Snatch | Dopamine125@0x_snatch
I'm starting a performance evaluation of currently available perpetual DEXs. Instead of relying on exchanges' marketing claims about speed, I'm measuring real user-perceived latency based on actual API trade responses.
Here are some interesting findings to share.
(For detailed Speed, Spread/Slippage, and Market Depth rankings, see the article below.)
Speed Rankings
1st — @nadoHQ : Trade cycle 268ms (dominant #1), fill_confirm 57ms (#2). Market ack at 211ms is average, but fast fill recognition delivers the shortest end-to-end latency.
2nd — @Lighter_xyz : Market ack 98ms (the only sub-100ms result), cancel ack 54ms — strongest in the ack category. Weakness: fill_confirm at 494ms (REST polling). Strong for signal-based trades.
3rd — @risextrade : Fill_confirm 41ms (shortest), trade cycle 435ms in the upper-mid range. Weaknesses: market ack 395ms, cancel ack 293ms (longest).
4th — @tradehotstuff : Aside from fill_confirm at 48ms (#2), everything else is weak. Market ack of 549ms is the worst (cumulative cost of EIP-712 signing + POST /info), cancel ack 195ms.
5th — @pacifica_fi : Only cancel ack at 45ms (shortest) is competitive. Trade cycle 849ms and fill_confirm 600ms make it the slowest overall.
These are the results from testing during the weekend when liquidity is thin. Additional tests are planned
Please let me know which exchanges you'd like me to include, and I'll add them to the testing lineup.