← Research logMicrostructure & Flow
KILLED · latency
#

One venue leads price discovery (e.g. the highest-volume perp); its moves predict laggard venues by tens-to-hundreds of milliseconds.

$$ \arg\max_{\ell}\ \mathrm{Corr}\big(r^{\text{lead}}_t,\ r^{\text{lag}}_{t+\ell}\big),\quad \ell>0 $$

Synchronized multi-venue tape, cross-correlation lag estimate, trade laggard on leader move.

Detected lead20–120 ms
Latency to act (public API)100–400 ms
Net edgenone
KILLED
The lead exists but is tens of milliseconds — far inside our public-API round-trip. This is a colocation/HFT trade; at our latency the laggard has already caught up. Killed.
A lead measured in milliseconds is only tradeable by someone whose latency is measured in microseconds. Know which game you are in.

We publish the failures too.

This is one of 100+ documented hypotheses. Browse the full lab notebook, or see the strategies that survived into production.