A backtest equity curve looks simple at first: a line that rises and falls over time, often compared against a benchmark. Knowing how to read a backtest equity curve properly — not just its endpoint but its shape, drawdowns, and comparison lines — separates a useful research tool from a source of bad expectations.
What a backtest equity curve actually shows
The Y-axis represents cumulative return indexed to 1.0 at the simulation start. A value of 2.0 means the simulated portfolio doubled from its starting level. The X-axis is time, typically plotted by month, since monthly rebalancing is the standard cadence for AI-curated equity strategies.
Each point on the line answers: "If you had held this strategy's composition at the start of each month, what would your notional be worth relative to where you began?" The curve rises when a month's return is positive and falls when negative, compounding every gain and loss against the prior balance.
That compounding is not cosmetic. A 35% drawdown in year two can still leave the five-year curve positive if the recovery is strong — and conversely, a curve that ends above its starting level can still have inflicted serious pain in the middle years.
What the benchmark line is measuring
Most backtest charts overlay a benchmark line for context. For US equity strategies, the benchmark is a broad market index. It answers: "What would passive index exposure have returned over the same period?"
When the strategy line sits above the benchmark, the simulation shows the strategy outperformed the index. When it dips below, the reverse. The gap between the two lines is the simulated excess return — and it varies considerably by time window. A strategy that ran well from 2013 through 2020 may have narrowed that gap sharply in the rate-shock year of 2022, and the chart at any given date reflects the entire path.
The benchmark line is a reference, not a target. Strategies with concentrated mandates — blue-chip only, mid-cap, or sector-focused — won't track a broad index closely by design. Comparing DOW10 against a cap-weighted index measures how ten Dow components fared against 500 names; the shape of that comparison tells you more about mandate differences than about quality.
How blended returns work across multiple strategies
VelaDeck's backtest simulator shows a weighted blend when you allocate across more than one strategy. If you assign $6,000 to US20 and $4,000 to SP20, the backtest weights US20's monthly return series at 60% and SP20's at 40%, then compounds the weighted sum each month.
That blended line is not an average of the two endpoints — it's a single compounded path that reflects your specific allocation split. Change the weight and the curve changes. US20 — Top Value Stocks — has accumulated approximately +145% over a five-year simulation window; SP20 — Beat the S&P 500 — is calibrated toward index outperformance with 20 monthly-rebalanced holdings. A blend of the two will sit somewhere between their individual paths, weighted by your notional. See the [strategies page](/strategies) for current composition counts and rebalance cadence.
The practical use of this blended view is scenario framing: how would your specific split have behaved in 2020 versus 2022? The simulation can answer that — with the caveats below.
Four things the curve cannot tell you
- **Fill timing.** The simulation assumes each month's composition is held from the start of the rebalance period. Your actual orders fill at the market price when VelaDeck submits them, which is after the composition publishes and after the ingest job runs. On volatile days, that spread matters.
2. **Cash drag.** Between a SELL order settling and a BUY filling, the notional sits as cash. The simulation treats the portfolio as continuously invested. In practice there's always some idle balance, especially around partial rebalances.
3. **Entry timing.** The simulation starts from a fixed historical date. You start from today. A strategy that returned +145% over a five-year window from its simulation start may produce a different result from your funding date, depending on valuations, macro conditions, and which monthly compositions fire in your first year.
4. **Drawdown correlation.** The blended backtest weights return series independently. In a broad market selloff, most equity strategies fall together. The diversification benefit of holding US20 and IT15 simultaneously looks larger on the backtest than it will feel when both strategies decline in the same month.
For more on what the simulation models and what it excludes, see the [backtest simulation explained post](/blog/backtest-simulation-explained).
How to use the curve without over-reading it
Read the drawdown depth, not just the endpoint. A curve that ends at +150% but passed through a -40% trough tests whether you'd have held through that trough in practice. If that drawdown would have made you pause the automation, the endpoint return is not achievable for you regardless of simulation history.
Compare curve shapes, not just final values. Two strategies can share similar five-year endpoints while one was smooth and the other volatile. The shape reveals the volatility profile you're opting into. [How monthly rebalancing works](/blog/how-monthly-rebalancing-works) affects both the smoothness of the return path and the turnover embedded in each cycle.
Use the benchmark gap as a diagnostic, not a scorecard. If a strategy consistently ran below the benchmark for multiple years, that's worth understanding — not necessarily a reason to avoid it, but a reason to ask why. A value-oriented strategy underperforming a growth-heavy index during a growth-led bull run is expected behavior. A strategy that underperforms across all market environments is a different situation.
VelaDeck does not provide investment advice. The backtest curve is a research tool for understanding historical behavior, not a prediction of future results. Live trading is off by default and requires explicit opt-in — start in paper mode to see how signals and fills behave before any real capital moves. [Create a free account to explore the backtest for your allocation.](/signup)
Frequently asked questions
What does the Y-axis of a backtest equity curve represent?
Cumulative return indexed to 1.0 at the simulation start. A value of 1.5 means the simulated portfolio sits 50% above its starting level. Each month's return compounds against the prior balance, so the Y-axis reflects the full path of compounded performance, not simply the sum of individual monthly returns.
Is the benchmark line in VelaDeck's backtest the S&P 500?
It's a broad US equity index used as a reference context line. It shows what passive index exposure would have produced over the same period. Strategies with concentrated mandates — Dow-only, sector-focused, or mid-cap — diverge from that benchmark by design. That divergence tells you about mandate differences more than it tells you about quality.
If the strategy line is above the benchmark, does that mean it will outperform going forward?
No. Historical simulation shows what the strategy's composition would have returned under past conditions. Market conditions change. A strategy that outperformed a broad index for a decade can underperform for several subsequent years without the underlying signal being broken. Past simulation results are not a forecast of future returns.
Does VelaDeck hold my money while running the backtest?
No. VelaDeck doesn't custody funds at any point. The backtest is a local calculation against your allocation inputs — no capital moves. Your money stays in your Alpaca brokerage account at all times. VelaDeck submits orders only after you've explicitly opted into live trading in Settings, and paper mode is always the default.
How far back does VelaDeck's backtest data go?
Most AI-curated equity strategy series in VelaDeck start around 2013, giving roughly twelve years of simulation history. Earlier data isn't available for all strategies. The simulation window for a blended allocation is bounded by the shortest-running series in your allocation mix.