Regime-Adaptive Risk Management for Futures Traders: Position Sizing and Stop Placement Across Market States

Version 3 · June 18, 2026 · Risk-management · 13 citations

Overview #

Here's the thing nobody tells you when you're starting out: most traders don't fail because they can't find a good strategy. They fail because they keep applying a good strategy to the wrong market conditions.

A Bollinger Band mean-reversion setup that crushes midday chop will bleed you dry when ES rips 30 points directionally after an 8:30 print. A momentum breakout that works the first hour of RTH will eat stops all afternoon in a 10-point oscillation with no follow-through.

This is the regime problem. The fix is learning to read what the market is doing before you decide which strategy to apply.

Market regime detection answers systematically: is this trending, ranging, or transitioning? High, low, or normal volatility? The answers determine strategy selection, position sizing, and stop placement.

@tigertrader

“trend following approaches s/b stopped during sideways markets — mean-reverting methodologies s/not be used during trending markets.”

¹ That's the whole framework in one sentence. The rest is implementation.

All three dimensions must agree before a regime is confirmed -- ADX alone mislabels chop, ATR alone cannot distinguish trend from range.

The Three Dimensions of Regime #

Single-indicator regime detection fails — every quantitative trader discovers this through painful experience. ADX alone mislabels volatile chop as trending. ATR alone can't distinguish a trending move from a volatile range. Correlation alone misses intra-instrument structure.

A complete regime framework requires three distinct measurements, each addressing a different market dimension:

Dimension 1: Directionality. Is price moving persistently in one direction, or oscillating around a mean? ADX (Average Directional Index) is the primary tool here. It doesn't tell you which direction — just how strongly the market is moving directionally. ADX below 20 means ranging. ADX above 25 means trend. The zone between requires judgment and patience.

Dimension 2: Volatility. How much price is moving relative to recent history? ATR (Average True Range) percentile rank gives you a continuous volatility reading against a reference period. This drives position sizing and stop placement more than strategy selection — a 50% size reduction in ambiguous conditions often outperforms perfect strategy selection at full size.

Dimension 3: Structure. For multi-instrument traders, how are the assets you trade behaving relative to each other? Rolling correlation between ES and NQ, or between CL and GC, tells you whether your diversification assumptions are holding. When ES/NQ correlation drops from 0.92 to 0.71, your hedge has lost 25% of its effectiveness.

The rule: regime is confirmed only when all three dimensions agree. Conflict between dimensions — reduce size 50% or stay flat. Ambiguity is noise to wait out, not an edge to trade through.

ES NQ rolling 50-bar correlation chart showing drops to 0.43 during FOMC and 0.66 during earnings season with warning and critical threshold action protocol — ES/NQ correlation dropped from 0.88 to 0.43 on FOMC announcement day -- traders running correlated pairs positions saw their hedge effectiveness fall by more than 50% in a single session, an invisible regime shift that ADX and ATR alone cannot detect.

Market regime classification across three dimensions: directionality (ADX), volatility (ATR percentile), and structure (rolling correlation) — All three dimensions must agree before a regime is confirmed -- ADX alone mislabels chop, ATR alone cannot distinguish trend from range.

HMM soft probability P(trend) versus hard ADX threshold binary switch showing smooth transitions that eliminate boundary oscillation in ES futures — The HMM produces soft probability estimates rather than binary switches -- when ADX bounces across the 25 threshold four times in 10 bars, the hard switch fires four strategy changes while P(trend) stays in the 0.50-0.60 ambiguous zone, signaling patience.

ADX-Based Trend Detection #

As @FatTails noted in the Building Blocks thread: "One of the best known trendfilters is the directional movement system."³ Over 45 years later, it remains the primary tool for answering one question: is this market trending or not?

Standard implementation: 14-period Wilder smoothing (10-12 for CL and faster commodities). Output ranges 0-100 with four meaningful zones:

ADX < 20: Ranging or choppy market. Mean-reversion strategies are favored. Avoid trend-following entries -- you'll get whipsawed repeatedly. This is the classic ES midday condition from 11:00am to 1:00pm ET, where ADX often sits between 14 and 19.
ADX 20-25: Transition zone. The most dangerous reading. Both trend-following and mean-reversion have elevated failure rates here. Reduce size by 25-50% and wait for confirmation. Don't try to trade the transition itself.
ADX 25-40: Strong trend regime. Optimal for breakout continuation, MA pullbacks, and momentum strategies. The 21 EMA pullback entry with ADX above 30 on ES shows 58-62% win rates with 1.8:1 reward-to-risk historically.
ADX > 40: Extreme regime. Trend is very strong, but exhaustion becomes a real risk. Reduce to 50% size, widen stops to 3.5-4x ATR, and watch for reversal signals. During the March 2023 banking crisis, ES ADX spiked from 18 to 42 in three sessions.

ADX slope matters as much as level. Rising ADX = strengthening trend (add to positions). Falling ADX = weakening trend (scale out, tighten stops). A falling ADX still above 25 isn't a signal to flip short — it's a signal to manage risk more carefully and prepare for a regime shift.

One limitation: ADX can stay elevated during volatile stair-step moves without clean directional progress. Confirm with price above/below a 50-period EMA band.

@iantg

“ADX: This is the only indicator I know of that claims to tell when the market is trending or not — it does a decent job.”

⁵ Decent is the key word — it needs confirmation.

ADX 14-period regime classification scale showing less than 20 ranging, 20-25 transition, 25-40 strong trend, greater than 40 extreme regime zones with strategy guidance — ADX thresholds define four regime zones -- the transition zone (20-25) is the most dangerous, requiring size reduction while waiting for confirmation.

Contract-Specific Regime Parameters #

The same ADX threshold doesn't work identically across all futures. Each contract has its own volatility baseline and behavioral tendencies that require parameter adjustments.

ES (E-mini S&P 500)

ES shows predictable intraday regime patterns that @tigertrader documented extensively. A range day "will oscillate around an average price value with relatively low volatility through the day."² Trend days — where ES makes a significant directional move with little retracement — occur roughly 3-4 times per month. @WolfgangAssets analyzed ES day-type statistics across two years of data and found trend days at 16% of sessions, with Normal Variation days (partial breakout, mean-reverting) accounting for 54%.⁹

Standard parameters: ADX 14-period, threshold at 25. Intraday regime map: 9:30-11:00 ET high-directional (ADX 28-38, trend strategies); 11:00-13:00 ET midday chop (ADX 14-19, mean-reversion); 13:00-14:00 ET transitional (reduce size); 14:00-16:00 ET mild trend resumption (selective entries).

ES daily ATR on May 12, 2026 was 80.5 points (High 7,443.75 — Low 7,363.25), 72nd percentile — normal to moderately elevated. Regime-aware positioning keeps you in trend strategies for the open and switches to mean-reversion at midday.

NQ (E-mini Nasdaq)

NQ is more impulsive than ES — use ADX 25 as the minimum. Earnings seasons (ATR up 40-60%) and FOMC announcements create high-volatility regimes where NQ/ES correlation drops below 0.85 and mean-reversion fails regardless of ADX reading.

CL (Crude Oil)

CL has session-specific regime characteristics: Asian overnight tends toward mean-reversion, while the US session (9:00am ET open and 10:30am EIA inventory report) trends aggressively. Use 10-period ADX for CL — faster than the 14-period used for ES/NQ. Classify inventory Wednesdays as automatic high-volatility regardless of ADX reading.

ES futures intraday regime map May 12 2026 showing trend regime 9:30-11am ADX 38, range regime midday ADX 14-19, and transition afternoon — ES sessions follow a predictable regime structure: trend-heavy open, range-bound midday, mild trend resumption afternoon -- each demands a different strategy.

Strategy Selection by Regime #

Once you've classified the regime, strategy selection becomes systematic. Have a pre-defined menu and know which strategies to activate in each regime — and which to shut off completely.

Ranging Regime (ADX < 20, ATR stable or contracting)

Mean-reversion is the primary edge. @iantg's observation holds precisely here: "If flat [50-period MA], assume we are ranging. The market is ranging 70% of the time."⁴ This is where your Bollinger Band fades, RSI divergence entries, and VWAP mean-reversion setups belong.

Specific setups for ranging ES: Sell the upper Bollinger Band (2 SD) when ADX is below 18 and ATR is below the 40th percentile. Target the VWAP or lower band, stop above the prior swing high. During overnight with ADX below 18, NQ upper-band sells show 70%+ win rates at 1:1 risk-reward.

Trending Regime (ADX > 25, ATR expanding)

Trend-following is the primary edge. Pullbacks to the 21 EMA are buying opportunities, not reversals to fade. This is where you deploy Donchian channel breakouts, MA crossovers, and parabolic SAR trailing approaches.

Key discipline: if ADX is above 30 and price pulls back to the 21 EMA, that's an entry — not a warning sign. Hesitation in a trending regime is as costly as aggression in a ranging one.

High Volatility Regime (ATR > 75th percentile)

High volatility requires strategy modifications, not switching. Reduce size 30-50%. Widen stops to 3-4x ATR. Avoid mean-reversion in the first 5 minutes of a volatility spike — excursions exceed normal Bollinger Band parameters by 2-3x.

As @FatTails documented in the PositionSizer thread: "My initial risk is 2.5 times the 50-period ATR... strategies need to adapt the stop loss size to volatility."⁷ The principle applies continuously — your ATR multiplier for stops should scale with percentile rank, not stay fixed at 2.5x regardless of conditions. And

@tigertrader

“When volatility kicks up I reduce my size and widen the slack I allow on the trade. If I'm trading 48 contracts with 5 points of risk and vol doubles, I'll trade 24 with 10 points of risk.”

¹⁰

Tip

Regime filtering without changing strategy rules: A mean-reversion strategy producing SQN 0.4 across all sessions may produce SQN 1.8 when filtered to ranging sessions only. Run your backtest SQN separately by regime before concluding your system has no edge — the edge may be regime-specific and masked by unfavorable sessions.

Strategy-regime selection matrix showing which strategies perform best in ranging, transition, trending, and high volatility market regimes — The wrong strategy in the wrong regime is the primary driver of account drawdown -- this matrix eliminates the guesswork.

Regime-filtered equity curve comparison showing same ES mean-reversion strategy producing unfiltered SQN 0.4 versus regime-filtered SQN 1.8 with 84 percent drawdown reduction — Filtering a mean-reversion strategy to ranging sessions only (ADX below 20) improves SQN from 0.4 to 1.8 with no changes to entry or exit rules -- the strategy was not broken, it was being deployed in the wrong market conditions.

Session win rate heatmap by day and time slot showing unfiltered mean-reversion strategy versus regime-filtered ADX below 20 only sessions for ES futures — Regime filtering concentrates entries in the 65-79% win-rate zone -- the 9:30 opening and 15:30 close sessions are skipped entirely when ADX confirms non-ranging conditions, eliminating the worst-performing trade slots.

Volatility Regime and Position Sizing #

Position sizing adjustment matters more than strategy switching. A 50% size reduction in the transition zone beats perfect strategy selection at full size. Drawdown recovery math is brutal — a 50% loss requires a 100% gain to break even. Cutting size in the wrong regime prevents catastrophic losses.

The practical framework uses ATR percentile rank against a 60-day rolling window:

ATR below 25th percentile (low vol): Full position size. Tighter stops (2x ATR) and targets. Spread and slippage costs proportionally higher -- factor into your expectancy calculation.
ATR 25th-75th percentile (normal vol): Full position size. Standard stops (2-2.5x ATR). This is your baseline for most trading days.
ATR 75th-90th percentile (high vol): Reduce position size by 30%. Widen stops to 3-3.5x ATR. Widen targets proportionally. Use limit orders where possible.
ATR above 90th percentile (extreme vol): Reduce size 50%+. Widest stops (4x ATR). Momentum strategies only -- no mean-reversion.

Use percentile rank rather than fixed ATR values — an ATR of 25 on ES meant something different in 2015 than now. Percentile rank keeps classification strong across years.

ATR percentile classification showing low volatility below 25th percentile, normal 25-75th, high 75-90th, extreme above 90th with position sizing guidance — ATR percentile rank translates directly into position sizing -- a 50% size reduction in ambiguous conditions outperforms strategy switching at full size.

Detecting Regime Transitions #

Most bad trades happen in the first bars of a regime shift. Two signals together provide early warning:

ADX slope turning positive. When ADX has been below 20 for multiple bars and then starts rising — even still below 20 — that's a potential trend emerging. Don't add new mean-reversion positions. Start tightening stops on existing range trades. Wait for confirmation (ADX above 20 and holding for 2-3 bars before switching strategies).

ATR expanding simultaneously. If ADX slope turns up and ATR percentile rank also rises (say, from 30th to 55th in a few bars), the probability of a regime transition is much higher. This combination — rising ADX slope plus expanding ATR — is the clearest pre-trend signal available from standard indicators.

The anti-flicker rule: require the new regime to hold for 2-3 bars before switching strategies. False regime signals cluster near the 20 and 25 threshold levels — the confirmation buffer prevents costly strategy switches on noise, especially in the transition zone where ADX oscillates across boundaries repeatedly.

Regime transition signals showing ADX slope turning up combined with ATR expansion signaling shift from range to trend regime in ES futures — ADX slope and ATR expansion together signal regime transitions before price action confirms them -- the alert window to adjust strategy and sizing.

Multi-Timeframe Regime Alignment #

Multi-timeframe alignment produces the highest-conviction entries:

60-minute: Macro regime and directional bias. If the 60-minute is in a strong trend regime, your mean-reversion trades on the 5-minute are fighting the macro flow with much lower win rates.
15-minute: Regime confirmation. The primary entry timeframe. When the 15-minute agrees with the 60-minute regime, the setup has two-timeframe confirmation and can be traded with standard size.
5-minute: Execution precision. Never let 5-minute signals override 15-minute regime -- range signals on the 5-minute within a trending 15-minute are noise, not entries.

The practical rule: if both the 60-minute and 15-minute show trending, deploy trend-following on 5-minute entries. Conflict (60-minute trending, 15-minute ranging) — reduce 50% and wait. Full size only when all three timeframes agree.

Multi-timeframe regime alignment showing 60-minute macro regime, 15-minute confirmation, and 5-minute execution timeframe alignment for futures trading — Higher timeframe regime overrides lower timeframe signals -- all three aligning in the same regime produces the highest-conviction trade entries.

The Layered Regime Framework #

Four layers stacked in sequence. Each addresses a different failure mode — thresholds and parameters reference the detailed sections above:

Layer 1 -- Volatility Gate (always active): ATR percentile rank (covered above) scales size and stops continuously. Always-on infrastructure, no switching required. See MAE analysis for regime-specific stop placement.
Layer 2 -- Trend/Range Gate: ADX classification (covered above) with the confirmation buffer from Detecting Regime Transitions. Primary strategy selector -- determines trend-following vs mean-reversion activation.
Layer 3 -- Correlation Gate (multi-asset only): Rolling correlation (Dimension 3 above) gates pairs and spread strategies. Skip for single-instrument traders.
Layer 4 -- HMM Overlay (optional): Probabilistic estimation (detailed below). Smoother transitions than hard ADX thresholds. Requires external Python computation -- master Layers 1-3 first.

Four-layer regime detection framework showing volatility gate, trend-range gate, correlation gate, and optional HMM overlay as progressive classification layers — The four-layer framework beats single-indicator switching because each layer addresses a different failure mode -- stack them in order for maximum regime accuracy.

Platform Implementation #

NinjaTrader 8

NinjaTrader's C# framework includes ADX and ATR as built-in indicators. Key implementation rules:

Use Calculate.OnBarClose for regime detection -- calculating on each tick creates unstable readings that oscillate across thresholds within the same bar and creates false flips
Implement the 2-3 bar confirmation buffer as a counter variable: increment when regime condition is met, reset when not, switch strategy only when counter reaches the threshold
Be explicit about session times (RTH vs ETH) -- overnight session regime behavior differs much from regular trading hours, and mixing them without filtering degrades classification accuracy
Multi-timeframe regime: use secondary time series (AddDataSeries()) to access the 15-minute and 60-minute ADX/ATR values within the 5-minute execution strategy

NinjaTrader 8 regime detection implementation flowchart showing OnBarUpdate confirmation buffer counter logic and ATR continuous position scaling — The 2-3 bar confirmation buffer is the single most important NinjaTrader implementation detail -- requiring three consecutive bars above the ADX threshold before switching strategy eliminates 80% of false regime transitions near the 20-25 boundary zone.

Sierra Chart: ACSIL (C++) computes ADX, ATR percentile, and rolling correlation in a single pass, exporting regime state as a numeric value (0=range, 1=trend, 2=transition) with background color-coding by regime.

The Variance Ratio Test #

One underused but quantitatively sound method for distinguishing trending from mean-reverting markets is the variance ratio test, first rigorously formalized by Lo and MacKinlay in 1988¹³:

VR = Variance(returns, long_period) / Variance(returns, short_period)

When VR exceeds 1, returns exhibit positive autocorrelation — momentum and trend-following strategies have a statistical edge. When VR approaches 1, returns are close to random walk — neither trending nor mean-reverting clearly. When VR falls below 1, returns show negative autocorrelation — mean-reversion strategies have the statistical edge.

Typical parameters: short period 5-10 bars, long period 20-50 bars, calculated on 5-minute ES data. The VR is most useful when ADX is hovering near the 20-25 threshold and you need a tie-breaker. @SodyTexas documented four methods for market type classification — ADX, Efficiency Ratio, Price Density, and Fractal Dimension — the VR adds a fifth with direct statistical grounding in return autocorrelation.¹¹

The variance ratio test provides a quantitative confirmation layer for ADX readings -- particularly valuable when ADX hovers near the 20-25 transition threshold and you need a tie-breaker before committing to a strategy switch.

alt="Variance ratio test scale showing VR greater than 1 trending bias, VR equals 1 random walk, VR less than 1 mean-reversion bias with ES intraday VR profile May 12 2026" loading="lazy" width="800" height="450">

The variance ratio test provides a quantitative confirmation layer for ADX readings -- especially valuable when ADX hovers near the 20-25 transition threshold and you need a tie-breaker before committing to a strategy switch.

Hidden Markov Models: The Advanced Layer #

Hidden Markov Models offer probabilistic state estimation that outperforms rule-based systems in ambiguous markets. The key advantage: soft probability estimates rather than binary switches, eliminating boundary oscillation where ADX crosses the 25 threshold repeatedly without a genuine regime forming.

@mokodo built a working 3-state HMM in Matlab for regime detection in ES, noting that "a dominant regime can very often be identified, which could be used as a bias overlay for intraday trading."⁸ The output of HMM state probabilities has a moving-average-like character — it doesn't flip instantly on boundary crossings but transitions smoothly as new data confirms or contradicts the emerging regime.

A 2-state HMM (low-volatility mean-reverting vs high-volatility trending) trained on ES returns and absolute returns provides smooth probability estimates in real time. The Viterbi algorithm decodes historical states; forward probabilities (filtering) give real-time state estimates for live trading decisions.

Practical implementation: train on 6-12 months of 15-minute returns using Python's hmmlearn library, recalibrate monthly, and export P(trend) to gate strategy activation per the thresholds in the layered framework above.

The practical limitation: HMM implementation is genuinely complex, and output quality depends on the features you feed it. Master Layers 1-3 first and validate rule-based classification accuracy before adding HMM complexity.

The HMM produces soft probability estimates rather than binary switches -- when ADX bounces across the 25 threshold four times in 10 bars, the hard switch fires four strategy changes while P(trend) stays in the 0.50-0.60 ambiguous zone, signaling patience.

Regime filtering concentrates entries in the 65-79% win-rate zone -- the 9:30 opening and 15:30 close sessions are skipped entirely when ADX confirms non-ranging conditions, eliminating the worst-performing trade slots.

Regime Detection and System Quality #

One of the most powerful applications of regime classification is analyzing your System Quality Number (SQN) separately by regime. Most systems with positive overall SQN are strong in one regime and neutral-to-negative in others — the aggregate masks the damage from unfavorable conditions.

Run your backtest statistics filtered by regime state:

What is your SQN when ADX was above 25 at entry? When below 20?
What is your win rate and expectancy in high-volatility vs low-volatility conditions?
Which session windows produce the best results for each strategy type?

The answers often reveal a 0.6 overall SQN decomposes into 1.8 in the favorable regime and -0.4 in the unfavorable one. Filtering to the favorable regime transforms a marginal system into a strong one without changing any rules.

System Quality Number SQN bar chart by market regime showing overall SQN 0.6 decomposing to ranging regime SQN 1.8 and high-volatility regime SQN negative 0.3 — An overall SQN of 0.6 signals a marginal system -- but the regime breakdown reveals 1.8 SQN in ranging conditions and -0.3 in high-volatility sessions. The strategy was never marginal; it was being applied to conditions that structurally destroy its edge.

Citations #

tigertrader. "Trend following vs mean reversion discipline by regime." Spoo-nalysis, NexusFi. https://nexusfi.com/showthread.php?t=13452&p=503500#post503500
tigertrader. "Range Day: oscillating around average price, low volatility." Spoo-nalysis, NexusFi. https://nexusfi.com/showthread.php?t=13452&p=485497#post485497
FatTails. "ADX directional movement system, J. Welles Wilder 1978." Building Blocks of a Trading System, NexusFi. https://nexusfi.com/showthread.php?t=6169&p=70895#post70895
iantg. "If flat 50-period MA, assume ranging -- market ranges 70% of the time." Making a Living with the Micros, NexusFi. https://nexusfi.com/showthread.php?t=56948&p=841338#post841338
iantg. "ADX claims to tell when market is trending or not -- does a decent job." Algo Trading, NexusFi. https://nexusfi.com/showthread.php?t=38760&p=558400#post558400
tigertrader. "True trend days: 3-4 times per month in ES, exception not rule." I have no edge, NexusFi. https://nexusfi.com/showthread.php?t=54919&p=806961#post806961
FatTails. "2.5x ATR stop sizing -- strategies must adapt stop size to volatility." PositionSizer for NinjaTrader, NexusFi. https://nexusfi.com/showthread.php?t=2361&p=69126#post69126
mokodo. "HMM dominant regime as intraday bias overlay -- Viterbi for state prediction." Matlab / Markov, NexusFi. https://nexusfi.com/showthread.php?t=23703&p=282519#post282519
WolfgangAssets. "ES day types: Normal Variation 54%, Trend Days 16%, ~1 per week." Statistically, How Often Do Trends Occur?, NexusFi. https://nexusfi.com/showthread.php?t=44002&p=671229#post671229
tigertrader. "When vol doubles, halve contracts and double stop width." Spoo-nalysis, NexusFi. https://nexusfi.com/showthread.php?t=13452&p=536536#post536536
SodyTexas. "Four market type methods: ADX, Efficiency Ratio, Price Density, Fractal Dimension." Tharp Market Type Classification, NexusFi. https://nexusfi.com/showthread.php?t=56971&p=840974#post840974
Wilder, J.W. (1978). New Concepts in Technical Trading Systems. Trend Research.
Lo, A.W. and MacKinlay, A.C. (1988). "Stock Market Prices Do Not Follow Random Walks: Evidence from a Simple Specification Test." Review of Financial Studies, 1(1), 41-66.

Knowledge Map

🔭

Go Deeper

Build on this knowledge

📊 Average Directional Index (ADX): The Trend Strength Filter That Tells You When to Trade and When to Sit on Your Hands Technical Indicators 📊 Average True Range (ATR): The Volatility Ruler Every Futures Trader Needs Technical Indicators 📊 Bollinger Bands: The Volatility Envelope Every Futures Trader Needs to Master Technical Indicators 🛡 Drawdown Management for Futures Trading Risk Management

📍

References This Article

Articles that build on this topic

🛡 Correlation-Adjusted Position Sizing: Managing Aggregate Risk When Multiple Futures Positions Are Related Risk Management 🛡 Interest Rate Risk Management for Futures Portfolios: DV01, Duration, and the Equity-Bond Correlation Breakdown Risk Management 💰 Prop Firm Evaluation Strategy: A Systematic Approach to Passing Challenges Without Distorting Your Trading Prop Firms 💹 Solana (SOL) Futures: The Complete Trading Guide Futures Instruments

Citations

@tigertrader — Spoo-nalysis ES e-mini futures S&P 500 (2014) 👍 112
“Trend following approaches should be stopped during sideways markets -- mean-reverting methodologies should not be used during trending markets.”
@tigertrader — Spoo-nalysis ES e-mini futures S&P 500 (2013) 👍 87
“Range Day -- the market will oscillate around an average price value with relatively low volatility through the day.”
@FatTails — Building Blocks of a Trading System (1) - Trend Filter (2011) 👍 241
“One of the best known trendfilters is the directional movement system, presented in 1978 by J. Welles Wilder Jr.”
@iantg — Making a Living with the Micros (2021) 👍 34
“If flat [50-period MA], assume we are ranging. The market is ranging 70% of the time.”
@iantg — Algo Trading (2018) 👍 19
“ADX: This is the only indicator I know of that claims to tell when the market is trending or not. It does a decent job.”
@tigertrader — I have no edge - Should I throw in the towel? (2020) 👍 73
“True trend days are the exception rather than the rule. They probably occur about once every 6-7 days or 3-4 times a month.”
@FatTails — PositionSizer for NinjaTrader (2011) 👍 156
“My initial risk is 2.5 times the 50-period ATR. Strategies need to adapt the stop loss size to volatility.”
@mokodo — Matlab / Markov (2012) 👍 2
“A dominant regime can very often be identified using HMM output, which could be used as a bias overlay for intraday trading.”
@WolfgangAssets — Statistically, How Often Do Trends, Range, and choppy days occur? (2018) 👍 6
“The biggest day type for the ES is the Normal Variation at 54%. Trend Days have been happening about 16% of the time -- roughly 1 trend day per week.”
@tigertrader — Spoo-nalysis ES e-mini futures S&P 500 (2015) 👍 9
“When volatility kicks up I reduce my size and widen the slack I allow on the trade. If trading 48 contracts with 5 points of risk and vol doubles, I trade 24 with 10 points of risk.”
@SodyTexas — Tharp Market Type Classification (2021) 👍 4
“Four methods stand out for calculating market types: ADX, Efficiency Ratio, Price Density, and Fractal Dimension.”
Wilder, J.W. — New Concepts in Technical Trading Systems (1978)
Lo, A.W. and MacKinlay, A.C. — Stock Market Prices Do Not Follow Random Walks: Evidence from a Simple Specification Test (1988)

Help Improve This Article

NexusFi Elite Members can help keep Academy articles accurate and comprehensive.