MARRESFLOWSHOCKRequest sample
HOW IT WORKS

See the liquidity behind the spread

Flowshock simulates trading in an order book. Follow a price move, inspect the shares available to buy or sell, and compare that same minute with the historical market.

See the one-cent spread example ↓
One session. Many possible paths.AAPL · MIDPOINT PRICE · USD
274276278280282284World 1World 2World 3Historical
10:05 NY16:00 NY
━━ Historical━━ Three sample worlds
THE MARRES METHOD

Simulate the orders behind the price

In Flowshock, simulated traders place orders into a shared book with finite buying power and inventory. As orders arrive, trade or are cancelled, the best quotes and the size available at each price change together. That lets you trace a price move back to the liquidity available at the time.

REFERENCE

Start from an observed book

The AAPL pilot starts from a historical snapshot. Each world begins with the same configuration, so differences between runs can be examined on a common footing.

SIMULATION

Let trading shape depth

A quote can stay in place while the shares behind it disappear. Follow both sides of the book to see whether liquidity holds as trading unfolds.

COMPARISON

Check against history

After warm-up, compare each world with the observed market at the same minute and across the same price levels. Then check how often shallow books occurred over the session.

WHAT FLOWSHOCK CAN DO

What can the order book tell you about liquidity?

Start with a quote, follow depth through the day, then compare each simulated world with the observed session.

INSIDE THE ORDER BOOK

What sits behind the best quote?

At 12:25 New York time, the simulated spread was one cent and the historical spread was two. Yet across ten levels per side, the simulated book held 93% fewer shares — 599 versus 8,116. Move the slider to see where that difference appears.

AAPL · 9 DEC 2025 · 12:25 NY

Cumulative bid and ask shares · common scale of 8,116 shares
ACROSS THE SESSION

When does liquidity thin?

Switch worlds to see how depth changes. Amber dots mark snapshots below that world’s 10th-percentile depth, calculated from the one-minute sample.

How to interpret the comparison

Each world uses the same configuration and a different random seed. A similar distribution of depth can still hide differences in when liquidity disappears.

DISPLAYED DEPTH · SHARES
10:05 NY15:55 NY
Five-minute preview · ten levels per side
HISTORICAL BENCHMARK

How much depth was typical?

A thin minute can be an exception. Compare median depth over the measured session to see whether each world usually held as much stock near the quotes as the historical book.

Inspect the distributions →
MEDIAN DISPLAYED DEPTH · AAPL
Historical8,488 shares
World 11,174 shares
World 21,310 shares
World 31,396 shares
10:05–16:00 NY · shares across ten price levels per side
HISTORICAL FIT

How closely does a world match the observed day?

Flowshock compares prices, depth and returns over the same trading window. Choose the measure that matters to your analysis. A world can track historical prices closely while offering much less liquidity.

Compare the three sample worlds →
WORLD 1 VS HISTORY · AAPL · 9 DEC 2025
Distribution distance (Wasserstein W₁)
MeasureDistance
Midpoint price1.64 USD
Displayed depth7,761 shares
One-second returns0.0578 bps
Lower values indicate a closer match within each measure. Zero means the two observed distributions coincide. Matching distributions can still have different timing.

What would you like to evaluate?

Inspect minute-level observations from three sample worlds. Send Marres your request for review.

Request sample →

Have a particular instrument or venue in mind? Tell Marres which market you follow and the liquidity question you want to explore.

Ask about your market ↗