2018issue C086-7
Regime-dependent odds after directional price runs
A directional run is counted from a zero reset and grows only while closes keep the same sign. Next-day odds are updated from that path alone, then the rule must stay with the move, fade toward the mean, or stand aside in a paired book.
- A directional run resets to zero when a close prints opposite the prior day and grows only while later closes keep the same sign.
- Conditional probability treats the next period's odds as contingent on the preceding run, unlike an unconditional probability that treats each day as an independent trial.
- An unbiased coin stays at a 50 percent chance after any streak; a stock is allowed to change those odds from one occurrence to the next, so a must-reverse assumption is not the same as a historical bias after a stated run length.
- Same-day long and short samples, including a dollar-neutral book or a volatility-neutral book, let the procedure stay, fade, or stand aside without relying on the market's overall direction.
One procedure for three actions
A directional run is a sequence of closes that keep the same sign. The count starts from a zero reset when a close prints opposite the prior day, then grows only while later closes keep the same sign.
As an editorial reading, TradersWeek treats each same-direction close as a change of statistical regime rather than another independent trial. The teaching task is to write one procedure that counts the run, updates the next-day odds only from that path, and then forces a three-way choice: stay with the move, fade toward the mean, or stand aside in a paired book. Continuation and reversion are never graded as if they were the same state.
Odds that depend on the path
Conditional probability is the chance of an event given that a prior event has or has not occurred. The next period's odds are therefore treated as contingent on the preceding run rather than as an unconditional probability, which would treat the day as an independent trial. A runs-test checks whether successive same-direction or opposite-direction outcomes behave like independent trials or like a clustered, path-dependent series.
An unbiased coin is presented as remaining a 50 percent chance of either face after any streak. That picture contrasts static independent trials with a stock whose odds are allowed to change from one occurrence to the next.
Assuming a run must reverse because it has already lasted is separated from the case in which a historical sample shows a bias better than 50 percent up or down after a run of a given length. The first belief is the gambler's fallacy: a static process is treated as if it must reverse after a streak even though the next-trial odds have not changed.
Whether a name continues a run or reverses after a stated number of same-direction days is treated as a historical quantity to estimate, not as a fact known in advance.
Continuation or reversion as opposing regimes
Two opposing regimes are stated. A run can keep raising continuation odds until an inflection, or each further day away from the mean can raise reversal odds, so that a six-day run is treated as increasing the chance the seventh day reverses.
Mean reversion is a rule that treats distance from a reference mean as raising the chance of an opposing move rather than further extension. Entering against a move solely because it has already extended is distinguished from estimating the odds of continuation versus mean reversion, including the path that produced an extreme oscillator reading.
A paired book as the stand-aside path
Same-day long and short run samples can be chosen without relying on the market's overall direction. A dollar-neutral book or a volatility-neutral book is described as a way to reduce that market's effect on the combined positions.
Pairs trading is a same-horizon long and short in related instruments intended to isolate relative movement while reducing shared market exposure. Pairing a stock with a related sector fund or market proxy is described as a way to stay more market-neutral while trading the relative spread, and the same construction is said to apply on intraday, overnight, or weekly swing horizons.
Comparing a symbol with a broad market proxy is offered as a check on ordinary versus current co-movement, with the caveat that the co-movement may not persist.
As an editorial reading, standing aside in that paired book is the third action of the same procedure, not a separate study. Stay, fade, and abstention are chosen together on the system holding period so continuation and reversion are never scored as one state.
All readings on this track · 15 readings
- 1986Constructing runs and persistence tests from labeled prices
- 1986Evaluating daily price and volume serial independence windows
- 1986Evaluating advance-decline plus-day runs against chance baselines
- 1986Weekly resamples as a diagnostic filter for statistical windows
- 1988Runs test as a critique of price-series memory
- 1989Evaluating weekday close direction with a counted baseline
- 1989Statistical windows for indicator time parameters
- 1992Channel-height ratios for equity trend evaluation
- 2001A runs test before volatility and expected-value sizing
- 2005Constructing runs-test z-scores for signed return persistence
- 2005Evaluating persistence with runs and autocorrelation
- 2005Weekday FX turning points and close run tests
- 2013Constructing a runs-test turn forecast
- 2017Star rating from slope and swing runs
- 2018Regime-dependent odds after directional price runs