Perp DEX vs CEX for Trading Competitions

Voltrade's venues directory currently carries 39 active venues — 23 decentralised and 16 centralised. Only a subset of either group can host a competition, and the reason is not partnership status or business development. It is a technical property: can an outside party establish what a given trader traded, over a given window, with the trader's consent and nothing more?

Answer yes and a competition is possible. Answer no and no amount of goodwill fixes it. This is the structural comparison behind every venue page on this site.

Custody, and why it decides the rest

The first-order difference is where the assets sit. On a perp DEX, positions are held against a wallet or an account the trader controls, and settlement is on-chain. On a CEX, the exchange holds the balance and the trader holds a claim on it.

That is the familiar framing, and it matters for its own reasons. But for competitions it matters mainly because of what it implies about identity. On a DEX, the trader already has a public, unforgeable identifier — the address — and proving control of it is one signature. On a CEX, the trader's identity inside the exchange is an account number that means nothing outside it, and there is no signature that proves control of an account.

Everything about how competitions differ across the two follows from that.

Eligibility: wallet-only versus account linking

This is the difference competitors feel first, in the entry flow.

Wallet-only venues need nothing beyond the wallet you register with. On Hyperliquid the trading address is your EVM wallet. On GMX the registered wallet is the GMX trading address. On Lighter the registration wallet resolves to Lighter account indices through a public endpoint. On Convallax, participants are simply those who registered with the wallet they trade with — nothing to link. On TurboFlow, scoring reads the registered wallet. In every case, connect, register, trade.

Linked-account venues need a second identifier that the wallet cannot supply. dYdX needs a dYdX address. Blofin needs a Blofin UID, linked after signing up under the affiliate code. KuCoin needs an OAuth consent that mints a per-user read-only key.

Two things follow that sponsors consistently underestimate.

Linking is where the funnel leaks, and where it leaks differs by venue. dYdX and Blofin enforce the link at the door: joining without one is rejected with a 400 naming the missing integration, so the loss happens before registration and shows up as a smaller participant count rather than a wrong one. KuCoin has no such gate — someone can register and only later discover they never granted the OAuth consent that makes them scoreable, which is the case where a registrant genuinely believes they entered and did not, in the only sense that counts. Either way the linking prompt is the single highest-leverage message in the whole campaign.

Linking is also a filter, and filters are not purely bad. Someone who finds and links their dYdX address or their Blofin UID has an account and intends to use it. You trade raw registration count for a higher share of registrants who actually trade. Which side of that trade you want depends on whether your campaign goal is reach or activity.

Note that "DEX" and "wallet-only" are not synonyms: dYdX is a DEX that requires linking, because its account address is not your EVM wallet. The custody model and the identity model are related but not the same thing.

Data availability: the property that actually gates a venue

Here is the hierarchy, from easiest to hardest to score, drawn from how the integrations are actually built.

Tier 1 — public fills. Hyperliquid publishes a full on-chain order book, so a trader's fills and open positions are readable from an unauthenticated endpoint. This is the only tier that comfortably supports profit competitions, because PnL scoring needs realized PnL, unrealized PnL on open positions and cumulative volume, all reconcilable per participant. One public endpoint provides all three.

Tier 2 — public aggregates. Lighter's per-account trade history is auth-gated behind each account's own API key, and its public recent-trades endpoint returns only the last 100 trades per market with no pagination. What is public is a wallet-to-index lookup and a monotonic lifetime volume counter, so per-day volume is derived from differences between two snapshots of that counter. Robust — a missed sample costs nothing because the counter catches up — but it is an all-markets aggregate, so per-market scoping needs a second mechanism entirely. Lighter's is a block-explorer history replay, which happens to serve complete history for any account unauthenticated.

Tier 3 — an external database. Convallax is scored from an external read-only Postgres rather than an HTTP API, summing trade premiums per participant wallet. Unusual, and it works, but it exists only because the venue was willing to expose it.

Tier 4 — authorised private reads. KuCoin's model. There is no public data at all, so the trader authorises a read: an OAuth flow mints a per-user API key restricted to a single read permission group, with spot trading, margin, futures trading, earn, transfers and withdrawals all explicitly disabled. The key is verified with a live call before storage, and stored encrypted.

Tier 5 — a third-party report. Blofin's model. Scoring never touches the trader's account; it reads affiliate reporting endpoints about traders under the affiliate. It works, but it inherits whatever shape the report has — which in Blofin's case means no per-market filter, no spot/perp separation, and daily buckets that land at 16:00 UTC rather than midnight.

Tier 6 — nothing. Most venues in the directory sit here. There is no public data, no partner reporting, and no per-user read path. Those venues appear in the catalog for comparison and are not competition venues, and that is why: not unwillingness, just no way to establish what happened.

What each tier can and cannot host

The row that surprises people is the last column. Affiliate-report venues can run a perfectly good whole-exchange volume competition. They cannot run a campaign about one token pair, because the data does not carry a market. If your brief is "drive volume into our pair", that rules out an entire tier before you have talked to anyone.

Fee models are not comparable across the line

A trap worth naming, because the catalog invites it.

The directory lists dYdX at 0.02% maker / 0.05% taker, Hyperliquid at 0.02% / 0.05% with an annotation that this is a base tier which is volume-scaled and subject to change, GMX at 0.01% / 0.06% with an annotation that its fees are dynamic by market, pool and action rather than fixed maker/taker, and KuCoin at 0.1% / 0.1%. Some venues — Blofin and Lighter among them — carry no fee figures at all, so we quote none.

Reading those as four comparable numbers is a mistake in at least three ways. A dynamic pool-and-action fee is not the same kind of quantity as a maker/taker split. A base tier is the top of a ladder that any serious competitor will be several rungs down. And a spot fee on a CEX and a perp taker fee on a DEX are charged against different notional bases entirely — a perp fee applies to leveraged notional, which is why Voltrade's own points model rates perpetual volume at one tenth the rate of spot volume.

For a sponsor the useful question is not "which venue is cheapest" but "what will the participants who actually move this leaderboard be paying", and the honest answer usually requires asking the venue. That is why every venue page here links the directory rather than freezing a number in prose.

Choosing a side

Pick a wallet-only DEX when turnout is the goal, when you want a profit competition rather than a turnover one, or when you want participants to keep custody throughout. The absence of a linking step is the biggest single lever on registration-to-participation conversion, and it removes the whole class of "registered but never scored" outcomes.

Pick a CEX when your token trades there and nowhere on-chain that matters, when your audience already holds balances there, or when the market you care about — a specific listed pair — only exists there. Accept that you are paying for it in funnel friction and, on affiliate-report venues, in scoring granularity.

Pick neither if the data does not exist. A competition that cannot be scored is not a competition; it is a promise. That is the whole reason this hierarchy is worth understanding before you pick a venue.

Going deeper

Venue-by-venue comparisons: Hyperliquid vs Lighter, Hyperliquid vs dYdX, Lighter vs dYdX, GMX vs Hyperliquid and KuCoin vs Blofin.

For the mechanics that sit above the venue choice, read how DEX trading competitions work, volume vs PnL competitions, and how VXP and daily volume caps work. The complete guide is the overview, and the venues directory is the live catalog.