How the dexwatch score works
Every exchange on dexwatch carries a single 0–100 score. It exists to turn a wall of figures into one honest, comparable read — so here is exactly how it's built, what goes into it, and where it falls short. No black box.
What the score is — and isn't
It is
- A composite of public, measurable figures — liquidity, trading cost, security and track record
- Computed identically for every venue, so scores are comparable
- A fast first read to shortlist exchanges worth a closer look
It isn't
- Financial advice, or a prediction of returns
- A guarantee of safety — no score can rule out a hack or exploit
- Influenced by affiliate deals; commercial relationships are never an input
The four ingredients
The score blends three weighted factors, adds a small baseline so a solid venue starts in fair territory, then subtracts risk penalties. The weights:
The largest factor, because thin liquidity is what actually costs a trader money in slippage and price impact. We blend 24-hour trading volume (60%) with total value locked (40%), each on a logarithmic curve so the gap between a $50M and a $5B venue is rewarded without letting the very largest dominate everything.
Limitation: volume can be inflated by incentives or wash trading, and TVL can be double-counted. Read it as depth, not virtue.
How much the venue has earned trust. Where we have a security read it's used directly; otherwise it's estimated from time in production and audit history — a protocol that has run for years without incident and carries reputable audits scores higher than a brand-new contract.
Limitation: audits and age reduce risk but never remove it. Established protocols have still been exploited.
The headline taker fee, scored so that lower is better and a typical 0.30% AMM fee sits at the bottom of the range. It's the smallest of the three factors because, for most traders, liquidity and safety matter more than a few basis points.
Limitation: it ignores network gas, maker rebates and the price impact you'll actually pay on size.
A fixed +8 baseline keeps a fundamentally sound venue out of unfairly low territory. Then we subtract penalties for genuine red flags: −6 for extreme leverage (above 100×, which invites outsized losses) and −5 for being very new and unproven. Final scores are clamped to the 40–99 range.
Limitation: these are deliberately blunt. They flag risk categories, not the specific risk of any one venue.
The tiers
For quick reading, scores fall into four bands:
Worked example
The clearest way to show there's no magic: two real venues from the current snapshot, scored line by line. Notice how Aster's strong liquidity is pulled back by the leverage penalty.
Figures are an approximate snapshot and change constantly, so a venue's score moves with its data. The calculation, however, never changes between exchanges.
Where the data comes from
Liquidity figures (spot volume and TVL) come from DefiLlama; token prices and market caps from CoinGecko. Fees, leverage, chains, audits and launch dates are compiled from each project's own documentation. Where a figure genuinely isn't available, we show "Not available" rather than guess — we never invent data to fill a gap.
The dataset is a periodic snapshot, so numbers go stale between updates; live price, volume and TVL are refreshed in your browser where possible. Perpetual-DEX volume is a snapshot figure, as the source isn't directly accessible from the browser.
Independence
dexwatch earns through affiliate links, which are always marked. To be unambiguous: affiliate status is not an input to the score or the ranking. The same formula runs over the same public figures for every venue, whether or not we have a commercial relationship with it. Full detail is on the legal & disclosure page.