Liquidity when a stock drops: the bid thins on 95% of sessions

Liquidity is usually discussed as a property of a stock. It is better understood as a property of a moment. The same ticker that shows a comfortable bid all morning can show a fifth of it during the ten seconds you actually want to sell. We measured how much money the bid actually holds, second by second. This is the companion to https://traderwe.com/research/stop-loss-slippage-the-spread-widens-with-the-drop-that-triggers-it which measured the price you cross when a stock falls. This one measures how much of it there is.

1. What is actually resting at the bid

Take the size quoted at the best bid, multiply it by the bid price, and you have the dollars a market sell can hit before it has to walk down to the next price level. The median second in our recording had $2,300 standing at the front of the book. That is the entire cushion at the quoted price. Anything larger than that is not a trade at the quoted price, it is a trade that starts at the quoted price.

2. It thins as the fall gets bigger

Split the falling seconds by how far the price actually fell in that second. Every column falls monotonically. The median drops by 3.8 times between the smallest and largest buckets, and the lower quartile drops from $780 to $190. Read the last row as the one that matters. In a quarter of the seconds where a stock fell 1% or more, the best bid holds less than $190. A position of any size is not exiting into that at the quoted price.

3. Put it next to the spread

The companion study measured the other half of the same instant. On falls up to 0.25% the quoted spread averaged 0.50%. On falls of 1.00% or more it averaged 2.11%. Both move against you at the same time, and they multiply rather than add. The price you must accept gets worse, and the amount available at that worse price gets smaller, so a larger order has to walk further down a book that is also more expensive to walk down. This is why exit cost is not a fixed percentage you can subtract from a strategy. It is a function of the event you are exiting into.

4. Does it hold on ordinary days

Pooling every falling second across the sample can flatter a result: days with violent falls also tend to have thinner books all day, so part of the 3.8x gap is composition between days rather than thinning within one. So we measured each session separately. On 35 of 37 sessions with enough observations, the median dollars at the bid was smaller on falls of 1.00% or more than on falls up to 0.25%. That is 95% of sessions, and the two exceptions were near ties rather than reversals, the weaker one at 0.95. The size of the effect inside a single session is smaller than the pooled figure. The median session showed 1.63 times less money at the bid on the big falls, against 3.8 times when every second is pooled. Both numbers are real and they answer different questions. The direction is what survives everywhere, and the pooled ratio is the one that describes the population a trader is sampling from across many days rather than what a single afternoon looks like.

5. The same measurement, by price

Higher-priced stocks carry more dollars at the touch, which is arithmetic more than quality: one round lot of a $100 stock is $10,000 and one round lot of a $2 stock is $200. The $100 and up row rests on 3,985 down-seconds, the thinnest slice in the table, and its upper quartile sits above the $20,000 ceiling we counted to, so that cell has no number rather than a guessed one. The practical reading is the middle of the table. Between $2 and $20, which is where most day trading scan lists live, the median second offered between $1,530 and $1,740 at the front of the book while the stock was falling.

6. The book behind the front row

The front row is not everything. Summing the size across all quoted bid levels and pricing it at the best bid, the median falling second showed $145,700 of total bid interest, and the lower quartile showed $32,200. So the book is deep and the front row is thin, at the same time. The gap between $1,810 at the touch and $145,700 across the book is the distance a large order has to travel, and every step of that travel is a worse price. Depth further down is not a substitute for depth at the touch. It is the schedule of what the rest of your order will pay.

7. What this does not measure

We recorded what was quoted, not what was executable. A quote can be pulled in the milliseconds between the snapshot and an order arriving, and our one-second grid cannot see that. The numbers here are therefore an upper bound on what was available, not a promise. Hidden and undisplayed liquidity is also invisible to us. Some orders will fill better than this table implies because something not shown in the book met them. Both limitations point the same way: the displayed book is the optimistic case, and it is already thin at the moment you most need it.

8. What to do with this

Size the exit, not just the entry. If a position is worth more than the front row of the book on a 1% down-second, the exit is a multi-level walk by construction, and it should be planned before the position exists rather than discovered during it. Check the touch before entering, not the average. A stock can average a healthy bid across the day and still show $190 in the seconds that matter. The relevant number is what is standing there when the stock is moving, and that is what this table reports.

9. Sample accounting

The sample is 38 sessions drawn evenly from 264 recorded sessions between 2025-04-02 and 2026-07-27, giving 745 symbol-days, 4,400,640 recorded seconds and 387,948 falling seconds. The universe is the stocks our own scan surfaced each day, which are the day's active movers rather than a cross-section of the market. Only the regular session is included. Seconds were used only where the quote was complete, the size was positive, and the two snapshots were exactly one second apart. Dollar figures are the quoted size multiplied by the best bid price. All figures are medians and quartiles rather than means, because a small number of very deep seconds would otherwise dominate the average and describe a stock nobody was trading. We did not measure returns, entries, or whether any of this can be traded profitably. This is a liquidity measurement.