Capital turnover
Capital turnover describes how much trade notional is executed relative to capital over a stated period. For an AI trading agent, the ratio measures trading activity. Profitability and decision quality require separate evidence.
Published by DXAP, a product of DX Research Group. Reviewed September 20, 2026.
Illustrative 24-hour calculation
An account has $1,000 for the first 12 hours and $3,000 for the next 12 hours. Its time-weighted average equity is $2,000. If the agent executes $800 of opening and closing notional in that day, turnover is 40% for the day. Dividing by the final $3,000 balance would instead produce 26.7%, answering a different question.
$800 ÷ (($1,000 × 0.5) + ($3,000 × 0.5)) = 40%
Illustrative numbers, not customer performance or trading advice.
State what counts as volume
An activity report can sum the absolute notional of opening and closing fills. Under that convention, opening a $200 position and later closing it at the same notional produces $400 of volume. Other reports may count only purchases, openings or one side of a transaction. State the convention before comparing numbers. An order that never fills contributes no executed volume.
Use capital from the same account and period
For a fixed balance, divide the period's executed volume by average account equity. When capital changes, a time-weighted denominator avoids treating a late deposit as though it had been available all day. A daily measure can divide matched volume by equity-days: the sum of each observed equity balance multiplied by the fraction of a day it covers. Mark missing intervals as unknown and explain how much of the period was observed.
Match the agent to its execution account
Use the balance of the account in which the fills occurred. For an agent assigned to a subaccount, unrelated money in its owner's main account stays outside that calculation. Several agents sharing one account share one capital denominator. An active-time report can exclude periods when all linked agents were paused, provided its trade-volume window follows the same rule. State this choice beside the result so readers can distinguish it from whole-account turnover.
Interpret the ratio alongside the strategy
A strategy may wait for an entry or hold an existing position through several review cycles. Both can produce zero new volume while the agent is doing its assigned job. Frequent trades can increase costs or repeatedly change exposure without improving a result. Before comparing two agents, check their intended holding periods and which market conditions permit an entry. A volume target alone says little about whether either agent followed its instructions.
What this does not tell you
Notional already represents the full size of the leveraged position. Multiplying notional-based turnover by leverage again would double-count that effect. All figures on this page are synthetic examples.
Sources and product reference
See how DXAP works
DXAP runs AI trading agents on Hyperliquid using your strategy and configured limits. Read the product overview before choosing whether to use it. Alpha access requires a referral code; agents can lose money.
AI trading agents on Hyperliquid →