MicrostructureKilled

Reading the order book to call the next tick on HYPE

1.8 bps

the strongest order-book signal we found, paid per trade, before costs

bps = basis points = a hundredth of a percent (100 bps = 1%).

In plain English

Reading the order book to call the next tick is the dream of every screen-watching trader. We ran five classic patterns over 7.8 million ticks. The signals are real and rock-solid, and still far too small to cover the cost of the trade. Real, measured, and not bankable.

What was claimed

The order book telegraphs the next tick, so read it fast enough and you profit.

The bar it had to clear

a signal big enough to clear the 6 to 9 bps it costs to trade it

What the data said

under 2 bps at its best, the trade still loses money

Why it's dead

Five classic order-book patterns (sweeps, quote-pulling, queue imbalance, quote staleness, spread regime) tested on 41 days of sub-second HYPE quotes, about 7.8 million ticks. The signals are statistically rock-solid: the best, an 8-tick sweep, showed +1.83 bps at 5 minutes on n=31,832 with t=+10. But rock-solid and tradeable are different things. Under 2 bps does not cover a 6 to 9 bps round-trip.

The detail

Strongest cell: 8 consecutive same-direction mid-moves inside 1 second, forward return rising steadily +0.22 / +0.49 / +0.75 / +1.21 / +1.43 / +1.83 bps from 100ms out to 5 min (n=31,832, t=+10.0). Queue-imbalance quintiles (moments ranked into five buckets by how lopsided the resting buy and sell orders were) gave +0.75 to +1.50 bps. Every cell sits 3 to 5x under the 6 to 9 bps round-trip cost and about 27x under the 50 bps promote bar. Documented, not promoted: the structure is genuine, the magnitude is not bankable.

Kill date: 2026-04-29
Sample: 41 days, ~7.8M ticks
Method: Documented kill
Verdict: real signal, too small

What it would have done to your money

You put in$10,000

You run it for200 trades

You would have

$8,869

−$1,131

lost (11%)

You started with$10,000

This strategy left you$8,869

Instead of not trading$10,000

This strategy had no real edge, so there is no return to compound. The only honest number is the cost: every trade just pays the 0.06% round-trip fee, compounded here. You choose the amount and the time; the bleed is what trading noise costs.