The Engineer's Guide to Systematic Investing

When people learn that investing is one of my interests, the most common reaction is disappointment that I do not have a hot stock pick for them. The expectation is that someone “into” investing should be ready with three tickers and a story about why each one is going up next quarter. That expectation comes from treating investing as entertainment. The thing I actually do is closer to engineering: define an objective, write down rules, control for known error sources, and stop touching the system between scheduled reviews.

That approach has a name. Systematic investing replaces ad hoc forecasts with explicit, repeatable rules. It can be as simple as contributing to a globally diversified index portfolio every payday and rebalancing once a year, or as complex as a multi-factor long/short strategy with daily risk targeting. What makes it systematic is the discipline, not the complexity.

Process beats prediction

The case for treating investing as a process problem rather than a prediction problem rests on three findings the academic record has replicated over decades.

Most active management trails the market after costs. William Sharpe’s 1991 essay The Arithmetic of Active Management made the structural point: before costs, the average actively managed dollar must equal the market average; after costs, the average actively managed dollar must lag the passive average. Sharpe (1991). The empirical record matches the math. S&P’s SPIVA U.S. Year-End 2025 scorecard found that 79% of active large-cap U.S. equity funds underperformed the S&P 500 in 2025, the fourth-worst result for active managers in the 25-year history of the report. Long-horizon underperformance is worse: across 15-year SPIVA windows, the share of active funds that beat their benchmark is consistently in the single digits to low teens. S&P SPIVA U.S. Scorecard.

Retail trading does worse than that. Brad Barber and Terrance Odean’s Trading Is Hazardous to Your Wealth looked at 66,465 households at a large discount broker from 1991 to 1996. The most-active households earned an annualized 11.4% while the value-weighted market earned 17.9%. Average annual portfolio turnover for the most-active group exceeded 250%. Barber & Odean (2000). The paper’s last sentence remains the cleanest summary in the literature: “Our central message is that trading is hazardous to your wealth.”

Even disciplined investors leak return through behavior. Morningstar’s Mind the Gap 2025 measured the difference between fund total returns and the dollar-weighted return actually earned by investors in those funds. Over the decade ending December 2024, the average dollar earned 7.0% per year while the funds themselves returned 8.2%, a 1.2 percentage-point shortfall that adds up to about 15% of fund gains over the period. The gap was smallest for set-and-forget allocation funds (97% capture) and largest for sector funds and high-turnover categories. Morningstar (2025).

Why this resonates with engineers

Systematic investing maps cleanly onto how engineers and other quantitative professionals already think. The investor specifies an objective, declares constraints, identifies error sources, and then designs a process robust to the largest of them. That mental model replaces the prediction game with a familiar exercise in expected value, sensitivity, and reliability.

• Falsifiable rules. A systematic plan can be written down and audited. “Buy when it feels right” cannot.
• Decoupling skill from luck. A multi-decade process produces enough trades to test whether the rules survive regime changes. A handful of stock picks does not.
• Explicit error budgets. Fees, taxes, and behavior gaps each have a measurable annual rate. Treating them as line items, the way an engineer treats power budgets or latency budgets, surfaces where the real money is being lost.
• Robustness over optimization. An over-fit backtest is just curve-fitting. Systematic investors who borrow robustness checks from engineering (out-of-sample tests, stress scenarios, sensitivity analysis) tend to ship plans that survive contact with reality.

The systematic-investing ladder

“Systematic” spans a wide range of sophistication. Placing yourself on the ladder helps decide where the next dollar of effort goes.

Level 1 captures most of the return available to most households. Level 2 captures most of the rest. Levels 3 onward are real but increasingly small marginal gains relative to the engineering cost. Skipping levels in favor of Level 6 strategies before Level 1 is operationally locked in is the most common DIY mistake.

The evidence on systematic factors

Decades of academic research have identified characteristics that appear to earn higher long-run average returns than broad market exposure: size, value, profitability, momentum, low volatility, carry, and trend-following. Eugene Fama and Kenneth French formalized the size and value factors in 1992 and added profitability and investment patterns in their 5-factor model. Fama & French (1992). Mark Carhart added momentum to the mutual-fund-performance literature and showed that common factors explained much of apparent persistence. Carhart (1997). Robert Novy-Marx showed that gross profitability had explanatory power comparable to book-to-market in predicting average returns. Novy-Marx (2013).

These results are real but should be held loosely. Campbell Harvey, Yan Liu, and Heqing Zhu cataloged hundreds of published return factors and showed that the multiple-testing problem alone makes many of them statistically suspect: with enough data mining, plenty of “significant” factors appear by chance. Harvey, Liu, & Zhu (2014). A defensible factor needs an economic rationale, long sample evidence, out-of-sample support, cross-market robustness, post-cost implementability, and a behavioral story for why an investor can survive the inevitable droughts. The Summitward guides on Fama-French factors and small-cap value work through which factors clear that bar and which do not.

Why broad diversification is the natural Level 1

Hendrik Bessembinder’s Do Stocks Outperform Treasury Bills? showed that long-run market returns are produced by an astonishingly small minority of stocks. From 1926 through 2016, the top-performing 4% of CRSP-listed U.S. companies accounted for the entire net wealth creation above one-month Treasury bills; the median listed stock failed to beat T-bills over its lifetime. Bessembinder (2018). The Summitward guide on most stocks lose to T-bills works through the math.

That skewness is the strongest mechanical argument for owning the market broadly. The investor who tries to pick the next handful of winners is fighting a distribution where missing the right few-percent matters more than avoiding the bad many-percent. Broad diversification guarantees you own the winners before anyone knows who they are.

Friction is the real enemy

Once a portfolio is broadly diversified and the contributions are on autopilot, the next dollar of improvement comes from cutting recurring frictions. Each one looks small per year and large over decades.

• Excess fees. The gap between a 5 bps total-market ETF and a 50-100 bps active fund compounds to six-figure differences over a working career.
• Behavior gap. The 1.2 pp/year Morningstar shortfall is the cost of buying high and selling low in aggregate. It is largest for sector funds and high-turnover categories.
• Tax drag. Realized short-term gains from high-turnover trading are taxed as ordinary income; even long-term gains compound less efficiently than unrealized ones.
• Tracking error from factor tilts you cannot stick with. A small-cap-value tilt that you abandon after seven years of underperformance is worse than the plain market-cap index you would have held instead.

Try it: Friction Cost Calculator

Set your starting balance, contributions, horizon, and expected return. Then move the friction sliders. The chart shows what each friction costs you in lifetime dollars when applied alone, and the verdict cards show the disciplined plan vs. the all-frictions plan side by side. The Morningstar 1.2 pp/year behavior gap is the default for that slider.

What a Level 2 plan looks like in practice

For most DIY investors, the boring version of systematic investing is also the most reliable.

1. Write an investment policy statement. One page. Goal, target allocation, rebalancing rule, contribution rule, tax-account rule, fun-money cap.
2. Build a globally diversified core. A U.S. + international split, often implemented with VTI plus VXUS in taxable for foreign-tax-credit and tax-loss-harvesting flexibility, or VT in tax-advantaged accounts. The case is in Case for Global Diversification.
3. Automate contributions. Paycheck deferrals, auto-invest on dividend payment dates, RSU sales on a written schedule.
4. Rebalance with rules, not feel. Annual or semi-annual checks at 5 percentage-point bands. The rebalancing guide covers the evidence.
5. Manage taxes intentionally. Asset location, tax-loss harvesting under wash-sale rules, RSU sell-at-vest policy.
6. Cap any speculation. If you keep a fun-money bucket for individual stocks or thematic bets, write down the dollar cap before you start, and never refill it from the core plan.

Who systematic investing fits, and who it does not

Frequently Asked Questions

Is systematic investing the same as passive investing?

Passive investing (index funds, market-cap weights) is the most common implementation of Level 1 and Level 2 systematic investing, but the categories overlap rather than match. Factor-tilted systematic strategies are still systematic without being passive. The Summitward guide on the passive-investing label works through where the discretion actually lives in “passive” funds.

Does automating my contributions count?

Yes, that is Level 1. Automating contributions to a globally diversified portfolio is already systematic investing, even without a written policy or rebalancing rule. Adding the policy and the rules moves you to Level 2 and captures most of the remaining benefit.

Should I add a factor tilt?

Only if you understand the tracking-error commitment. Small-cap value, in particular, has trailed the broad market by a wide margin for stretches of seven years or more in the historical record. A tilt you abandon in year six is worse than the plain-market portfolio you would have held instead. The Summitward guides on AVUV / AVDV / AVGV and the profitability factor cover the practical implementation.

What about momentum, trend-following, or managed futures?

These are real systematic strategies with academic support, but they live at Level 4-5 of the ladder. Implementation matters even more than for value or profitability tilts. Most DIY investors capture the bulk of available compensated risk before they need to consider any of them.

How small should my fun-money bucket be?

Small enough that a total loss would not change your retirement timeline. A common boundary is 5% or less of liquid investable assets, refilled only from earnings (never from the core plan). The point is to scratch the stock-picking itch without putting the plan at risk.

Related Guides

• How Financial Sales Pitches Hide the Real Cost of Investing is the hub on the marketed products a systematic process helps you ignore, with a hurdle calculator.
• Most Stocks Lose to T-Bills covers the Bessembinder skewness result that anchors the case for broad diversification.
• Fama-French Factor Analysis works through the academic factors and how to read your portfolio’s exposures.
• Passive Investing Is a Label covers where discretionary choices hide inside ostensibly passive index funds.
• How Often to Rebalance is the practical execution rule for Level 2 systematic plans.
• One Household Portfolio, Not One Per Account frames asset location and rebalancing at the household level rather than per-account.
• When to Sell a Winning Stock works through the cognitive-load case for systematic investing from the other direction: the sell decisions indexing offloads, and what to do with the old individual stock positions you already hold.
• Robinhood’s Agentic Trading applies the process-not-prediction lens to AI agents: why automating a sound policy helps and automating discretionary trades does not.

Key Takeaways

• Systematic investing is process discipline, not complexity. Level 1 (automatic contributions to a diversified portfolio) is already systematic investing.
• Three independent results favor process over prediction. SPIVA shows 79% of active large-caps trailed the S&P 500 in 2025; Barber-Odean showed retail traders earned 11.4% vs the market’s 17.9%; Morningstar measured a 1.2 pp/year behavior gap over the decade ending 2024.
• Friction is the lever most DIY investors leave uncovered. Excess fees, behavior gap, and tax drag each compound to six-figure differences over a working career. The calculator above quantifies your specific exposure.
• Factor tilts are optional, not magic. Hold loosely, demand economic rationale, and only commit to a tilt you can survive seven years of underperformance in.
• Cap entertainment investing. A small stock-picking bucket is fine if it is capped, refilled only from earnings, and walled off from the plan that has to work.