Methodology

About

Touchstone screens major-index US equities by Mark Minervini's SEPA: stage-2 trends, the market's strongest relative strength, bases that tighten into a pivot, and earnings and sales that grow faster quarter on quarter. Every price measure is computed from bars up to the day screened, and every fundamental from the company's own SEC filings as they stood on the day they were filed. Every pick is tracked afterwards at 1, 4, 13 and 26 weeks.

Price momentum persists for three to twelve months (Jegadeesh and Titman, 1993); nearness to the 52-week high predicts returns beyond it (George and Hwang, 2004); prices under-react to earnings news for a quarter or more (Bernard and Thomas, 1989). Momentum crashes in the rebound after a market fall (Daniel and Moskowitz, 2016), so every pick records the market's direction on the day it was made.

Beaten-down companies have one screen of their own. Distressed stocks as a class have underperformed since 1981 (Campbell, Hilscher and Szilagyi, 2008); within an already-cheap population, financial strength separates winners from losers by roughly 15 points a year (Piotroski, 2000). Repair to Leader asks for both, and for the price to have turned.

The three screens

Each screen is ranked on its own. The three are never combined into one score: they carry different base rates and different holding periods, and a blended number lets the weakest borrow credibility from the strongest. A company with growth, a pattern or accounts that cannot be measured is not ranked low; it is not ranked.

Stage 2 Leader: Companies in a confirmed stage-2 uptrend — all eight of Mark Minervini's trend-template conditions — with relative strength among the market's best, and earnings and sales growing faster than a year earlier and faster than the quarter before. Momentum persists for months, stocks near their highs keep outperforming, and prices under-react to earnings news for a quarter or more.
Contraction Setup: Stage-2 companies pausing in an orderly base: each pullback shallower than the last, volume drying up as the swings tighten, and the price within 5% of the pivot — the top of the last contraction. The pivot is a level on the chart, not a recommendation.
Repair to Leader: Companies that fell 35% or more within the year while their accounts stayed sound — a Piotroski F-score of 6 or more, outside Altman's distress zone — priced cheaper than the median company, and now turning: above the 50-day average, the 200-day rising, at least five of the eight trend conditions. Distressed stocks as a class underperform; inside a cheap, beaten-down population, financial strength separates the winners from the losers.

Picks made before Oct 18, 2026 came from four earlier screens — Quality Drawdown, Turnaround, Event Shock, Trend Reversal - and stay in the record under those names.

Coverage

Constituents of the S&P 500 and Nasdaq-100 by published membership; the Dow 30 as a subset of the S&P 500; and a Russell 3000 sleeve reconstructed as the 3,000 largest US operating companies by market value, because FTSE Russell does not publish its list. Companies that leave an index remain in the historical record, so the outcome record is not survivorship-biased by construction.

Insurers, asset managers and REITs present no classified balance sheet or gross profit and cannot be scored on the Piotroski tests. Foreign private issuers reporting solely in a currency other than the US dollar are excluded rather than mixed with dollar prices.

Sources

Financial statements: SEC EDGAR XBRL company facts. Each fact carries its filing date and is used only from that date forward; nothing is joined on period end. Every figure links to the filing it came from.
Prices: End-of-day bars, split- and dividend-adjusted. Refreshed nightly.
Disclosures: 8-K current reports with SEC item codes; 6-K reports for foreign private issuers. Refreshed nightly.
Earnings dates: Nasdaq's published calendar, refreshed nightly. A date without a confirmed time is provisional.
Option chains: CBOE end-of-day delayed chains for S&P 500 constituents: settled volume and open interest, snapshotted nightly. No trade tape.
Market breadth: Counted over the screened universe, not all US listings. Not comparable to a market-wide breadth service.

Definitions

Quality and valuation

Piotroski F-score: Nine accounting tests of profitability, leverage and efficiency, published by Joseph Piotroski in 2000. Among beaten-down stocks, the high scorers have historically beaten the market and the low scorers trailed it.
Altman Z-score: A bankruptcy-prediction model from 1968, accurate to roughly 80–90% a year ahead. Used here only to exclude companies, never to rank them.
Earnings yield (EBIT / enterprise value): Operating profit measured against the whole cost of buying the business — its shares plus its debts, less its cash. A higher figure means more earnings for the price. Around 3.4% is typical; above 9% is in the cheapest tenth.
Valuation rank: How cheap this company is against every company screened, averaging its earnings yield, free cash yield and book value. Ranked against the whole market rather than against other fallen shares — cheap relative to other wreckage is not cheap.
Drawdown: How far the share price sits below its highest close of the last 12 months.
The company's own move: Of everything this share price did over the last year, the part that is not what the market did. A company down 57% while the market rose 20% has fallen 77 points on its own — nothing about that is being sold with everything else. Straight subtraction, not risk-adjusted, so a naturally volatile company shows a larger own move than a risk model would attribute to it.
Relative strength: Performance against the S&P 500 over three months compared with twelve. A positive figure means the company is losing ground more slowly than it was.

Trade setup

Average true range, as a percentage of price: The size of a normal day's move: the 14-day average of each day's full range, gaps from the previous close included, over the last close. What a position is sized by and a stop is set by - a stop two ATRs away survives an ordinary day.
Relative strength rank: Where this company's weighted return over the last year ranks among every company screened, 1 to 99: the latest quarter counts double each earlier one, as IBD's rating does. Momentum is one of the best-documented effects in stock returns; the latest month on its own tends to reverse.
Trend template: How many of Mark Minervini's eight conditions for a stage-2 uptrend hold: price above the 50, 150 and 200-day averages in that order, the 200-day rising, at least 30% above the year's low, within 25% of its high, and an RS rank of 70 or more. A filter for setups, not a forecast.
Volatility contraction: Mark Minervini's base pattern: after a high, a series of pullbacks each shallower than the last - 25%, then 12%, then 5% - as selling is absorbed. Shown with each contraction's depth, oldest first. Measured on the last six months of split-adjusted highs and lows; a dip under 2% is a pause, not a contraction.
Pivot: The top of the base's last, tightest contraction: the level the price has to clear for the base to be resolved upward. A level on the chart, not a recommendation.
Volume dry-up: The last 10 sessions' average volume over the last 50's. Below 1, fewer shares are changing hands as the swings tighten - supply thinning.
Relative volume: The latest session's volume over its average for the 50 sessions before. Above 1.5 a move is being made on unusual participation.
Up/down volume: Over 50 sessions, the volume traded on days the stock rose divided by the volume on days it fell. Above 1 more shares changed hands on the way up: accumulation.
Move in ATRs: The move over the period divided by the stock's average daily range (ATR%): how many normal days' worth of movement it was. A week's move of one or two ATRs is routine; five or more is a repricing.
Two ATRs: Twice the average daily range, as a share of the price: the distance a position sized on volatility is commonly given. A measurement, not a stop to place.

Stock table

Sector: One of eleven GICS-style sectors, assigned from the company's SEC industry code by Touchstone's own map. GICS itself is licensed, and SEC codes describe what a company makes rather than how it earns, so a few companies sit elsewhere than S&P places them - Alphabet and Meta under Information Technology, Visa and Mastercard under Industrials. Blank where no code is filed.
Price change: The change in the adjusted closing price over one week, one month and one year, measured back from the latest close. Adjusted for splits and dividends. Blank where the company has no close that far back.
Price-to-earnings (P/E): Market value divided by net income for the latest fiscal year as filed. Blank for a company that lost money: a negative P/E is not cheap, it is undefined.
Price/earnings-to-growth (PEG): P/E divided by the growth in earnings per share from the prior fiscal year, in percent. Around 1 is conventionally read as growth priced at par. Blank unless both years were profitable and earnings per share grew.
Return on equity (ROE): Net income for the latest fiscal year divided by stockholders' equity, averaged over the year's two balance sheets where both are filed. Blank when equity is negative, which turns a loss into an apparent return.
Price-to-book (P/B): Market value divided by stockholders' equity at the latest fiscal year-end. Blank when equity is negative, as it often is after years of buybacks.
Debt-to-equity (D/E): Long-term debt divided by stockholders' equity at the latest fiscal year-end. Blank when either is not filed or equity is negative - an unreported debt figure is not evidence of no debt.
EPS growth: The latest quarter's earnings per diluted share against the same quarter a year earlier, from the 10-Qs as filed. Beside it, in points, how much faster or slower that is than the quarter before grew: acceleration. No figure where the year-earlier quarter was a loss - a turnaround has no growth rate.
Sales growth: The latest quarter's revenue against the same quarter a year earlier, and in points how that compares with the quarter before's growth.
Leader criteria: 8/8 trend template · RS 80+ · EPS growth 20%+ y/y · sales growing · ROE 17%+ · not in distress. Latest nightly bars and the filings as filed; a figure that could not be computed fails.

Market

Market conditions: A label for what breadth shows, from the up/down ratio and the share of companies above their 40-day average. Descriptive, not predictive: the thresholds are round numbers, not a fitted result, and nothing here enters a company's rank.
Market direction: Whether the index is in an uptrend, under pressure or in correction, read from its own closes and volume: a correction is a 10% fall from its high, and ends only on a follow-through day - from the fourth day of a rally, a 1.25% gain on higher volume. Under pressure is five or more distribution days (falls of 0.2% on higher volume) in 25 sessions, or a close under the 50-day average.
Companies above their 40-day average: What share of the screened universe is trading above its own average price of the last 40 trading days. The most-watched single breadth figure. Readings under a fifth mark indiscriminate selling; over half is a market where most things are working.
Up/down ratio: Companies that rose 4% or more in a day, divided by those that fell 4% or more, summed over the window. Above 1 means more companies had a strong day than a bad one. A 4% move is roughly a three-sigma day, so a company making one has usually had news.
VIX: The volatility the S&P 500's options price over the next 30 days, annualised, and its change in points over the last five sessions.
VIX against VIX3M: One-month implied volatility over three-month. Below 1 the curve slopes up, as it does in calm markets; above 1 the near term is priced as the riskier stretch, as it is in a selloff.
Sectors against the market: Each sector ETF's return over three months and one, less SPY's: where money moved relative to the index. Beside it, the sector's S&P 500 members: how many meet the stock table's leader criteria, the three with the strongest RS, and the share above their 50-day average. Sectors are mapped from SEC industry codes, not GICS.
Sector trend: The 3-month return against SPY, now and four weeks ago. Ahead means more than a point ahead. Leading: ahead and gaining. Fading: ahead but slipping. Improving: behind but gaining. Lagging: behind and slipping.
Above the 50-day: The share of the sector's S&P 500 members closing above their 50-day average: whether the whole sector is moving or a few names are carrying it.

Financials

Revenue: What customers paid the company over its fiscal year, before any cost, as filed in its annual report. The change is against the year before.
Net income: Profit left after every cost, interest and tax for the year. A column below the line is a loss.
Free cash flow: Cash from running the business less what it spent on equipment and buildings: the cash a year actually left behind, which profit alone does not show. Not shown for banks and insurers, whose cash moves with their loans.
Margins: Each as a share of revenue. Gross: left after the direct cost of what was sold. Net: left after every cost. Free cash: free cash flow over revenue. A margin rising over years means the business keeps more of each dollar it sells.Revenue of $100 with $60 left after the cost of goods and $20 after everything is a gross margin of 60% and a net margin of 20%.
Cash and long-term debt: Cash on the balance sheet at the year's end beside debt due after a year. Net cash is cash less that debt; net debt is the shortfall. Debt some filers do not tag shows as a gap, not as zero.$10B of cash and $7B of long-term debt is net cash of $3B; $4B of cash against $9B of debt is net debt of $5B.
Shares outstanding: Shares in issue at the year's end. A rising count means each share owns a smaller slice of the company (stock pay, share sales); a falling one, buybacks. The chart's scale is fitted to the count, so a small change looks large: read the percentage.

Options

Expected move: The move, either way, the option market prices in through the next scheduled report: the at-the-money straddle for the first expiry that contains it, over the price. It includes the days around the report, so it slightly overstates the report alone.A stock at $200 with an expected move of ±7% is priced to end the report somewhere between about $186 and $214.
Usual move: The median, over the company's last four reports, of how far the stock moved either way from the close before the filing to the close of the session after it.Moves of +3%, -8%, +5% and -4% after the last four reports make a usual move of ±4.5%: the middle of the four sizes, either way.
Expected against usual: The move the options price through this report over the move the last four reports delivered. Above 1 the event is priced bigger than the company's own record; below 1, smaller. A price, not a forecast: it says what protection through the report costs, not which way the stock will go.Expected ±7% against a usual ±5% is 1.4×: options price this report bigger than the company's last four.
Expected 1-month range: The at-the-money straddle for the expiry nearest a month out, as a share of the price: the move, either way, the option market is pricing by then. The range is the price plus and minus that move; the stock ends inside it a little more often than not.At $200 with an expected move of ±8% by next month's expiry, the range is about $184 to $216.
Swings expected vs recent: The daily swings the options are priced for (implied volatility) against the swings the stock actually had over the last 20 sessions (realised volatility). Bigger means options charge for more movement than the stock has been delivering; smaller, less. On a contract row, that contract's own implied volatility over the stock's realised.Options priced for 40% a year on a stock that swung 28% a year over the last 20 sessions: 1.4×, bigger.
Next month vs later: At-the-money implied volatility for 8-45 day expiries over 46-90 day ones: whether options price the coming month as choppier or calmer than the two after it. Ahead of a report it usually reads choppier, and the extra comes out once the report is out.
Crash protection: What options that pay off in a fall cost against options that pay off in a rise (the 25-delta skew), compared with this company's own normal - the median of its stored nightly readings, once there are 20 - and with five sessions ago. Protection costs a little more than upside on most shares. More than usual means readers of the chain are paying up to insure against a drop; it does not say which way the stock will go.Puts priced at 46% and calls at 40% is a skew of 6 points; if this company's normal is 4, protection costs more than usual.
Open interest added: The change over five sessions in option positions left open overnight, calls and puts. A rise means positions were added. It cannot say whether they were bought or sold, so more calls is not a vote for a rise: a call is as often sold against shares already held.Call open interest going from 7.0m to 7.6m in a week is +0.6m added - by buyers or sellers, the count cannot say.
Crosses the report: The contract expires after the next scheduled report, so its price carries the report's move - and the extra implied volatility that usually comes out once the report is out.
Options activity: The week's most significant option contracts from the nightly end-of-day chain snapshot of the S&P 500, soonest expiry first. A contract counts when its volume reached its open interest - it traded at least every position standing in it - or when the next morning's open interest confirmed the day opened or closed positions. Listed only when it can be traded and held for a swing: 21 to 120 days to expiry, open interest of 1,000 or more, 500 contracts traded, a spread within 5% of mid, on a stock trading 2M shares a day. End of day, not a trade tape: no sweeps or blocks.
Next day: opened, closed or mixed: A large option trade on its own is ambiguous: it may have opened a position, closed one, or been one leg of a spread. Open interest is published overnight, so the next morning settles it. Opened: open interest rose by a good share of the volume traded, so positions were added. Closed: it fell, so positions were taken off. Mixed: it barely moved, as when one side opens while the other closes. A dash: the next morning's count is not in yet. None of the three says whether the contracts were bought or sold.
Change in open interest: How much the standing book at this horizon grew or shrank against yesterday. Positions, not trades: volume says something changed hands, open interest says whether anything was added. Because a bucket is defined by days to expiry, its contents roll forward with the calendar — this measures the book at that horizon, not a fixed set of contracts.
Volume against open interest: The day's contracts traded, divided by the positions already standing. Above 1 the day's trading exceeded every position in existence, which cannot happen without something new — though only tomorrow's open interest proves it opened rather than closed.
Bid-ask spread: The gap between the bid and the ask as a share of the midpoint: what a round trip costs before the position has moved at all.
Priced vs rest of chain: The contract's implied volatility less the at-the-money reading for the same company, in points: how this strike is priced against the rest of the chain.
Price against value: The midpoint against the vendor's theoretical value for the contract. Positive, the market is paying more than the model; negative, less.
Move to break even: How far the stock must move by expiry for a buyer at the midpoint to get the premium back: strike plus premium for a call, strike less premium for a put.
Days to earnings: Calendar days to the company's next scheduled report. A hold that crosses it carries the event, and the implied volatility that falls after it.
Puts against calls, by premium: Money spent on puts divided by money spent on calls. Measured in dollars rather than contracts on purpose: counting contracts treats a hundred cheap lottery tickets as more meaningful than one large position, which is backwards.
Implied volatility: How much movement the option market is charging for, at the money and a month or two out. It is a price, not a forecast — high implied volatility means protection is expensive, which is as often a reason to sell it as to buy it.
Volatility skew: What 25-delta puts cost against 25-delta calls, in volatility points. Positive is the normal state for shares: downside protection costs more than upside participation. The reading worth noticing is how far it sits from that company's own normal, and a negative number means calls are bid over puts.
Net gamma exposure: Dollar gamma across the chain per 1% move, assuming dealers are long the calls and short the puts the public trades. Positive, their hedging leans against a move; negative, it leans with it. An assumption about who holds what, not a measurement.

Disclosures

Nothing on this site is investment advice or a recommendation to buy or sell any security. Rankings are within a screen and are not comparable across screens. Prices and option data are end-of-day and delayed. The outcome of each pick is recorded on its company page; no aggregate performance claim is made for the screen.