Slippage on news events: evaluation model vs. live execution parity
I've been stress-testing a few of the newer two-phase evaluation models that claim direct market access, but there's a recurring discrepancy during high-impact liquidity gaps. On the demo servers used for the challenge phase, fill latency during CPI or FOMC seems artificially optimized, giving near-instant fills with minimal slippage on instrument pairs like $EURUSD and $GBPUSD. However, once transitioned to the simulated live phase—which supposedly routes through their primary liquidity provider—the execution delays stretch to 180-250ms, resulting in significant negative slippage that destroys the edge of any news-momentum strategy. I'm trying to figure out if this is a deliberate synthetic delay introduced via their bridge software to manage risk on funded books, or simply a byproduct of poor prime-of-prime integration during peak volume. Are any of you running cross-broker latency arbitrage checks during evaluations to verify if the feed matches true retail raw spreads, or are we just accepting these wider B-book spreads as a cost of doing business with these firms?