Important: an indicator is a rule for organizing historical prices, not a prediction. TickRun tests one long position at a time and applies a signal to the following session’s close-to-close return. Results do not include taxes, bid–ask spreads, slippage, or market impact.
How to read these strategies
TickRun converts each indicator into a buy event and a sell event. A buy opens a 100% long position; a sell closes it. While a position is open, repeated buy signals are ignored. While no position is open, sell signals are ignored.
A crossover requires the indicator to move from one side of a level to the other. Optional signal delay and minimum holding-period controls are then applied. Transaction costs are deducted on entry and exit.
Strategy index
Momentum and oscillator strategies
1. Relative Strength Index (RSI)
RSI compares recent upward and downward price changes on a 0–100 scale. TickRun buys when RSI crosses below the lower threshold and sells when it crosses above the upper threshold, testing a mean-reversion interpretation: weakness triggers entry and later strength triggers exit.
Defaults: 14-session lookback, 30 lower threshold, 70 upper threshold.
2. Moving Average Convergence Divergence (MACD)
MACD subtracts a slow exponential moving average from a fast one. TickRun buys when the MACD line crosses above its signal EMA and sells when it crosses below. It follows changes in trend momentum but can switch repeatedly in sideways markets.
Defaults: 12-session fast EMA, 26-session slow EMA, 9-session signal EMA.
3. Stochastic Oscillator
The oscillator places the close inside its recent high–low range. TickRun buys when %K crosses above its smoothed %D line while still below the lower threshold. It sells when %K crosses below %D while above the upper threshold.
Defaults: 14-session %K, 3-session %D, thresholds 20 and 80.
4. Williams %R
Williams %R expresses where the close sits within the recent range on a scale from −100 to 0. TickRun buys when it crosses upward through −80 and sells when it crosses downward through −20, looking for recovery from an oversold region and reversal from an overbought region.
Defaults: 14-session lookback, entry −80, exit −20.
5. Commodity Channel Index (CCI)
CCI measures how far typical price has moved from its recent average relative to mean deviation. TickRun buys on an upward cross of −100 and sells on a downward cross of 100.
Defaults: 20-session lookback, entry −100, exit 100.
6. Rate of Change (ROC)
ROC measures the percentage price change over a selected lookback. TickRun buys when ROC crosses above the entry level and sells when it crosses below the exit level. With both levels at zero, it follows positive versus negative momentum.
Defaults: 12-session lookback, entry 0%, exit 0%.
7. Awesome Oscillator
This oscillator subtracts a slow simple average of median price from a fast one. TickRun buys above the entry level and sells below the exit level. A zero threshold treats positive short-term momentum as bullish.
Defaults: 5-session fast average, 34-session slow average, zero entry and exit.
8. TRIX
TRIX applies three exponential smoothings to price and measures the one-session rate of change of the result. TickRun buys when TRIX crosses above its entry level and sells when it crosses below its exit level.
Defaults: 14-session smoothing, zero entry and exit.
9. Ultimate Oscillator
The Ultimate Oscillator blends buying pressure across short, medium, and long windows. TickRun buys when it crosses upward through the lower level and sells when it crosses downward through the upper level.
Defaults: 7, 14, and 28 sessions; entry 30; exit 70.
10. Detrended Price Oscillator (DPO)
DPO removes a shifted moving average from price to emphasize cycles rather than the long trend. TickRun buys when DPO crosses above the entry level and sells when it crosses below the exit level.
Defaults: 20-session lookback, zero entry and exit.
11. True Strength Index (TSI)
TSI double-smooths price changes and divides them by double-smoothed absolute changes. TickRun buys on an upward cross of the entry level and sells on a downward cross of the exit level.
Defaults: 13-session fast smoothing, 25-session slow smoothing, zero entry and exit.
Trend and moving-average strategies
12. SMA Crossover
Two simple moving averages represent shorter- and longer-term price trends. TickRun buys when the fast SMA crosses above the slow SMA and sells when it crosses below. Simple averages react evenly to every observation in their windows.
Defaults: 20-session fast SMA, 50-session slow SMA.
13. EMA Crossover
This uses the same crossover rule as the SMA strategy, but exponential averages weight recent observations more heavily. That makes signals more responsive and potentially more sensitive to noise.
Defaults: 20-session fast EMA, 50-session slow EMA.
14. ADX / Directional Movement
ADX estimates trend strength while +DI and −DI estimate direction. TickRun buys when +DI crosses above −DI and ADX exceeds the minimum strength threshold. It sells when +DI crosses below −DI.
Defaults: 14-session lookback, minimum ADX 25.
15. Ichimoku Cloud
Ichimoku combines recent high–low midpoints into conversion, base, and cloud boundaries. TickRun buys when closing price crosses above the cloud’s upper boundary and sells when it crosses below the lower boundary.
Defaults: conversion 9, base 26, Span B 52 sessions.
16. Parabolic SAR
Parabolic SAR creates a trailing level that accelerates as a trend develops. TickRun buys when price crosses above SAR and sells when price crosses below it. The acceleration settings control how quickly the level approaches price.
Defaults: acceleration step 0.02, maximum 0.20.
17. Aroon
Aroon Up and Aroon Down measure how recently the period’s high and low occurred. TickRun buys when Aroon Up crosses above Aroon Down while Up meets its minimum. It sells on the reverse cross while Down meets its minimum.
Defaults: 25-session lookback; minimum Up 50 and Down 50.
18. Know Sure Thing (KST)
KST combines four smoothed rate-of-change components with increasing weights. TickRun buys when KST crosses above its signal average and sells when it crosses below.
Defaults: fixed KST components and a 9-session signal smoothing.
19. Double Exponential Moving Average (DEMA)
DEMA combines a single and double EMA to reduce lag. TickRun buys when price crosses above DEMA and sells when it crosses below.
Defaults: 20-session window.
20. Triple Exponential Moving Average (TEMA)
TEMA combines single, double, and triple exponential averages to reduce lag further. TickRun buys when price crosses above TEMA and sells when it crosses below.
Defaults: 20-session window.
21. Vortex Indicator
The positive and negative Vortex lines compare directional movement with true range. TickRun buys when the positive line crosses above the negative line and sells on the reverse cross.
Defaults: 14-session lookback.
22. Zero-Lag Exponential Moving Average (ZLEMA)
ZLEMA adjusts the price input before applying an EMA to reduce apparent delay. TickRun buys when price crosses above ZLEMA and sells when it crosses below.
Defaults: 20-session window.
Volatility and channel strategies
23. Bollinger Bands
Bollinger Bands place upper and lower standard-deviation envelopes around a simple moving average. TickRun buys when price crosses below the lower band and sells when it crosses above the upper band, using a mean-reversion rule.
Defaults: 20-session average, two standard deviations.
24. Donchian Channel
The Donchian boundaries are the prior period’s highest high and lowest low. TickRun buys when price breaks above the previous upper boundary and sells when it breaks below the previous lower boundary.
Defaults: 20-session channel.
25. Keltner Channel
A Keltner Channel surrounds an EMA with Average True Range envelopes. TickRun’s mean-reversion version buys when price crosses above the lower channel after being below it and sells when price crosses below the upper channel after being above it.
Defaults: 20-session center EMA, 10-session ATR, multiplier 2.
26. Average True Range (ATR) breakout
ATR measures range-based volatility. TickRun builds ATR envelopes around an EMA, then buys when price crosses above the upper envelope and sells when it crosses below the lower envelope. Unlike Keltner’s rule here, this is a breakout strategy.
Defaults: 20-session center EMA, 10-session ATR, multiplier 2.
Price-and-volume strategies
27. Money Flow Index (MFI)
MFI combines typical price and volume into a bounded oscillator. TickRun buys when MFI crosses above the lower threshold and sells when it crosses below the upper threshold, looking for recovery from weak money flow and reversal from strong money flow.
Defaults: 14-session lookback, thresholds 30 and 70.
28. On-Balance Volume (OBV)
OBV cumulatively adds volume on non-declining sessions and subtracts it on declining sessions. TickRun buys when OBV crosses above its signal EMA and sells when it crosses below.
Defaults: 20-session signal EMA.
29. Volume-Weighted Average Price
TickRun uses a rolling VWAP based on typical price and daily volume. It buys when closing price crosses above VWAP and sells when price crosses below it. This is a rolling daily-data interpretation, not an intraday institutional VWAP.
Defaults: 14-session rolling window.
30. Chaikin Money Flow (CMF)
CMF estimates accumulation or distribution from the close’s location within each range, weighted by volume. TickRun buys when CMF crosses above the entry level and sells when it crosses below the exit level.
Defaults: 20-session lookback, zero entry and exit.
31. Accumulation/Distribution Index (ADI)
ADI accumulates range-positioned volume. TickRun compares the cumulative line with its signal EMA, buying on an upward crossover and selling on a downward crossover.
Defaults: 20-session signal EMA.
32. Ease of Movement
Ease of Movement relates movement in the high–low midpoint to volume and trading range. TickRun smooths it, buys when it crosses above the entry level, and sells when it crosses below the exit level.
Defaults: 14-session smoothing, zero entry and exit.
33. Force Index
Force Index multiplies price change by volume, then smooths the result. TickRun buys when it crosses above the entry level and sells when it crosses below the exit level.
Defaults: 14-session EMA, zero entry and exit.
34. Negative Volume Index (NVI)
NVI changes only on sessions when volume declines from the previous session. TickRun buys when NVI crosses above its signal EMA and sells when it crosses below.
Defaults: 20-session signal EMA.
35. Volume–Price Trend (VPT)
VPT cumulatively adds volume multiplied by percentage price change. TickRun buys when VPT crosses above its signal EMA and sells when it crosses below.
Defaults: 20-session signal EMA.
Experimental strategies
36. Moon Phases
This deliberately experimental calendar rule divides each lunation into ten numbered stages. Stage 0 is New Moon, stage 5 is Full Moon, and stages 1–4 and 6–9 describe the intervening waxing and waning portions. Choose one Buy phase and a different Sell phase. TickRun holds continuously between them. The indicator plot steps through stages 0–9 on a numeric 0–10 y-axis; 10 is the cycle endpoint before it wraps to 0.
The saved ten-year database contains exact New Moon, First Quarter, Full Moon, and Last Quarter timestamps from the U.S. Naval Observatory in Universal Time. USNO does not publish ten separate event instants. TickRun assigns the four anchors coordinates 0, 2.5, 5, and 7.5, interpolates between consecutive timestamps, and rounds to the nearest numbered stage. The intermediate subdivisions are therefore an explicit TickRun research convention, not independently measured events.
The phase is sampled at 00:00 UTC for each date-only stock row, so an event later that day cannot affect the same row. If a weekend skips a stage, the next session still detects that the selected boundary was crossed. The following close-to-close return is the first return attributed to the updated position.
This rule has no established causal market mechanism and should be treated as a falsifiable research example, not evidence that lunar cycles predict returns. Multiple-testing and overfitting risks are especially important when comparing stage pairs.
Defaults: Buy phase 0 (New Moon); Sell phase 5 (Full Moon).
37. Weather in Wall Street
This experimental rule buys after one completed Manhattan weather category and sells after another. The five choices are Sunny / clear, Cloudy / dry, Rain / drizzle, Snow / ice, and Fog / haze. The indicator chart shows the category available to the strategy for each market session.
The information comes from the U.S. National Oceanic and Atmospheric Administration’s National Centers for Environmental Information, specifically Local Climatological Data Version 2 for station USW00094728 — NY City Central Park. The station is at 40.77898, −73.96925, approximately 8.8 km (5.5 miles) in a straight line from the New York Stock Exchange at 11 Wall Street. It provides a stable Manhattan record, but it is a proxy: localized rain and clouds can differ in Lower Manhattan.
TickRun derives precipitation categories from NOAA daily precipitation, snowfall, and weather codes. On dry days, hourly sky observations from 09:00 through 16:59 NOAA local standard time distinguish sunny/mostly clear from cloudy. A dry day needs at least four sky observations and at least half must report clear, sky clear, or few clouds to be called Sunny / clear; insufficient sky coverage falls conservatively into Cloudy / dry. Snow/ice takes precedence over rain, rain over fog/haze, and then the dry sky rule applies.
Report-day bias is explicitly prevented. A complete daily summary includes observations that were not known at the market open and may include conditions after the close. Therefore weather dated on day D can never affect market day D. TickRun uses only the newest completed weather date strictly before a session and maps it to the first later trading session. Monday normally uses Sunday’s report. If NOAA’s newest report is delayed, the old report may remain visible but is not emitted repeatedly as a new signal. TickRun’s usual convention then applies the resulting position to the following close-to-close return.
Weather has no established causal relationship with broad stock returns. The strategy should be treated as a research exercise, and its relatively small snow and fog samples make optimized results especially vulnerable to chance and multiple testing.
Defaults: Buy Sunny / clear; Sell Rain / drizzle.
Calendar strategies
These rules schedule a nominal calendar boundary, then execute its signal on the first saved market session on or after that boundary. This makes weekends, exchange holidays, and missing calendar dates explicit instead of silently dropping them. Buy and sell values must differ. At the beginning of a dataset, TickRun reconstructs whether the most recent boundary was a buy or sell boundary, so the initial position does not depend on the file starting at a convenient date.
38. Day of the Week
Choose Monday through Friday for the Buy day and a different trading weekday for the Sell day. Saturday and Sunday are not offered. If the selected weekday is absent from the saved data because of an exchange holiday, its signal maps forward to the next available session. For example, a Monday signal appears on Tuesday when Monday is a market holiday. The stepped indicator reports the actual session weekday from Monday through Friday.
This strategy can test recurring weekly holding windows, but apparent weekday effects can be unstable and sensitive to transaction costs, market regime, and the exact sample period.
Defaults: Buy Monday; Sell Friday.
39. Day of the Month
Choose calendar days 1–31 for Buy day and Sell day. Each target maps to the first saved session on or after it. When a selected number does not exist in a month—such as day 31 in February—the nominal target becomes the first calendar day of the following month, then moves forward again if that day is not a session. This exact rule prevents short months from disappearing from the test.
The strategy is suitable for examining turn-of-month and mid-month hypotheses. Results can be distorted by a small number of month boundaries, distributions, or large overnight moves, so inspect individual trades as well as the aggregate return.
Defaults: Buy day 1; Sell day 15.
40. Month of the Year
Choose a Buy month and a different Sell month. The signal is scheduled for the first calendar day of the selected month and maps to the first saved session on or after that date. The position may cross the year boundary; with the default settings, TickRun enters in November and exits in May.
This rule can test broad seasonal windows, but ten years of data contains only about ten observations of each annual boundary. Treat results as a small sample and check whether performance depends on one exceptional year.
Defaults: Buy November; Sell May.
41. Daily Geomagnetic Activity (Ap)
This experimental rule uses the planetary Ap index as one daily measure of disturbances in Earth's magnetic field. It enters when the preceding completed UTC day's Ap value is at or below the buy threshold and exits when it is at or above the sell threshold. Values between the thresholds retain the existing position.
The source is the GFZ German Research Centre for Geosciences. Kp is measured in eight three-hour intervals and is quasi-logarithmic, so directly averaging Kp would be misleading. GFZ converts Kp to the linear planetary ap scale and publishes Ap as the mean of the day's eight ap values. TickRun stores GFZ's published Ap value and whether it is definitive or preliminary.
A session dated D uses only Ap from UTC day D−1. Missing days produce no signal and are not forward-filled. TickRun then applies the resulting position to the following close-to-close return. Preliminary recent values can change when GFZ publishes definitive values.
No established causal relationship links planetary geomagnetic activity to broad stock returns. Treat results as exploratory, account for multiple testing, and compare performance across unseen periods and assets.
Defaults: Buy at or below Ap 7; sell at or above Ap 30.
Ap direction: Quiet conditions buys at or below the lower threshold and sells at or above the upper threshold. Disturbed conditions reverses the rule, buying at or above the upper threshold and selling at or below the lower threshold.
42. Ocean Tide Gauge Heights
This experimental rule uses the previous completed UTC day's mean observed water level at NOAA station 8518750, The Battery in Lower Manhattan. In lower-water mode it buys at or below the lower threshold and sells at or above the upper threshold. Higher-water mode reverses those actions. Values between the thresholds retain the existing position.
TickRun aggregates NOAA CO-OPS 6-minute observations in metres relative to Mean Lower Low Water (MLLW). A daily value is accepted only when at least 216 of the expected 240 observations are present. The saved record also retains the minimum, maximum, range, sample count, and whether its source observations are preliminary or verified.
A market session dated D uses only the exact tide value from UTC day D−1. Missing or incomplete days produce no signal and are not forward-filled. NOAA can revise preliminary readings during verification, so refreshing the snapshot may change recent results.
No established causal mechanism makes local tide height a predictor of broad stock returns. Treat the strategy strictly as an exploratory multiple-testing experiment and validate any apparent result outside the optimized sample.
Defaults: Lower-water mode; buy at or below 0.65 m; sell at or above 0.90 m MLLW.
43. Commercial Aviation Flight Counts
This experimental rule measures U.S. domestic commercial-flight activity. TickRun counts completed flights in the Bureau of Transportation Statistics Reporting Carrier On-Time Performance table, calculates the average per calendar day for each source month, and divides it by the mean of the preceding 12 monthly daily averages. A ratio of 1.0 means activity matched its trailing-year baseline.
Lower-activity mode buys at or below the lower ratio and sells at or above the upper ratio. Higher-activity mode reverses those actions. Values between the thresholds retain the existing position.
Publication timing is part of the signal. BTS releases flight-level data after the month being measured. Every saved observation includes the public timestamp of its official BTS file. A market session can use that month only when its date is strictly later than the recorded release date. If files are revised or republished later, TickRun keeps the conservative later timestamp, and an older source month cannot displace a newer month already available. This prevents future flight records from leaking backward into the backtest.
The reporting-carrier population and coverage rules can change, and pandemic-era traffic is an exceptional structural break. No established causal relationship makes flight activity a predictor of stock returns. Treat optimized results as exploratory and subject to multiple-testing bias.
Defaults: Lower-activity mode; buy at or below 0.90; sell at or above 1.10.
44. Friday the 13th
This experimental calendar rule tests the superstition that Friday the 13th is associated with bad luck. Choose an entry from 0 through 7 calendar days before each Friday the 13th and an exit from 0 through 7 calendar days after it. The default holds from seven days before until seven days after.
Each nominal date maps to the first saved market session on or after it, so weekends, exchange holidays, and missing rows do not silently remove a signal. The indicator shows signed calendar days from the nearest Friday the 13th: negative before the date, zero on it, and positive afterward. As elsewhere in TickRun, a signal changes the position at that session's close and the new position first affects the following close-to-close return.
The 0/0 combination is rejected because daily close data cannot honestly represent a buy and a sell at the same close. Friday-the-13th dates are fully determined by the calendar, so the rule uses no future market observations and introduces no data-publication look-ahead. The superstition has no established predictive mechanism; sparse events and optimization across offsets make chance findings particularly likely.
Defaults: Buy 7 calendar days before; sell 7 calendar days after.
45. Sunspot Momentum
This experimental rule measures how far the current daily sunspot number is above or below its recent level. TickRun compares the current SWPC sunspot number with the fixed average of the preceding seven calendar days, excluding the current day, and calculates 100 × (current − previous average) / (current + previous average). This symmetric percentage stays between −100 and +100; if both values are zero, momentum is zero.
Rising-activity mode buys at or above the high momentum threshold and sells at or below the low threshold. Falling-activity mode reverses those actions. Values between the thresholds retain the current position. The preceding average requires complete daily reports; missing dates are neither interpolated nor forward-filled.
The source is NOAA/SWPC's preliminary daily sunspot number in the NCEI archive of Solar and Geophysical Activity Summary reports. TickRun preserves each report's exact observation date, Issued timestamp, and source URL. A report is eligible only for market dates strictly later than its UTC issue date, and the changed position affects the following close-to-close return.
Solar activity has an approximately 11-year cycle and shorter rotational structure, but no established causal mechanism makes this momentum measure a predictor of stock returns. The large optimization space makes out-of-sample validation and nearby-parameter stability especially important.
Defaults: Rising activity; fixed previous 7-day average; low momentum −10; high momentum +10.
46. Sports Misery Index
This experimental rule asks whether recent results for major New York sports teams coincide with stock-market behavior. Its fixed basket contains the Yankees, Mets, Knicks, Nets, Giants, Jets, Rangers, Islanders, and New York Liberty. For each market date, TickRun examines final games dated within the preceding seven calendar days. A win scores +1, a loss −1, and a tie zero.
Each participating team's scores are averaged first, and those team averages receive equal weight. TickRun then reverses the sign and multiplies by 100: +100 means every participating team lost, while −100 means every participating team won. Equal-team weighting prevents the much denser baseball schedule from automatically dominating football, basketball, or hockey. At least two teams and four unique games are required; otherwise the indicator is missing.
Team-misery mode buys at or above the high threshold and sells at or below the low threshold. Team-success mode reverses those actions. Values between the thresholds retain the existing position.
The checked-in source is ESPN's public team-schedule data. Only completed regular-season and postseason games are retained. Live, scheduled, postponed, cancelled, incomplete, and suspended records are excluded. ESPN does not expose a dependable historical result-publication timestamp, so a game can influence only market dates strictly later than its UTC game date. This conservative rule prevents same-date outcomes from entering the signal. The public JSON endpoint is not a formally supported developer API, so refresh validation must fail rather than silently accept a changed or incomplete schema.
There is no established causal mechanism connecting New York sports results to broad stock returns. League seasons overlap unevenly, the active team mixture changes through the year, and optimized thresholds can easily capture chance. Treat the strategy as experimental research requiring out-of-sample validation.
Defaults: Team misery; fixed previous 7 calendar days; low misery −25; high misery +25.
47. Lottery Jackpot Fever
This experimental sentiment rule combines the latest advertised annuity jackpots for Powerball and Mega Millions. It compares the combined value with the median combined jackpot over the preceding 365 calendar days and reports 100 × (current − median) / (current + median), bounded between −100 and +100. At least 180 prior daily values are required.
High-fever mode buys at or above the high threshold and sells at or below the low threshold. Low-fever mode reverses those actions. Values between thresholds retain the current position.
The checked-in records come from the Texas Lottery's dated historical winning-number pages. Although an estimated jackpot is normally advertised before its drawing, the archive does not preserve a dependable publication timestamp. TickRun therefore uses a draw's value only on dates strictly after that draw, a conservative one-day rule that prevents future jackpot records from leaking backward. Missing jackpot values are never invented.
Game rules, starting jackpots, inflation, and public interest change over time. There is no established causal mechanism connecting lottery jackpots to stock returns, so treat apparent performance as experimental evidence requiring out-of-sample testing.
Defaults: High fever; low fever −20; high fever +20; fixed 365-day median baseline.
48. Box Office Optimism
This experimental consumer-mood rule uses Box Office Mojo's domestic overall gross for each completed Friday–Thursday week. Each gross is compared with the median of seasonally comparable weeks within ten calendar days in the previous five years, using at least three earlier observations. The bounded score is 100 × (gross − baseline) / (gross + baseline).
Optimism mode buys at or above the high threshold and sells at or below the low threshold. Pessimism mode reverses those actions. The most recently available weekly score is carried forward until another score becomes available.
To avoid look-ahead, TickRun withholds every completed week for seven additional calendar days. All baseline observations come strictly from earlier years. This buffer is conservative, but the current historical website can still contain later corrections that were not known in real time; that revision risk is documented rather than hidden.
Ticket prices, release schedules, streaming competition, strikes, and the pandemic create structural changes that a seasonal median cannot remove. No established causal mechanism makes cinema grosses a stock-market predictor. Validate any result outside the optimized sample.
Defaults: Optimism; low optimism −15; high optimism +15; trailing five-year seasonal baseline.
49. Wikipedia Panic Index
This experimental attention indicator uses daily English Wikipedia readership for a fixed basket of four pages: Recession, Bankruptcy, Unemployment, and Stock market crash. For each page, TickRun compares the current user-view count with the median of the preceding 28 complete calendar days and calculates 100 × (current − median) / (current + median). The four bounded page scores are averaged equally so the most popular page cannot dominate.
Panic mode buys at or above the high threshold and sells at or below the low threshold. Calm mode reverses those actions. Values between the thresholds retain the current position. The calculation requires all four current observations and every observation in all four preceding 28-day windows; missing values produce no signal and are never filled.
The source is Wikimedia's open Analytics API for English Wikipedia, all device types, and the user agent category. This excludes traffic classified by Wikimedia as spiders or automated. The published counts are aggregate totals and do not identify individual readers.
A completed UTC day's count normally loads afterward, but the API does not preserve a historical publication timestamp for each record and exceptional processing delays can occur. TickRun therefore withholds every observation for three calendar days before use. This conservative buffer greatly reduces ordinary publication look-ahead, but current historical records may still contain later corrections. Page moves and redirects are another limitation because redirect traffic does not automatically transfer to the destination article.
Reading about economic stress can reflect news coverage, school assignments, curiosity, or genuine anxiety. There is no established causal mechanism making these page views a stock predictor. Treat results as an exploratory attention experiment and validate them outside the optimized period.
Defaults: Panic; fixed four-page basket; previous 28-day median; three-day availability delay; low panic −10; high panic +10.
50. Insider Confidence Index
This experimental filing-based indicator measures the balance between purchases and sales reported by corporate insiders. TickRun uses original Form 4 submissions and keeps only non-derivative transactions coded P (open-market or private purchase) or S (open-market or private sale). Grants, gifts, exercises, tax withholding, derivatives, amendments, and rows without positive reported shares and price are excluded rather than estimated.
For each calendar day, TickRun examines SEC filings from the preceding 28 complete calendar days. Purchase and sale dollars are summed separately for every issuer. Each issuer receives 100 × (purchases − sales) / (purchases + sales), and those bounded issuer scores are averaged equally. This prevents one exceptionally large transaction from dominating the raw balance. At least 20 issuers and 50 eligible transactions are required.
Because reported sales normally exceed purchases, TickRun then ranks the raw balance against its own preceding 365 calendar days. The historical percentile is mapped to −100 through +100, with no future observations in the baseline. Positive values mean confidence is unusually high relative to the recent year; negative values mean it is unusually low. At least 180 earlier daily balances are required.
Confidence mode buys at or above the high threshold and sells at or below the low threshold. Caution mode reverses those actions. Values between thresholds retain the current position.
The source is the U.S. Securities and Exchange Commission's quarterly Insider Transactions Data Sets, extracted from the structured portions of Forms 3, 4, and 5. Look-ahead protection uses the filing date—not the earlier transaction date. A filing dated D can first affect the indicator on calendar day D+1, and the changed market position first affects the following close-to-close return.
The SEC publishes the records as filed and warns that filer or extraction errors may exist. Present-day bulk files can include later corrections, and some purchases or sales may arise from prearranged plans or personal liquidity needs. Insider trading has a plausible information channel, but this particular aggregate has no established causal relationship with future broad-market returns. Treat it as experimental research and validate results outside the optimized sample.
Defaults: Confidence; fixed previous 28 calendar days; trailing 365-day percentile; low confidence −50; high confidence +50.
51. IPO Hype Index
This experimental capital-markets indicator measures unusual activity in public registration filings that are commonly associated with prospective offerings. TickRun uses original SEC forms S-1 and F-1 and excludes amendments. For each calendar day, it counts distinct issuer CIKs with an eligible filing in the preceding 28 complete calendar days; multiple filings from one issuer within that window count once.
The raw issuer count is ranked against only its preceding 365 calendar days, and the historical percentile is mapped to −100 through +100. Positive values mean registration activity is unusually high relative to the recent year, while negative values mean it is unusually low. At least 180 earlier daily counts are required.
High-hype mode buys at or above the high threshold and sells at or below the low threshold. Quiet-market mode reverses those actions. Values between thresholds retain the current position.
The source is the official SEC EDGAR quarterly master indexes. The SEC describes Form S-1 as a registration statement often filed in connection with an initial public offering; Form F-1 covers offerings by certain foreign private issuers. These forms are not an exact list of IPOs: they can cover other offerings, a filing may be withdrawn, and confidential draft submissions are absent until made public. TickRun therefore labels the count as an IPO-market activity proxy, not a completed-IPO series.
Look-ahead protection uses the public EDGAR filing date. A filing dated D can first affect the indicator on D+1, avoiding assumptions about its same-day acceptance time, and the changed position first affects the following close-to-close return. Present-day master indexes can include later EDGAR corrections.
Strong registration activity may reflect issuer confidence, investor demand, financing conditions, or temporary industry waves, but this aggregate has no established causal relationship with future broad-market returns. Treat it as experimental research and validate results outside the optimized sample.
Defaults: High hype; fixed previous 28 calendar days; trailing 365-day percentile; low hype −50; high hype +50.
52. Turtle Breakout (System 1)
This experimental trend-following strategy uses fixed System 1 signal rules. Buy when the daily close exceeds the highest high of the previous 20 trading sessions, unless the last completed hypothetical 20-day breakout was a winner. A close above the previous 55-session high provides a fallback entry regardless of that filter. Exit when the close falls below the previous 10-session low or reaches the fixed stop at entry close minus 2N.
The filter tracks hypothetical long and short breakouts independently of actual holdings, including skipped entries. A hypothetical trade ends at its opposite 10-session channel or a 2N loss. A profitable channel exit counts as a winner; a stop, loss or breakeven exit permits the next breakout. Its direction does not matter. With no earlier result, the first valid breakout is allowed. Only completed results are used.
N is the 20-session average true range: seed with the mean of the first 20 true ranges, then update as (19 × previous N + current true range) / 20. Freeze N at each entry for its stop. Channels exclude the current session, and signals need 20 complete prior bars; the fallback needs 55.
TickRun uses daily-close confirmation for entries, channel exits and stops. Signals first affect the following close-to-close return. This is a long-only adaptation: hypothetical shorts inform the filter but never create short holdings. The original Turtles used intraday breakout orders, volatility position sizing and pyramiding; those features are not simulated here. Stops do not guarantee an intraday fill at the stop price.
Defaults: Fixed 20-day entry; 10-day exit; 55-day fallback; N window 20; stop 2N; winner filter always enabled. No strategy parameters are optimized.
Read the dedicated Turtle System 1 history and backtesting guide for a worked example and research workflow.
Rule reference: Curtis Faith’s Original Turtle Trading Rules.
How to use the results responsibly
- Compare a strategy with Buy & Hold and inspect drawdown, not return alone.
- Include realistic transaction costs.
- Treat optimized parameters as hypotheses; optimization can overfit the same history it measures.
- Prefer rules that remain reasonable across different assets, periods, and nearby parameter values.
- Remember that corporate actions, liquidity, spreads, taxes, and execution constraints can change real outcomes.
Before optimizing a rule, read why the best backtest can be misleading. Then open the TickRun backtester, or consult the user manual.