The ticker regime, explained and tested

The ticker regime is computed for each of the 3,056 liquid U.S. stocks we tested, every night, from the stock's own price path. It reuses the four states of the Undercurrent, our market model, so a stock's label answers “which market state does this stock's chart look like?”

From 1997 to 2026 it sorted stocks by risk, steadily: next-month volatility of 28%–30% in Bull and Bear, 43% in Recovery and 72% in Crisis (2010–2022). It did not sort them by direction. As a buy-or-sell rule it trailed a plain 200-day average on the same stocks.

So read it as a risk label for the name in front of you, alongside the market regime for the tape as a whole.

1. 18 features from the stock's daily prices, plus two market ETFs (HYG for credit, TLT for Treasuries): drawdowns, trend, returns, volatility and cross-asset spreads.
2. Scaling. Each feature becomes z = (value − median) / IQR, using the medians and interquartile ranges of the market model's training data.
3. Distance. For each of the four states, add up weight × |z − centroid| over the 18 features. The centroids are the market model's state averages; the weights are Fisher ratios, how well each feature separates the states.
4. Scores. A softmax turns the four distances into scores out of 100: score = exp(−distance / 50), normalised. These are the bull, recovery, bear and crisis scores on the screener.
5. Pick. Crisis wins when it has the top score; otherwise Bear, then Recovery, then Bull.
6. Stickiness. The current label stays unless another state beats its score by more than 15 points, so labels do not flicker day to day. Days in the label are counted as “days in regime”.

The constants were taken from the market model trained in April 2026 and are frozen: retraining the market model does not move them.

Swipe the table for every column.

Volatilities are annualised and multiplied by 0.5 in the code. Two features (HYG's 5-day return, and TLT minus HYG) use only market ETFs, so they are the same for every stock on a given day.

Swipe the table for every column.

• Bull: above its 200-day average, positive 20-day return, shallow drawdowns, volatility below normal.
• Recovery: still deep below its 60-day and 1-year highs, but bouncing hard (5-day return up, volatility high).
• Bear: falling over 5 and 20 days, drawdowns building, Treasuries outpacing the stock and credit stress up; volatility only slightly above normal.
• Crisis: far out on every axis: drawdowns 3 to 5 ranges deep, volatility about 4 ranges above normal.

Next-month volatility

• Bull28%
• Recovery43%
• Bear30%
• Crisis72%

Odds of falling more than 10% at some point in the next month

• Bull13%
• Recovery26%
• Bear15%
• Crisis39%

Swipe the table for every column.

Risk sorts cleanly: Crisis stocks moved 2.6 times as much as Bull stocks. Most of that is volatility persisting, since each state's next-month volatility is close to its last month's. Direction does not sort: about 54% of Bull stocks and 56% of Bear stocks were up a month later. Crisis has the highest average but by far the widest spread, and survivorship flatters it.

Swipe the table for every column.

Close to close, each signal earning the next day, cash at the 13-week T-bill, 5 bps per switch. Profit factor is per round trip, 2010–2022, on the public track record's rules.

Every timing rule trails buy-and-hold here, partly because the test uses today's surviving stocks: names that fell and never came back are missing, which rewards holding through drops. Between the timing rules the comparison is fairer, and the ticker regime is the weakest: it holds stocks 65% of the time yet earns less than the 200-day average. It also stays in volatile Recovery names and steps out of calm Bear ones.

Swipe the table for every column.

Option prices already know the state. Implied volatility rises from about 38% in Bear to 77% in Crisis. The premium over the volatility that followed accounts for 17–22% of implied volatility in every state. The label tells you how much volatility you are buying or selling, not that options are cheap or rich.

• Size by state. A Crisis or Recovery stock has moved 2.6 and 1.5 times as much as a Bull stock. Holding equal risk rather than equal dollars means smaller positions there. The default leverage map sizes Recovery like Bull (3x); the data says Recovery is the second-riskiest state.
• Do not read Bear as sell. Bear stocks were about as calm as Bull stocks and returned as much the next month. Bear describes a drifting-down chart, not a coming drop.
• Expect the drawdowns. About 39% of Crisis stocks fell more than 10% at some point in the next month, against 13% of Bull stocks. Stops and position sizes should leave room for that.
• Options: the state sets the volatility you trade. Selling premium on a Crisis stock collects about 1.8 times the points of a Bull stock, against proportionally larger moves. Defined-risk structures (spreads rather than naked options) cap the loss where moves are largest. Buying options is cheapest on Bull and Bear names because those names move less, not because the options are mispriced.
• Pair it with the market regime. The market regime reads the tape as a whole; the ticker regime reads one name. A Crisis stock in a calm market is a stock-specific story.

• Survivorship. Today's liquid stocks only. Failed companies are missing, which flatters long trades and buy-and-hold and penalises exit rules.
• Market scaling on single stocks. The medians, ranges and centroids come from market data. Stocks are more volatile than the index, so they sit in Recovery and Crisis far more often than the market does.
• Overlapping windows. Next-month statistics overlap day to day; they describe, they do not prove.
• Two years of option data, nightly weekly-expiry IV, which includes earnings premium.
• Not advice. A research note on a public model.