We published a report yesterday taking a look at the state of on-chain margin trading and perpetuals to try and find the most likely winners. The report — available for Delphi Pro subscribers — runs a valuation exercise on DYDX and looks at the the potential for growth for three other protocols we believe stand the best chance of taking a large part of the market.

For those of you not subscribed to Delphi Pro, the other three protocol are: Synthetix, Vertex, and GMX. It’s a pretty even split of ideologies. GMX and Synthetix are pool-style venues that rely on synthetic pricing. DYDX and Vertex are orderbooks.

In my opinion, the future of this space lies with the orderbook protocols. And it comes to down to one very simple thing — price discovery.

GMX and Synthetix facilitate huge volumes, but they are incapable (intentionally; by design) of facilitating price discovery. They rely on oracles — and indirectly, CEXs — in order to set fair prices for their users. This reliance on CEXs for arguably the most important part of the DEX (pricing assets) is a crutch that hinders these platforms. Quite simply, they can never be bigger or even as big as the venues they pull their pricing from. LPs become vulnerable to rather simple economic exploits if liquidity on GMX and Binance are the same.

But wait, don’t orderbook protocols also use oracles? Yes, but not in the same way. dYdX uses oracles to establish an index. That index is compared against dYdX’s native pricing to determine the funding rate for traders. Every single CEX also does this; it’s a necessity to determine optimal funding rates. GMX and Synthetix use oracle price feeds as the price of assets on the venue. The oracle price is thus the price at which you can long/short an asset.

Now, this isn’t an immediate concern. DEXs own such a tiny share of the market right now that their best chance of taking market share from CEXs is probably through protocols like Synthetix and GMX. So this doesn’t affect my view on mid-term growth whatsoever. But over the long long term, dYdX has no ceiling — the virtual pricing protocols like GMX do.

dYdX is already the biggest of the on-chain perpetuals protocols. But one could argue that their most successful iteration isn’t fully on-chain — and they would be right. dYdX v3 relies on a centrally-run matching engine and orderbook. But dYdX v4 makes the switch to a fully on-chain architecture. If v4 can attract users the way v3 did, there’s little to stop dYdX from being the reigning king of on-chain perps for the foreseeable future.

I could be wrong with my view here — as I often am. But as far DeFi’s best bet at beating CEX volumes go, I am firmly in the orderbook camp. For now at least.