About Risk-Adjusted Returns

The Risk-Adjusted Returns & Probability Dashboard is an institutional-grade suite that evaluates whether an asset's price appreciation adequately compensates investors for its inherent volatility. Moving beyond raw performance metrics, this dashboard deploys traditional quantitative models including the **Sharpe Ratio**, **Sortino Ratio**, and **Calmar Ratio**, alongside a proprietary **Risk-Adjusted Composite** that fuses all three into a single consensus gauge. By mapping these metrics over dynamic rolling windows - and ranking each against its own complete history through the **Percentile** view - analysts can objectively identify periods where the market is generating unsustainable excess returns or offering deep-value risk discounts. Each ratio serves a specific analytical purpose. The Sharpe model evaluates general volatility, whilst the Sortino model isolates downside deviation to avoid penalising explosive upside moves. The Calmar model specifically measures returns relative to maximum drawdown, making it highly effective for cyclical assets. The **Risk-Adjusted Composite** blends all three on an equal-weighted basis and fires a high-conviction reading only when the Sharpe, Sortino and Calmar signals agree. When these ratios trigger their upper thresholds, it mathematically signals that the asset is severely overbought relative to its risk profile. Every ratio can be viewed in two modes via the **Display** control. **Raw** plots the ratio in its native units against a zero baseline. **Percentile** - the default lens - re-expresses the same ratio as its rank within the asset's entire history on a fixed 0–100% scale, so a reading is always anchored to where it sits across every prior cycle rather than merely the visible window. Distinct **Oversold** (0–15%) and **Overbought** (85–100%) bands, a dashed median at the 50th percentile, and each metric's own colour gradient make cross-cycle positioning instantly legible. Optional overbought/oversold markers on the price line remain available in both modes, and every metric carries a bespoke **Insights** panel and **AI Summary** that surface the statistics specific to that ratio rather than a single repeated set. To complement the oscillator models, the dashboard features a robust empirical probability engine. This module scans the entire price history to calculate the exact historical likelihood of positive or negative returns across specific holding horizons. Paired with the Positive **HODL** Days metric, investors can optimise their time-in-market expectations and size positions based on absolute statistical precedent rather than emotion. The **Risk-of-Ruin** module completes the suite by answering the question the ratio metrics deliberately abstract away: *what does the pain actually feel like along the way?* Three sub-views deliver this in three distinct ways, all built from lived historical data with zero simulation. The **Underwater Curve** plots the asset's percentage from its running peak over time, with the deepest historical drawdowns annotated and a 'you are here' marker showing where current pain sits in the historical distribution. The **Recovery Atlas** scatters every historical drawdown episode by depth and days-to-recover, with the current ongoing drawdown highlighted as a diamond so users can see exactly where today's pain ranks against history. The **Recovery Paths** view traces every historical drawdown from its peak forward as a 'spaghetti chart' of actual trajectories, overlays a median recovery shape computed from all paths, and highlights the current drawdown in bright green so users can see exactly where their position sits on a path the asset has traced before.

Signal Zones

Trading Signals by Regime

How It Is Calculated

Frequently asked questions

What is the Percentile view, and why is it the default?

Every ratio has a different natural range - a Sharpe of 2 and a Calmar of 10 can both be 'extreme', which makes raw values hard to compare across metrics, assets and timeframes. The **Percentile** view solves this by ranking the current reading against the metric's entire history on a universal 0–100% scale. Because the rank is computed over all history and only then sliced to your chosen window, the scale is absolute: a 92% reading is genuinely in the top 8% of everything the asset has ever recorded, not merely the top of what is currently on screen. It is the default lens because it turns four differently-scaled ratios into one consistent, instantly-comparable cycle gauge. Switch back to **Raw** at any time from the Display control to see the ratio in its native units.

What is the Risk-Adjusted Composite?

It is a single consensus gauge that fuses the **Sharpe**, **Sortino** and **Calmar** ratios. Each is divided by its own 'strong' reference level so they carry equal weight despite their different scales, then the three are averaged. The Composite reads near zero in neutral regimes, pushes toward 1 when all three ratios are simultaneously strong, and turns negative when risk-adjusted performance is poor across the board. Its power is consensus: it only reaches an extreme when Sharpe, Sortino and Calmar agree, filtering out the single-metric head-fakes that any one ratio can throw on its own.

How are the Overbought and Oversold bands defined in the Percentile view?

They are fixed percentile zones, not raw thresholds. The **Oversold** band covers the bottom 15% of the metric's historical distribution and the **Overbought** band the top 15%, with a dashed median line at the 50th percentile. A ratio sitting in the Overbought band is historically stretched - risk-adjusted returns this high have rarely persisted - while the Oversold band marks the depressed, deep-value end of the range. Because the bands are anchored to the full history rather than the visible window, they retain their meaning across every cycle.

Why does each ratio show different statistics in its Insights panel and AI Summary?

Each ratio answers a different question, so the Insights panel and AI Summary surface statistics tailored to that question rather than one repeated set. **Sharpe** reports its Quality Days (time spent above 1, the institutional-quality line) and time-positive share. **Sortino** reports its Upside Premium - how far it sits above the same-window Sharpe, which reveals whether recent volatility has been upside or downside. **Calmar** anchors on the maximum drawdown driving its denominator. The **Composite** breaks out the individual percentiles of its three components plus an Agreement readout showing whether they are aligned or in conflict. The AI Summary then writes a regime read built from those same metric-specific figures, so the narrative is genuinely about the ratio you are looking at.

What is the difference between the Sharpe and Sortino ratios?

The **Sharpe Ratio** penalises all volatility equally, meaning explosive upside moves actually reduce the score. The **Sortino Ratio** resolves this by only factoring in downside deviation, providing a more accurate measure of true risk for highly convex assets. The gap between them - surfaced as the Sortino Upside Premium - is itself a signal: a wide positive premium means recent volatility has been mostly upside, the hallmark of a clean trend.

Why use the Calmar ratio for crypto assets?

Crypto is uniquely prone to severe, prolonged drawdowns. The **Calmar Ratio** divides annualised returns by the maximum drawdown, offering a highly precise metric for evaluating whether the pain of holding through a bear market was adequately compensated.

How should I use the rolling time windows?

Shorter windows (such as 90-day) are highly responsive and useful for timing swing trades or local mean reversion. Longer windows (like 365-day or 2-year) smooth out intraday noise, making them ideal for identifying macro cycle tops and generational bottoms.

What does the Positive HODL Days metric show?

It calculates the exact percentage of days in the asset's history that would currently be in profit if a position was initiated on that day and held to the present. A low percentage indicates deep structural capitulation.

What does the Risk-of-Ruin tab actually tell me?

It answers the question every other metric on this dashboard deliberately abstracts away: *what does the pain feel like along the way?* The ratio metrics tell you whether returns compensated for risk; the Ruin tab tells you exactly what risk you would have endured. Three sub-views cover this differently - the **Underwater Curve** shows the time series of percent-from-peak, the **Recovery Atlas** plots every historical drawdown as a dot by depth and recovery time, and **Recovery Paths** traces every drawdown's actual trajectory from peak forward. All three use lived historical data with no simulation.

Why is the current drawdown highlighted in the Atlas and Paths views?

Most drawdown analysis treats the present as separate from history. The Ruin tab deliberately collapses that distinction: in the **Recovery Atlas** the ongoing drawdown is rendered as a green diamond with a glow ring so you can see exactly where today's pain sits in the cloud of historical episodes - depth and elapsed time since peak. In **Recovery Paths** the ongoing drawdown is drawn in bright OCM green on top of all the historical paths, so you can see if your current trajectory is tracking better or worse than the asset's typical recovery shape. This is the single most useful piece of context the tab provides: *given how this asset has drawn down before, where am I on a familiar shape?*

What does the Median Path on the Recovery Paths chart represent?

For each day-since-peak that is present in at least 3 historical episodes, we take the median underwater value across those episodes. The resulting dashed white line is the asset's 'typical' recovery shape - derived purely from lived historical data. It is not a forecast and contains no statistical or distributional assumption. Its value is comparative: when the current drawdown is tracking *above* the median, the recovery is unfolding better than typical; when it is tracking *below*, the current episode is worse than usual and probably has more pain to come.

Why does the Ruin tab use ≥10% depth for Recovery Paths but ≥5% for the Recovery Atlas?

Different visualisations need different episode thresholds to read well. The Atlas uses ≥5% so the dot cloud is well-populated - you want to see the mid-sized drawdowns as well as the catastrophic ones to understand the full distribution of historical pain. Recovery Paths uses ≥10% because every sub-10% episode would clutter the spaghetti chart with daily wiggles that have no analytical value - what you want there is the universe of *meaningful* drawdowns the asset has traced out. The Underwater Curve uses a separate threshold (≥10%) for selecting which episodes get hover annotations, since space for annotations is limited.

Why is the underwater curve always ≤ 0%?

By construction. The underwater series at any day t is defined as (price_t − peak_t) / peak_t, where peak_t is the running all-time high up to and including day t. Whenever a new all-time high is set, peak_t = price_t and the underwater value hits exactly 0%. Between new highs, peak_t stays fixed at the previous high, so price_t < peak_t and the underwater value is negative. The series can never go positive because the moment price exceeds the prior peak, the peak is updated. This is the standard definition used in institutional risk analytics, and it is why the chart's filled red area always lives below the zero line.