Dynamic & universe strategies

Who it’s for

Users who want one strategy to trade a selected set of instruments rather than one

Assumes

You have built a single-instrument strategy

A dynamic strategy trades a universe rather than a fixed symbol: it evaluates every candidate on each bar, ranks them, and trades the top-ranked ones. Instead of naming an instrument, you name a selection rule.

Triggering one​

Give an instrument selection rule instead of an instrument:

build a momentum strategy across all large-cap crypto
buy the top 5 NIFTY 50 stocks by relative strength
trade the most active NSE stocks on a breakout signal

The builder recognises that you gave a selection rule and resolves a dynamic universe rather than asking you for a single asset. Dynamic strategies carry a Dynamic badge on the strategy card.

Not the same as a multi-symbol backtest​

The distinction is subtle and it matters more than anything else on this page.

How the universe resolves​

The candidate pool​

The pool comes from the platform's supported-asset catalogue, the single source of truth for every tradeable asset, with nothing hardcoded per universe. Both the live path and the backtest resolve members through the same code, so a universe means the same thing researched and deployed.

Note the third row: "NSE stocks" means the supported NSE universe, not an index. If you want an index's constituents, name the index.

Crypto pairs are de-duplicated by base asset. The catalogue lists both BTC_USDC and BTC_USDT; for ranking purposes one is used, because ranking the same asset twice would distort the selection.

Membership over time​

Universe membership has a timeline, and there are two modes:

Ranking​

Each bar, candidates are ranked by a metric you choose.

Why volume means traded value​

Worth reading, because it is a genuinely good decision:

"Most active / most traded" is liquidity, which is a multi-day property measured by average daily value (close × volume) over about 20 days. Not one bar's raw share count. A single freak bar would otherwise crown an illiquid name.

Using traded value also makes equity and crypto comparable, which raw share counts are not. And it is the same measure as the eligibility floor, so "most active" and "tradeable" agree rather than selecting different sets.

Missing metrics fail closed​

A metric whose input is absent (no stats feed, or warm-up incomplete) returns NaN, and the resolver treats NaN as rank-last or drop. It never treats a missing value as zero or as neutral.

delivery_volume and oi_change have no feed yet, and the validator says so rather than silently ranking everything equally.

Selection and allocation​

Once ranked, the strategy trades the top N. Capital is divided across selected positions from the strategy's single allocation, the allocation is never exceeded, and the two capital gates that enforce that cannot be disabled.

Selection is re-evaluated per bar, so the traded set changes as ranks change. The practical consequences:

Turnover. A strategy that reselects frequently trades more, and cost drag scales with turnover. Watch the gross-to-net gap. See Fees & charges

Correlated exposure. The top five by momentum in one sector is one bet with five names. Ranking does not diversify.

Reading a dynamic backtest​

The Markets tab shows per-asset performance, which instruments carried the result and which dragged.

Also read:

Limitations​

Next​

• Advanced strategies
• Backtest limitations
• Position sizing