TVL, discounted by how
long it lasts.
Liquidity discounted by expected persistence. A $500M pool on a nine-day horizon at confidence 0.7 reports $35M. The same pool at ninety days reports $350M.
TVL counts the wrong thing.
Protocols evaluate incentive programmes on TVL acquired per dollar spent, which rewards buying liquidity that leaves the moment payment stops.
A campaign that triples TVL while dTVL stays flat has bought a number on a dashboard.
TVL counts the wrong thing.
Aggregated across a chain, dTVL behaves nothing like headline TVL during a programme.
How this works
dTVL discounts liquidity by its horizon relative to a reference duration and scales by the published confidence. Emission efficiency then measures durable liquidity-days bought per dollar of emission.
01
Takes the horizon
The published horizon and confidence for each pool, as they stood at each point in time.
02
Discounts the liquidity
Scaled against a ninety-day reference, with an exponent controlling how sharply short horizons are penalised.
03
Publishes it unrequested
Emission efficiency is published for every observable campaign, whether or not the protocol asks. Computed only on request is computed only when flattering.
What it reads, derives,
and does with it.
What it reads
Published horizons and realised liquidity.
What it derives
What it computes from them.
What it does with it
What gets published.
“The visible depth of a pool and the durability of that depth are different quantities, and on-chain systems currently conflate them.”
Every covered pool, scored in public.
Liquidity Intelligence Infrastructure. Read the horizon before the capital moves, not after.