Simulated line, live
Observed price (the line)
True signal, fully revealed
Pure noise benchmark
Signal correlation (r)
—
line moves vs. true signal moves
Informativeness (r²)
—
% of line movement explained by signal
Aggregate position (Q)
0.00
net crowd lean, this tick
Ticks simulated
0
rolling ~10s read window
WARMING UP
Collecting ticks before the correlation read stabilizes…
What's actually being simulated: a common signal (slow-decaying, everyone sees it) that agents forecast over their own horizon and trade on — plus a fraction of their own anticipated market impact. Their aggregate position pushes the observed price. As impact-awareness rises, agents increasingly hedge against the footprint they know they're creating, which — per the paper's balance condition — makes the direct transmission of the signal cancel out of the price, leaving mostly the martingale (pure noise) component behind.
Low awareness
Positions track the signal directly. The line leans toward truth and stays there — informative, but easy to front-run.Rising awareness
Agents shade their trades to offset their own impact. Signal and impact partially cancel in the price.Full awareness
Cancellation is complete for the direct term. What's left is close to a random walk — the line moves, but it isn't telling you anything.