← All articles

The real cost of copier latency (it's slippage, multiplied)

When your leader fills and a follower's order arrives 500ms later, the follower isn't trading your setup anymore — they're trading your setup half a second late, at whatever the book looks like now. That difference has a price, and because a copier multiplies it across every account, it compounds faster than most traders expect.

The napkin math

NQ moves fast: in an active regime it's common to see several ticks change hands inside half a second. Say a 500ms-late entry costs just 1 tick of NQ slippage ($5) on average. Run 8 follower accounts, 3 round-trips a day (entry + exit both slip): that's 8 × 3 × 2 × $5 = $240/day, roughly $5,000/month — paid invisibly, one tick at a time, by a copier that "feels fine." A copier that holds its own share of the delay to ~40ms puts most of that back: at those speeds the book usually hasn't moved a full tick yet.

Your numbers will differ — instrument, session, size, regime. The point isn't our estimate; it's that the cost scales with account count, which is exactly when you're using a copier.

Separating the copier's share from the broker's

Total delay = the copier's processing + the broker's network round-trip. The second leg belongs to the broker: it varies by gateway and by session, and it hits every copier equally. So the number a copier owes you is its own share — detection, sizing, risk checks, submission. That's the number we measure on every production order and publish live at /latency (median ~42ms), with the broker's wire attributed separately on each Copy Log row, so a slow gateway is never disguised as engine speed — or vice versa.

Questions to ask any copier (including us)

• Show me per-order latency in the product. Not a benchmark — my orders.
• Show me your live aggregate numbers, from production, where I can check them.
• How do you attribute broker wire time vs your own processing?
• What did your worst 5% of copies look like last week?

Any vendor with good numbers will be glad you asked. Any vendor without them will change the subject to features. That asymmetry is the whole reason we publish ours: the first month is free, so the cheapest way to know is to run one real trade through it and read your own receipts.