What the Acubic Liquidity Filter Does

Key takeaways

• Every Acubic ranking method applies a liquidity floor: an asset is eligible only if its median daily dollar volume over the last 60 trading days is at least $1 million as of the ranking date.
• Dollar volume is calculated as closing price times shares traded in US dollars. An asset without enough recent trading history to compute the 60-day median is also excluded.
• The floor runs before any other filter, including your sector, region, and asset class preferences. An asset that fails it cannot appear in a proposal regardless of how it would rank otherwise.
• Currency pairs and any ticker you have locked into your portfolio are exempt from the floor.
• The floor applies to strategies created or regenerated since October 5, 2026. A strategy saved before that date uses its original rules in live re-rankings until you regenerate it; a new backtest of that strategy applies the floor, so the two can select different assets.

If you have ever noticed a ticker missing from an Acubic proposal that you expected to find, one of the most common reasons is the liquidity floor. Acubic applies a minimum trading activity threshold to every asset before any ranking or weighting begins, and assets that fall below it are removed entirely. This article explains what the floor is, how it is calculated, why it exists, what it typically excludes, and what changes if your strategy predates October 5, 2026.

What the floor measures

The liquidity floor uses a specific measure: median daily dollar volume, defined as closing price times shares traded in US dollars, calculated over the asset's last 60 trading days as of the ranking date. The result must be at least $1 million for the asset to be eligible.

Using the median instead of the average makes the threshold more stable. A single unusually heavy trading day, such as a news event or an index rebalancing, can push an average well above the typical level for a thinly traded stock. The median takes the middle value across all 60 observations, so a handful of outlier sessions do not carry an asset across the floor.

The 60-day window corresponds roughly to three calendar months of trading activity. That span is long enough to capture a consistent picture of how actively a market trades a given asset, rather than a snapshot tied to one unusual session. As of the ranking date means the measurement always reflects recent conditions: an asset that was liquid several years ago but has since lost trading volume does not carry the benefit of its historical activity forward.

If an asset does not have enough price history to compute the 60-day figure at all, whether it was recently listed or whether Acubic's data series has an insufficient number of observations, it is treated as not eligible.

Why the floor exists

The floor exists so that any holding the optimizer selects can actually be traded at a reasonable size. As described in the article on Acubic's risk constraints, a portfolio that includes a highly illiquid name looks fine on paper but becomes a problem at execution time, when the order size is larger than the typical market for that name can absorb without moving the price. The floor removes that class of holding before any other logic runs, so the system never has to manage a position the broker cannot fill cleanly.

This is an execution constraint, not a quality judgement. An asset below the floor is not necessarily a poor investment: it may be a smaller company whose shares simply do not trade in large volume, or a recently listed instrument that has not yet built a trading history. The filter asks only one question: is this asset traded actively enough to be held in a managed portfolio under normal market conditions? It makes no prediction about which assets will perform well or poorly.

Acubic does not predict short-term market moves at any stage of this process. The floor is a practical constraint on what can be held, applied before any scoring or weighting begins.

Where the floor runs in the selection process

The selection process runs in a fixed order: screen, rank, allocate. The liquidity floor is the first step in the screening stage, applied to every asset in the starting universe before anything else runs. As described in the article on asset selection, screening is a hard filter: a ticker either passes or it does not, and those that fail are removed before ranking begins.

The liquidity floor runs before your sector, region, and asset class preferences. This matters: it is not possible for a constraint you set during the quiz to override it. If an asset you expected fails the floor, it will not appear in a proposal even if your preferences would otherwise allow it.

After the floor passes its eligible set to the screening stage, your universe constraints are applied, and the surviving assets are ranked using whichever ranking method you chose: market capitalisation, Sharpe ratio, momentum, or the composite quantitative ranking. The allocation step, which assigns weights using the method you chose such as hierarchical risk parity or mean-variance optimisation, only ever sees assets that cleared every prior stage. The full pipeline is documented at Acubic's methodology page.

Keeping the floor at the front of this sequence is a deliberate choice. If the floor ran later, after your sector or region filters, it would only catch illiquid assets that survived those filters. Running it first means no subsequent step ever evaluates an asset that cannot be traded, which removes a whole class of construction edge case from the ranking and allocation stages.

What the floor typically excludes

In practice, the floor removes very small companies and recently listed stocks whose trading activity has not yet reached a consistent level. Micro-cap equities, thinly traded sector funds, and instruments that trade primarily on non-US exchanges in smaller volumes often fall below the threshold. Any ticker for which Acubic does not have a full 60-day price series is also excluded.

The floor does not remove assets on the basis of sector, country, asset class, or any ranking signal. Those are separate filters applied in the screening stage that follows. An asset that clears the floor may still be removed by a sector exclusion you specified, by an asset class filter, or by insufficient price history for the evaluation window on a Sharpe or momentum ranking. The floor only addresses one question: trading activity.

The floor also does not distinguish between stocks and exchange-traded funds, or between different market segments within those categories. A major ETF tracking a broad index will typically have daily dollar volume many orders of magnitude above $1 million and clear the floor without difficulty. A niche ETF in a specialist sector with limited daily trading may not.

The two exemptions: currency pairs and locked tickers

Two categories are exempt from the liquidity floor. The first is currency pairs. Currency pairs serve a specific role in portfolio construction that differs from equity and fund holdings, and the trading volume of a currency pair is not measured in the same way as the daily dollar volume of a listed equity. They are excluded from the floor's scope.

The second exemption is any ticker you have locked into your portfolio. A locked ticker is one you have explicitly pinned in place, telling the optimizer to hold that name regardless of how the ranking process would otherwise treat it. Because you have made a direct, deliberate decision about that holding, it bypasses the floor. The rationale is that the liquidity floor protects the optimizer's decisions from an execution problem you are unaware of; a ticker you have consciously chosen to include is a different kind of decision.

If a specific ticker matters to your strategy and you want to include it despite falling below the floor, the locked ticker feature is the route. The trade-off is that you are overriding a constraint designed to protect execution quality. It is worth checking why a ticker falls below the floor before deciding to override it: a name with very low recent trading volume may be harder to execute at the quantities a rebalance requires.

Strategies created before October 5, 2026

The liquidity floor was introduced for strategies created or regenerated on or after October 5, 2026. The methodology page documents the transition rule precisely: a strategy saved before that date keeps the rules it was saved with in its live re-rankings until the strategy is regenerated. A new backtest of the same strategy applies the floor, because backtests always reflect the current methodology.

This creates a situation where, for a pre-October 5 strategy you have not regenerated, the backtest and the live portfolio can select different assets. The backtest applies the floor and excludes assets below it. The live re-ranking does not, because it is running under the rules that were in effect when the strategy was saved.

If you want your live portfolio to reflect the same methodology as a backtest run today, regenerating the strategy from your current inputs brings the two into alignment. The methodology change log on the methodology page records the October 5 update alongside the other changes introduced on that date: risk-level caps on by default and consistent closing prices across stock splits.

For strategies created from October 5 onward, there is no transition to manage. The floor applies to both the live re-rankings and to any backtest, so the two always use the same eligibility rules.

What you see when an asset fails the filter

An asset that fails the liquidity floor is simply absent from the proposal. It does not appear at a reduced weight, and no specific warning identifies it as the reason. The most visible outcome is that a ticker you expected to see is not there.

The guide to reading an Acubic portfolio proposal covers the main reasons a ticker might be missing. The liquidity floor is one; a universe constraint that removed the asset before ranking is another; a minimum weight floor that the optimizer could not satisfy given the other constraints is a third. These can overlap: an asset that cleared the liquidity floor may still be removed by a sector exclusion you set.

If the data pipeline and the methodology page do not resolve a question about why a specific ticker is missing from your strategy, regenerating the strategy and observing which assets change between the old and new proposals is a practical diagnostic step. Any asset that was present before regeneration and absent after was not screened out by a user constraint, and the liquidity floor is the most likely explanation.

The filter in the broader construction context

The liquidity floor is a data quality and execution constraint. It does not select assets for their return potential, it does not predict which assets will perform well, and it makes no claim about which assets are more or less attractive investments in a broader sense. Its function is narrow: ensure that the assets the optimizer selects are ones that can be traded.

The complete guide to how Acubic works places the filter in the context of the full pipeline: the goal throughout is a documented, repeatable process, not a hidden signal or a forecast. The liquidity floor is part of that transparency, stated in the methodology, applied consistently, and exempt only in the two cases where the rationale for the floor does not apply.

Build a portfolio at the Acubic portfolio builder, inspect the resulting proposal against your inputs, and read the methodology page if you want the full definition of every filter that runs before the first weight is set. The article on what data Acubic uses covers the broader data pipeline, including where historical prices and company data come from.

Acubic does not promise returns and cannot predict short-term market moves. The liquidity filter is one part of a construction process designed to produce a portfolio that can be held and managed in practice, not an assurance about what any selected or excluded asset will do.