Skip to content

๐Ÿงช Correlation Tab

The Correlation tab of the Assets page asks the same questions of a selection of assets you put together โ€” assets you own, assets you only watch, or a broker's holdings. Open it from Assets in the sidebar, then the Correlation tab in the toolbar. A selection has no weights, so the tab shows percentages and ratios, never money: for your own money, see the Dashboard's Risk tab.

Each block below follows one pattern: what it answers, a screenshot, its tools and how to read them.

๐Ÿงบ Building the Selection

The card at the top answers which assets are compared? Each chip is one selected asset, and the tab remembers your selection in this browser.

The + panel open over the selection card: search, Type and Currency filters, two assets ticked with Add 2, and, read-only, the assets that cannot be analysed in the period, each with its reason

Tools shown

  • Eligibility check โ€” whether an asset has enough prices in the period to be analysed โ†’ Data Quality

How to read it

  • Add assets with the +, remove one with its ร—, or change many at once with Select all, Deselect all, Invert and My assets โ€” up to 100 assets.
  • The counter โ€” 3 in the analysis, of 42 that can be analysed โ€” counts the selected assets being analysed, out of all those that could be over the period set in the toolbar.
  • A dashed chip cannot be analysed in this period: it stays selected, but out of the results. An amber chip is analysed, with a warning. Hover or tap a chip to read why.
  • If the period leaves some assets without prices, an amber strip offers Use the period in which all have prices, which moves the page's dates to fit them all.

๐Ÿ•ธ๏ธ Correlation

This section answers which of these move together? โ€” to spot the assets that are really the same bet, and any that cushion the others.

The Correlation section with five assets: the triangular matrix with its legend, order buttons and sector and region badges; The most alike pairs RE Loan Roma with RE Loan Milano at 0.94, near-identical, and The ones that offset is empty

Tools shown

  • Correlation coefficient (ฯ) โ€” how closely two assets moved together, from โˆ’1 (always opposite) to +1 (always together) โ†’ Correlation
  • Order by similarity โ€” the default order: closely linked assets sit side by side, so redundant groups show up as blocks โ†’ Correlation

How to read it

  • One square per pair: blue when the two assets move together, red when they move in opposite directions, pale when they move independently. Hover or tap it for the value in words.
  • Start from the two lists: The most alike names the pairs that are almost the same exposure, The ones that offset the pairs that cushion each other. An empty list is a finding. Click a pair to find it in the matrix.
  • A dash is not a zero: the pair could not be measured โ€” too little history, or a price that never moved.
  • The order buttons rearrange the matrix; they never change a value.

๐Ÿ“‰ How Much Did Each of These Hurt?

This section puts every asset on the same scale of harm: its worst losses in the period, and how far it still stands below its peak.

The How much did each of these hurt? table: one row per asset with Bad day, Bad month, Worst fall and how long it lasted, Below peak now and Rise to peak

Tools shown

  • Bad day โ€” the average loss on the worst 5% of days (CVaR at 95%) โ†’ Conditional Value at Risk
  • Bad month โ€” the same over real one-month stretches โ†’ Conditional Value at Risk
  • Worst fall โ€” the deepest drop from a peak; lasted N d counts the days until the asset was back there, or until the end of the period โ†’ Max Drawdown
  • Below peak now โ€” how far it stands below its highest level in the period โ†’ Current Drawdown
  • Rise to peak โ€” the gain needed to get back there: 20% below takes +25% โ†’ What It Takes to Get Back

How to read it

  • Each column is its own horizon, from a day to the whole period: compare the assets down a column, never add the columns up.
  • Not a ranking: the table opens in your selection's order. Click a column title to sort it, or hover it for what it measures.
  • On a short period, Bad day and Bad month can stay empty while the drawdown columns are filled: they need more history.

โš–๏ธ What Did Each of These Pay for Its Risk?

This section sets risk against reward โ€” how much each asset swung, and what it returned on average per year โ€” in a table and a chart. With a benchmark chosen under Compared with, it also shows how each asset moved with it.

What did each of these pay for its risk? with S&P 500 as the benchmark: its tinted row opening the table, the period line, and the chart with the asset circles, the benchmark's diamond and the dashed line

Tools shown

  • Volatility โ€” how much the returns swung, on a yearly basis โ†’ Volatility
  • Ann. return โ€” the average annual return: the period's average return, scaled to a year โ†’ Observed Annualization
  • Sortino โ€” the return earned per unit of downward swing โ†’ Sortino Ratio
  • Sharpe โ€” the return earned per unit of volatility โ†’ Sharpe Ratio
  • Beta โ€” with a benchmark: how much the asset moved when the benchmark moved โ†’ Beta & Active Return
  • Correlation โ€” with a benchmark: how closely the two moved together โ†’ Benchmark Selection
  • Risk/return line โ€” with a benchmark: the dashed line drawn through it on the chart โ†’ The Risk/Return Line

How to read it

  • Not a ranking: the table opens with the benchmark's row, then your selection in its own order. Click a column title to sort it.
  • The line under the table gives the exact period behind the figures, and says when it is shorter than yours โ€” usually because of an asset with a shorter history.
  • On the chart, further right means more swing and higher means more average return; the benchmark is the diamond. A circle above the dashed line was better paid for its risk than the benchmark. Click a row or a circle to highlight that asset in both.

The average annual return is not the return you lived through

On a very volatile asset, the return actually lived through is lower: do not read the height of a dot as what the asset earned. It also comes from prices alone โ€” coupons and dividends are not included yet.

๐ŸŽฏ Choosing the Benchmark

The Compared with picker open: search, Type and Currency filters, the usable assets, then, read-only, those not usable over this period, each with its reason
  • One benchmark for every risk page: your choice under Compared with also applies to the Dashboard, broker and asset risk pages, so that they stay comparable.
  • Its row opens the table, with a dash under Beta and Correlation: the benchmark is not compared with itself. It may also be one of your selected assets.
  • Assets that cannot be measured over the period are listed apart in the picker, read-only. A benchmark among them stays chosen but is not used until you pick a period that suits it.

โฎ๏ธ What Ifโ€ฆ?

This section answers how did each of these come through a real past episode? โ€” a built-in crisis or a period you choose, replayed with real returns. Click its title to open it. Only the historical replay is offered: a hypothetical shock or a simulation would need weights.

What ifโ€ฆ? after Run replay: the box of assets left out, grouped by reason, with the button that replays the shorter period, above the table with Return and Effect

Tools shown

  • Historical replay โ€” each asset's real return over a past period โ†’ Historical Replay

How to read it

  • Choose a Preset, such as the Global Financial Crisis, or set the Period, then press Run replay. Changing the selection or the page's dates clears the result.
  • The table shows each asset's Return, worst first, with a bar: red for a loss, green for a gain. There is no total: a selection has no weights to add up.
  • Assets without prices at the ends of the period are left out and listed above the table, with the reason. When possible, a button replays the shorter period in which they all have prices.

๐Ÿšฉ Notices and Missing Data

When something is wrong with the data, the tab says so once, right below the selection card, and names the results it affects.

Top of the Correlation tab: the folded data-quality banner, the notice Some results are partial with its Measurements affected badges, and the Correlation section's amber banner Unavailable for the selected data

Tools shown

  • Data quality โ€” missing or outdated prices and exchange rates, and the results they make partial โ†’ Data Quality

How to read it

  • The data-quality banner lists what to fix, each problem with its action, such as Sync prices or Sync rates. Click it to open the list.
  • The notice below it names the partial results and their cause. A partial result still shows its figures: read them with that cause in mind.
  • A section's own amber banner means a result did not come back at all, often because the period is too short: choose a longer one, or sync the prices.
  • What ifโ€ฆ? reports its own problems inside its section.

๐Ÿ”Ž Good to Know

  • One period for all the assets. Every figure covers the same days for every selected asset, so the rows can be compared. An asset with a shorter history shortens that period for all of them, and their figures change: this is expected โ†’ Data Quality
  • One currency. Returns are measured in your Default Currency, set in Preferences (the instance's default if you never chose one), exchange-rate moves included.
  • A dash is not a zero. The figure could not be measured for that asset over this period; hover or tap it to read why.
  • Another page, another figure. Another page or another selection measures over different days: compare figures only when they cover the same period.
  • ๐Ÿ“‹ Asset List โ€” the Assets tab, its toolbar and its date range
  • ๐Ÿ›ก๏ธ Dashboard Risk Tab โ€” the same questions, for your own portfolio
  • ๐Ÿ“Š Risk Metrics โ€” the theory behind every section of this tab
  • ๐Ÿ› ๏ธ Technical details โ€” for developers: how this tab works inside