Quantis Trade
← Back to all metrics

Metrics / Turnover

Turnover

4 min read · Metric

How much you traded relative to your capital. The lever that converts gross alpha into net P&L.

What it is

Turnover measures the dollar volume traded over a period, normalised by initial capital and annualised. A 100%/yr turnover means you bought and sold the equivalent of your starting capital once over the year; 600%/yr means you turned over six times.

Turnover is the number that decides whether your gross backtest survives the transition to live trading. A 1.5 gross Sharpe at 100%/yr turnover loses very little to costs; the same Sharpe at 800%/yr turnover often nets out at 0.5 or worse depending on the round-trip cost assumption.

Formula

# Total dollar volume bought (or equivalently, sold)
total_buys = sum(absolute_value(trade) for trade in buys)

# Annualised, expressed as a fraction of initial capital
turnover = total_buys / initial_capital / years_in_sample

# Sometimes reported as round-trip turnover:
round_trip_turnover = (total_buys + total_sells) / initial_capital / years

Quantis uses the one-way (buy-side) convention so the number maps directly to “how many times did I cycle through my capital this year.”

Typical ranges

  • Buy-and-hold: 0–10%/yr (only corporate actions and rebalances).
  • Quarterly-rebalanced factor portfolio: 50–150%/yr.
  • Monthly cross-sectional momentum: 300–600%/yr.
  • Daily mean-reversion: 600–1500%/yr.
  • High-frequency intraday: 5000+/yr.

Common mistakes

  • Quoting gross Sharpe at zero costs. For a high-turnover strategy this is fiction. Always net by realistic per-trade cost (typically 5–15bps round-trip on liquid US equities; more on small caps or international).
  • Confusing one-way and round-trip. 300% one-way = 600% round-trip; some sources use one, some the other. Quantis uses one-way; cost models usually want round-trip.
  • Ignoring market impact for size. A 600% turnover strategy that backtests fine at $100k might be impossible at $10M because your trades start moving the bid. Liquidity caps matter.
  • Zero turnover with non-zero return. Suspicious — usually means the strategy never actually traded but corporate-action returns leaked into the equity curve. Investigate.

What the platform flags

Quantis applies a per-trade cost model and reports both gross and net Sharpe so the impact of turnover is explicit. A “0%/yr” turnover badge gets a special “— no trades” treatment because it usually indicates the backtest never executed any rebalance — the metric is unreliable rather than impressive.

Further reading