connect Claude to Quantery over MCP
Turn on Quantery's local MCP server and connect Claude Code or Claude Desktop, so your assistant can build theses, run screens, and backtest on your machine.
-
When capex outruns cash flow: a screen for the AI buildout
Alphabet borrowed $25 billion the same summer its free cash flow went negative. How to turn debt-funded capex into a screen you can test.
-
Net current asset value: pricing the business at zero
What net current asset value is, how to compute it from the balance sheet, what the net-net record shows, and why a live screen finds almost none today.
-
Your first scan in 10 minutes
Install Quantery, activate a trial key, run a proven starting-point screen, and read the survivors table. Then change one rule and make the screen yours.
-
How to build a thesis worth testing
From a hunch to testable criteria: choosing measurable proxies, killing vague rules, and one worked example built from raw fundamentals.
-
What makes a backtest honest
Point-in-time data, survivorship bias, look-ahead, excluded costs, and walk-forward testing: what a good backtest result does and does not tell you.
-
Free cash flow: the number that's hardest to fake
What free cash flow is, how it's derived from the actual filings, why it diverges from earnings, and how to express an FCF screen from raw fundamentals.
Follow along via RSS.