Systematic vs Discretionary Trading: Which One Should You Build?

Two broad approaches to trading: systematic (rules-based, mechanical, every trade follows a deterministic process) and discretionary (judgment-based, contextual, the trader uses expertise to decide each trade). Both can produce profitable edge. Both have specific failure modes. Most retail confuses the two, ending up with the worst of both worlds. Knowing which one you're actually building, and what validation it needs, clarifies a lot.

What systematic trading is

A systematic strategy is a set of rules that produces a trade decision from inputs without requiring judgment. Given the same market conditions, the system always produces the same trade.

Examples of systematic rules:

• "Buy when 4-hour MACD crosses up AND daily 50 SMA is rising"
• "Short when RSI(14) > 75 AND price hits prior swing high"
• "DCA $100 into BTC every Tuesday at 12:00 UTC"

Systematic strategies have several properties:

• Backtestable mechanically
• Repeatable across operators (anyone running the rules gets the same trades)
• Auditable post-hoc (was this trade per the rules or not?)
• Often automatable (a script can execute the rules)
• Explicit about exactly when they apply and don't apply

The trader's role in a systematic strategy is operational: following the rules, monitoring for execution issues, periodically reviewing whether the rules still produce edge.

What discretionary trading is

A discretionary strategy uses the trader's judgment to make trade decisions. Given the same market conditions, two different discretionary traders might take different trades, both potentially profitable, based on their specific reads.

Examples of discretionary trade rationale:

• "I see structural weakness here, with a confluence of factors I've learned to recognize over years"
• "This setup is technically valid but the news context worries me, I'm passing"
• "The chart looks like the 2022 bottom did, and I think the same setup is in play"

Discretionary strategies have:

• Reliance on developed pattern recognition
• Adaptability to unusual circumstances
• Difficulty of articulation (the trader often can't fully explain why this setup looks better than that one)
• Hard to backtest mechanically
• Harder to audit (was this a "valid" discretionary call?)
• Almost impossible to automate

The trader's role in discretionary trading is cognitive: applying expertise developed over years of pattern recognition.

Where each one wins

Systematic wins when:

• The strategy is genuinely mechanical (clean rules with edge)
• The trader's psychology is unreliable (can't follow judgment without bias)
• Trade frequency is too high for human attention
• The trader wants to reduce decision fatigue
• Validation via backtest is critical
• Multiple traders need to execute the same strategy
• The strategy is part of a larger automated framework

Discretionary wins when:

• The strategy depends on context that's hard to encode
• The trader has developed real pattern-recognition expertise (years of deliberate practice)
• Edge cases and unusual conditions matter
• The trade frequency is low enough for thoughtful per-trade analysis
• The "rules" can't be specified precisely enough to systematize without losing the edge

Both can work. Both have produced extremely successful traders.

The most common failure: pretending one is the other

A common pattern in retail:

The "discretionary" trader who's actually undisciplined. They claim to use "judgment" but really they're trading on intuition that hasn't been validated. They have no checklist, no consistent setup definitions, no post-trade review. Each trade is rationalized post-hoc with whatever reasoning sounds good. This isn't discretionary trading, it's improvisation with a sophisticated label.

The "systematic" trader who keeps overriding the system. They have a strategy with rules. Most of the time they follow the rules. But the trades that "don't feel right" get skipped, and the setups that aren't strictly valid but "look really good" get taken. The rules end up being suggestions; performance ends up being mediocre because they're neither cleanly systematic nor genuinely discretionary.

The fix: pick one, build it deliberately, and execute it consistently. The hybrid where you have rules but ignore them is worse than either pure approach.

How to validate systematic strategies

Systematic strategies validate via:

• Honest backtest (per the backtesting chapter)
• Walk-forward validation (per that chapter)
• Paper trading to verify executability
• Live trading at small size to measure friction
• Ongoing monitoring of expectancy

The key advantage: you can answer "does this work?" quantitatively because the strategy is testable.

If a systematic strategy isn't producing edge in walk-forward, you have a clear answer: it doesn't work, move on. The data is unambiguous.

How to validate discretionary strategies

Discretionary strategies are harder to validate because you can't backtest a judgment call. But you can:

1. Track expectancy across many trades. Same R-multiple framework. Across 100+ live trades, your discretionary calls either produce positive expectancy or they don't.

2. Tag setup categories. "Pullback to support in trending market," "Range break," "Reversal at major level," etc. Track expectancy by category. Some categories will be your real edge; others will be where you bleed.

3. Compare your calls to a baseline. Define a simple systematic version of what you do (the "naive" version of the same setup). Compare your discretionary results to the systematic baseline. Your judgment should be adding value (better expectancy) relative to the simple rules.

4. Review trades against post-hoc data. Did the trades you passed on (because of judgment) tend to fail? Did the trades you took because of context that the simple rules missed tend to work? If yes, your discretion is adding value. If not, your "discretion" is removing value relative to a simple rule.

The validation is harder but possible. Without it, "discretionary" is indistinguishable from "winging it."

A common mistake: building systematic without enough data

A trader specifies a system, backtests on 100 historical trades, sees positive expectancy, deploys live at full size. The live performance is much worse than the backtest. They're confused.

100 trades is a small sample. The backtest could easily have been favorable noise. Walk-forward validation across multiple regimes would have surfaced this; the trader skipped it.

The fix: systematic strategies need substantial validation before deployment. Don't shortcut walk-forward.

A common mistake: claiming discretion without expertise

A trader has been live for 8 months. They claim to be "discretionary" because they "just have a feel for the market." Their P&L is negative. They keep tweaking without any specific framework.

Discretionary trading depends on developed expertise. Without years of deliberate practice and reflection, the "feel" is mostly cognitive bias. Calling improvisation "discretion" is a way to avoid building the discipline that systematic trading would require.

The fix: be honest about expertise. If you don't have 5+ years of deliberate practice in a specific approach, default to systematic, the rules will protect you from your own limitations. Earn discretion through validated expertise.

A common mistake: blending without discipline

A trader runs a systematic strategy but "uses discretion" to skip some trades. They claim this hybrid is the best of both worlds. But there's no consistent framework for when to override; the overrides correlate with recent performance and emotional state more than with any objective improvement criterion.

The fix: if you want a hybrid, formalize the override conditions. "I skip systematic trades when X is true." The override condition should itself be specific and auditable. Without that, the "hybrid" is just systematic with arbitrary skipping, which usually performs worse than the pure system.

Mental model, systematic as the recipe, discretionary as the chef

A recipe produces consistent results regardless of who follows it. The cake comes out the same way every time. This is systematic trading, the rules produce the output.

A chef can improvise based on the ingredients, the context, what the diners want today. The chef's expertise adapts to circumstances the recipe couldn't anticipate. This is discretionary trading, judgment adapts.

But the chef earned that ability through years of cooking recipes. A novice who claims to be a chef is just guessing. And the chef who has earned the skill but ignores it (skipping basic preparation, not tasting, improvising without reason) produces worse food than either the recipe-follower or the genuine improviser.

In trading: most retail should follow recipes. The chef move is earned, not declared.

Why this matters for trading

Most retail traders are implicitly discretionary without realizing it, they've never built a system, so every trade is a judgment call by default. This usually produces poor results because the judgment hasn't been trained. Hex37 supports both approaches: bracket orders and consistent risk sizing for systematic execution; the journal and breakdown tools for discretionary performance review. Pick which you're building. Validate it appropriately. Don't pretend to be the other.

Takeaway

Systematic strategies execute mechanically; discretionary strategies use judgment. Both can produce edge, with different validation requirements. Most retail "discretionary" traders are actually undisciplined; most retail "systematic" traders override the system. Pick one deliberately. Validate systematic via backtest, walk- forward, and live-small. Validate discretionary via journaled expectancy across categories, comparison to naive baselines, and review against systematic counterfactuals. The hybrid approach can work but only with formalized override conditions, otherwise it's worse than either pure approach.