The Intraday Path Of Spreads And Depth

Trading conditions in a stock are not the same at 9:35am as they are at 1:00pm. Anyone who has watched a quote screen through a full session notices spreads flare up right after the open, settle down through the middle of the day, and widen again as the close approaches — often traced out as a rough "U" or "smile" shape when plotted across the day. Depth follows a mirrored pattern: thin at the open, building through the day, and often thinning again right before the close as market makers manage risk ahead of the final print. Knowing this shape is one of the simplest, most reliable pieces of microstructure knowledge a trader can use.

Why the pattern exists

At the open, overnight news and pre-market activity have to be absorbed, uncertainty about the "right" opening price is high, and market makers widen spreads and post less size to protect themselves from getting run over by informed order flow before the day's information has been digested. As the session progresses, information gets incorporated into prices, uncertainty falls, more participants are actively quoting, and competition among liquidity providers compresses spreads while depth builds up. Near the close, two forces push the other way: many funds route size into the closing auction rather than continuous trading (reducing continuous-market depth), and market makers face renewed uncertainty about overnight risk they're about to be stuck holding, so they widen out again in the final minutes, especially right before the market-on-close order imbalance is published.

Worked example

A liquid mid-cap stock typically quotes a 4-cent spread with 500 shares displayed at the open, tightens to a 1-cent spread with 3,000 shares displayed by 11am, holds that through early afternoon, and widens back to 2.5 cents with 1,200 shares displayed in the last 10 minutes before the close. A trader needing to execute a large order with minimal cost, and with no informational urgency, would get a materially worse fill trading the same size at 9:31am than at 11:15am — roughly 4x the spread cost and a sixth of the available depth at the open compared to midday. Scheduling that trade for the midday window, or splitting it to avoid both the open and the last few minutes, directly exploits this well-documented pattern rather than fighting it.

What this means in practice

Execution algorithms bake this pattern into their scheduling by default, participating less aggressively in the volatile open and close and more in the calmer midday window, unless the trader has a specific reason to prioritize the open or close (an index rebalance, a signal that decays fast, a need to trade at the closing auction price). Cost estimation models that don't account for time-of-day are missing a real and predictable driver of execution cost — the same order can cost meaningfully more or less purely based on when in the session it's placed, independent of the stock's average daily spread or volume.