Results & Analytics

TL;DR. The Analytics view turns a finished backtest into something you can act on. The headline metric is rarely the most important number on the screen — the rolling diagnostics, the worst-10 trade list, and the period heatmap are where you decide whether the result is a real edge or a curve-fit.

What you'll be able to do

• Read the equity curve and drawdown curve together to spot the patterns that don't show up in the summary table.
• Use the MAE/MFE distribution to refine stop and target placement instead of guessing.
• Recognize edge decay in a rolling window before it shows up as a live drawdown.

What it lets you do

A backtest result is a multi-megabyte data packet: every trade, every signal evaluation, every per-bar indicator value, plus a few dozen summary statistics. The Analytics view organizes that into four lenses you'll use in roughly this order every time:

1. Summary metrics — does this look promising enough to spend more time on?
2. Equity and drawdown — how did it make money? Steady grind or one or two big wins?
3. Rolling diagnostics — does the edge persist across time, or did it concentrate in one regime?
4. Trade list and distributions — what do the worst trades look like, and would I have actually held them?

Skip any of these and you're auditing a strategy with half the lights off.

Where to find it

Every completed backtest opens directly into the Analytics view. From the dashboard, Strategies → [strategy name] → Latest Run also takes you straight there. Multiple runs of the same strategy are listed side-by-side so you can pick one or compare two.

Lens 1 — the summary metrics

The top of the page is a grid of ~30 numbers organized into four blocks. The full definitions live on the trading metrics page — what to look at first:

• CAGR and max drawdown together. CAGR alone is meaningless without the drawdown that produced it. A 20% CAGR with a 15% max DD is a different strategy from a 20% CAGR with a 60% max DD, even though the headline number is the same.
• Sharpe and Sortino. Sharpe penalizes all volatility; Sortino only penalizes downside. If the two diverge sharply, the strategy has fat upside tails — usually good.
• Profit factor (gross wins ÷ gross losses) and expectancy (average $/trade). A profit factor under 1.3 with realistic costs is fragile; over 2.0 is suspicious and warrants a curve-fit check.
• Win rate is the metric beginners overweight. A 35% win-rate strategy with a 3R reward-to-risk profile beats a 70% win-rate strategy with 0.5R reward-to-risk in the long run — and survives drawdowns better.

Lens 2 — the equity and drawdown curves

The equity curve answers a question the summary table can't: how did the strategy make its money?

Three patterns to recognize:

• The clean stair-stepper. Monthly equity additions, drawdowns that recover in weeks. This is what a robust strategy looks like. Most look-too-good-to-be-true strategies don't look like this.
• The big-trade carry. Equity is flat for months, then jumps on a single trade. Check the trade list — if one or two trades account for >40% of total profit, you're looking at a lottery ticket, not a strategy.
• The smile. Strong early period, flat middle, strong late period. Often a sign that the strategy works in two specific regimes and is dead in everything else. Cross-reference the market regimes page.

The drawdown curve sits below the equity curve on the same time axis. Pay attention to:

• Drawdown duration, not just depth. A 12% drawdown that recovers in 2 weeks is a different psychological experience from a 12% drawdown that takes 11 months.
• Number of new equity highs. A strategy that puts in 30 new highs over 5 years is fundamentally healthier than one that puts in 4.

Lens 3 — the rolling diagnostics

This is where edge decay shows up before it shows up in your live account.

The rolling window panel lets you choose a metric (Sharpe, win rate, profit factor, expectancy) and a window length (30, 60, 90, 180 trades), then plots that metric across time. A strategy that performed well in 2018–2020 but whose rolling Sharpe has been below 0.5 for the last 18 months may have already stopped working — even if the lifetime summary still looks great.

What to look for:

• Drift. A slowly declining rolling Sharpe across years is the classic signature of overfitting being slowly exposed by changing market structure.
• Regime sensitivity. A bimodal distribution (long stretches of high Sharpe, long stretches of negative) means the strategy works in one regime and bleeds in the other. Pair it with a regime detection filter or accept that it's a sometimes-on strategy.
• Recovery. After a drawdown, does the rolling Sharpe come back to its prior level, or does it settle lower? Settling-lower is a yellow flag.

Lens 4 — the trade list and MAE/MFE

Open the trade list and sort by P&L ascending. Look at the worst 10 trades.

For each, ask:

• Would I actually have held this in real life? If the answer is "no, I would have closed it on Day 3," the backtest is overstating its real performance.
• Is there a structural commonality? (All on Mondays. All during FOMC week. All on the day after a 2σ move.) If yes, you've found a regime filter waiting to be added.

The MAE (Maximum Adverse Excursion) and MFE (Maximum Favorable Excursion) distributions answer two surgical questions:

• Are my stops too tight? If most winning trades had to first dip to -0.8R before going to +2R, a -0.5R stop would have killed half your winners.
• Am I exiting winners too early? If most winning trades had a peak MFE of +3R but you took +1R, there's profit on the table.

Period heatmap

A grid of returns by month-of-year × year. The single fastest way to see whether a strategy has a structural weakness in specific calendar windows (the August doldrums, the November-December tape, FOMC weeks).

A healthy strategy is mostly green with red months sprinkled in randomly. An unhealthy one has a recognizable pattern — and that pattern is usually fixable with a time filter using the Current Time node and a Comparison against the month.

Comparison mode

Select two backtests (typically two parameter variants of the same strategy) and the equity curves overlay on a single chart. Use this for:

• Sensitivity check. If the only difference between two runs is RSI period 12 vs 14, the equity curves should look very similar. If they diverge wildly, the strategy is over-tuned to that one parameter. The avoiding curve-fit page covers what to do about it.
• Before-and-after. Added a regime filter? Toggle the comparison on to see exactly which trades it removed and what that did to the equity curve.

Prop-firm overlay

If your account is configured as a prop-firm account (prop firm traders use case), the Analytics view adds:

• Daily loss limit tracking. A horizontal threshold line on the daily P&L panel, with violations highlighted in red.
• Trailing max drawdown. The "relative drawdown" rule most prop firms use, modeled exactly.
• Phase progress. How close the strategy is to the profit target while staying in compliance.

These aren't simulations of the prop firm's rules — they're the actual rules of FTMO, TopStep, Apex, etc., wired into the metrics so a backtest that would have failed the evaluation is flagged before you take it live.

Configuration reference

Tips and pitfalls

• Don't anchor on the headline metric. Whatever the platform shows biggest at the top, treat it as one of 30 numbers, not as the number.
• Always open the worst-10 trade list. Five minutes here catches more flaws than any other audit.
• Export the trade CSV. Even with everything the in-app analytics gives you, there's no substitute for slicing the data yourself once you have a hypothesis. The export button sits in the top-right of the trade list.
• Compare two runs only when one variable changes. Otherwise you're measuring noise.

How this fits the workflow

You design in the Visual Builder, validate in the Backtest Engine, and read the results here. When a strategy passes the audit, it's a candidate for the Portfolio Builder and eventually Forward Testing. When it fails, the diagnostics here usually tell you which page to read next: overfitting, risk management, or regime detection.