Parabolic Stop And Reverse

The Trailing Stop Problem

Every trader faces a dilemma: how do you protect profits without exiting too early?

Set your stop too tight, and you get shaken out by normal volatility. Set it too loose, and you give back too much profit when the trend finally ends. J. Welles Wilder (creator of RSI, ATR, and ADX) solved this with an elegant insight in 1978:

What if the stop automatically tightened as the trend matured?

Early in a trend, you need room to breathe — the move hasn't proven itself yet. But as the trend extends and you accumulate profits, your tolerance for giving them back should decrease. The stop should accelerate toward price, eventually catching up and signaling an exit.

This is the "parabolic" in Parabolic SAR — the stop follows a parabolic curve that gradually steepens, chasing price with increasing urgency.

---

How It Works

flowchart LR
    subgraph Uptrend["Uptrend Phase"]
        direction TB
        U1["Dots BELOW price"]
        U2["SAR rises toward price"]
    end
    
    subgraph Flip["Trend Reversal"]
        F["Price crosses SAR"]
    end
    
    subgraph Downtrend["Downtrend Phase"]
        direction TB
        D1["Dots ABOVE price"]
        D2["SAR falls toward price"]
    end
    
    Uptrend --> Flip --> Downtrend

The SAR value represents your trailing stop:

• Dots below price → You're in an uptrend, SAR is your support/stop-loss
• Dots above price → You're in a downtrend, SAR is your resistance/stop-loss
• Price crosses SAR → Trend has reversed, flip your position

---

The Mathematics

Uptrend SAR Calculation

$$\text{SAR}_{t+1} = \text{SAR}_t + AF \times (EP - \text{SAR}_t)$$

Downtrend SAR Calculation

$$\text{SAR}_{t+1} = \text{SAR}_t - AF \times (\text{SAR}_t - EP)$$

Where:

• SAR — Current Stop and Reverse point
• AF — Acceleration Factor (starts at step, increases by step each new extreme, caps at max_step)
• EP — Extreme Point (highest high in uptrend, lowest low in downtrend)

---

The Acceleration Factor: Why "Parabolic"?

The AF is the secret sauce. Each time price makes a new extreme (new high in uptrend, new low in downtrend), AF increases by the step value:

Event	AF	SAR Behavior
Trend starts	0.02	SAR moves slowly, giving room
New high/low	0.04	SAR accelerates
Another new high/low	0.06	Faster still
...	...	...
Maximum reached	0.20	Maximum acceleration

This creates a parabolic curve — starting flat, then steepening dramatically. The longer the trend persists, the more aggressive the stop becomes. Eventually, SAR catches up to price and triggers a reversal signal.

The intuition: Profitable trends deserve tight stops. Young trends deserve loose stops. The AF encodes this wisdom mathematically.

---

Trading Signals

flowchart TD
    Price["Current Price"] --> Compare{"Price vs SAR"}
    
    Compare -->|"Price > SAR"| Uptrend["UPTREND<br/>SAR = Support"]
    Compare -->|"Price < SAR"| Downtrend["DOWNTREND<br/>SAR = Resistance"]
    
    Uptrend --> Cross1{"Price crosses<br/>below SAR?"}
    Cross1 -->|Yes| Sell["SELL Signal<br/>Dots flip above"]
    
    Downtrend --> Cross2{"Price crosses<br/>above SAR?"}
    Cross2 -->|Yes| Buy["BUY Signal<br/>Dots flip below"]

Signal	Condition	Action
Buy	Price crosses above SAR (dots flip below)	Enter long / Exit short
Sell	Price crosses below SAR (dots flip above)	Enter short / Exit long

---

The Outputs

This module produces four outputs:

Output	Description
PSAR	The SAR value (trailing stop level)
Up	SAR value when in uptrend (0 otherwise)
Down	SAR value when in downtrend (0 otherwise)
Down Indicator	Binary: 1 if in downtrend, 0 if in uptrend

---

Missing Data

A closed-market or missing-data bar is skipped when advancing the SAR state machine — the state (trend, Extreme Point, Acceleration Factor) carries across the gap and resumes on the next observed bar, never resetting or fabricating a bar in between.

• Gaps are bridged, not broken: a Friday-to-Monday transition is evaluated as adjacent bars; the closed weekend itself reports null on all four outputs, never a repeated or interpolated value.
• A reversal that occurs during a gap surfaces one bar late: if the trend would have flipped over the weekend, the flip is only detected once trading resumes (the next observed bar) — a known, accepted lag, not a bug.
• Warm-up (first 2 observed bars) is null on all four outputs, not an echoed close or a fabricated trend flag — there is no real SAR value until the state machine has enough history to establish a trend.
• 0 means genuine opposite-trend, never missing: Up/Down read exactly 0 while the trend is on the other side — null is the only "no value" signal.
• Live execution returns exactly one datapoint per output — the newest bar's value, or null when that bar itself is missing or still in warm-up.

---

Parameter Tuning

Parameter	Default	Purpose
Step	0.02	AF starting value and increment
Max Step	0.20	Maximum AF cap

Effect of Parameters

Setting	Effect
Higher Step	Faster acceleration, tighter stops, more signals
Lower Step	Slower acceleration, looser stops, fewer signals
Higher Max Step	Eventually very tight stops
Lower Max Step	Stops never get too aggressive

Recommended Configurations

Market Type	Step	Max Step
Strong trends	0.025	0.25
Normal markets	0.02	0.20
Choppy/Volatile	0.01	0.10
Crypto	0.01	0.10–0.15

---

When It Works, When It Struggles

Excels In:

• Strong, directional trends — The classic use case
• Trailing stop management — Automatic, systematic exit
• Trend identification — Instant visual of current trend
• Momentum continuation — Riding winners

Struggles In:

• Ranging/sideways markets — Whipsaws repeatedly as price oscillates
• Gap environments — Gaps can trigger false reversals
• Choppy trends — Frequent small reversals generate excessive signals

The ADX Filter

Wilder himself recommended combining PSAR with ADX (which he also invented):

ADX	PSAR Signal	Recommendation
> 25	Any	Take the signal — strong trend
< 25	Any	Ignore — market is ranging

---

Creative Applications

Trend-Phase Position Sizing

Use the AF as a proxy for trend maturity:

AF Range	Interpretation	Position
0.02–0.06	Young trend	Full position
0.08–0.12	Mature trend	Reduce position
0.14–0.20	Extended trend	Minimum position

As AF approaches the max, the trend is "old" and more likely to reverse.

Multi-Timeframe Trend Alignment

Check PSAR on higher timeframes to establish bias:

Weekly PSAR	Daily PSAR	Action
Uptrend	Buy signal	Strong long entry
Uptrend	Sell signal	Ignore or reduce
Downtrend	Sell signal	Strong short entry

Volatility-Adjusted Parameters

Adjust Step based on ATR:

• High ATR → Lower Step (wider stops for volatile markets)
• Low ATR → Higher Step (tighter stops for calm markets)

---

Summary

Aspect	Key Point
Core insight	Stops should tighten as trends mature
What it measures	Dynamic trailing stop with parabolic acceleration
SAR below price	Uptrend — SAR is support
SAR above price	Downtrend — SAR is resistance
Best for	Trailing stops, trend-following, exit management
Worst for	Ranging markets (use ADX filter)

---

Further Reading

• Parabolic SAR - Investopedia
• Wilder, J.W. (1978). New Concepts in Technical Trading Systems
• Combining PSAR with ADX - Investopedia

---

Indicator Assessment

%%{init: {"quadrantChart": {"titleFontSize": 14}, "themeVariables": {"quadrant1Fill": "#e8f5e9", "quadrant2Fill": "#fff3e0", "quadrant3Fill": "#ffebee", "quadrant4Fill": "#e3f2fd"}}}%%
quadrantChart
    title Parabolic SAR Assessment
    x-axis Low Popularity --> High Popularity
    y-axis Low Reliability --> High Reliability
    quadrant-1 High potential  
    quadrant-2 Established
    quadrant-3 Niche
    quadrant-4 Widely used
    Stocks: [0.70, 0.65]
    Crypto: [0.55, 0.45]

Dimension	Stocks	Crypto	Notes
Popularity	●●●●○	●●●○○	Classic Wilder indicator, widely available
Reliability	●●●◐○	●●◐○○	Excellent in trends, poor in ranges
Complexity	●●●○○	●●●○○	AF acceleration logic requires understanding
Versatility	●●●○○	●●○○○	Primarily trailing stop; struggles in choppy crypto markets
Recent Interest	●●●○○	●●◐○○	Stable classic; crypto interest lower due to volatility

Why lower crypto reliability: PSAR generates many whipsaws in crypto's frequent consolidation phases. The ADX filter is especially important when using PSAR for digital assets.