Bid, Ask, Spread and Depth

A quote is two prices, and a spread

The **bid** is the highest price someone is currently willing to pay. The **ask** (or offer) is the lowest price someone will sell at. The difference is the **spread**, and the midpoint between them is the usual reference for a “fair” price. If you buy at the ask and immediately sell at the bid, you lose the full spread. So the spread is the price of immediacy: what you pay for trading **now** rather than resting an order and waiting your turn. Worked example: mid-cap stock — Bid × ask: $42.10 × $42.13 · Quoted spread: $0.03 · Midpoint: $42.115 · Spread in basis points (spread ÷ mid × 10,000): 7.1 bps · Cost of a 100-share round trip: $3.00 ← Basis points (bps) let you compare spreads across prices. One bp = 0.01%. A one-cent spread is 20 bps on a $5 stock but 0.2 bps on a $500 stock — which is why professionals never quote a spread in cents.

Depth, effective spread and price improvement

The quote shows only the best prices. **Depth** is how many shares are available at each price. If you want more than the best level holds, you pay progressively worse prices — that is the subject of lesson M4’s order-book lab. The **effective spread** measures what you actually paid: 2 × |your price − midpoint|. If you buy at $42.12 when the mid is $42.115, your effective spread is one cent, better than the three-cent quoted spread. That difference is **price improvement**, and it is the number an execution-quality report is really about. • Quoted spread: what the market shows • Effective spread: 2 × |fill − midpoint|, what you paid • Realised spread: what the venue earned after the price moved — the market maker’s side of the same story

Why spreads differ, and why they widen

Spreads are wide where providing liquidity is risky: small, volatile or rarely traded stocks, the first minutes after the open, and around news. Market makers widen quotes to cover three costs — holding inventory that may move against them, trading with better-informed counterparties (**adverse selection**), and processing costs. Competition narrows spreads; risk re-opens them. Rules sit at the edges of this. The SEC’s 2024 amendments redefined round lots for high-priced stocks and introduced a half-penny tick for some highly liquid names, with the tick-size and access-fee changes currently slated for November 2026. None of that changes the arithmetic in this lesson: your cost is still spread plus whatever your size costs you. A one-cent spread in a $190 mega-cap is about 0.5 bps. The same one cent in a $2 micro-cap is 50 bps — 100 times more expensive relative to the position. Relative cost, not cents, is the number that matters.

A liquidity tier you can price in one line

You can sort almost every stock you will ever trade into three tiers and carry the expected spread with you. **Tier 1** is the mega-cap with a one-cent spread on a $200 price — about 0.5 bps, the cheapest execution on earth. **Tier 2** is the ordinary operating company, a nickel wide on $50, or roughly 10 bps. **Tier 3** is the illiquid small cap, a quarter wide on $4, or north of 600 bps. The comparison is not academic because the spread is paid on every round trip. Twenty tier-2 round trips a year at 10 bps is 400 bps of return handed to the market — 4 points, before fees, before being wrong about anything. Ten round trips in tier 3 is a different order of magnitude entirely, and it is the single most common reason a strategy that backtests beautifully on liquid names loses money on the ones a small account can actually afford.

Displayed size is a promise with an asterisk

The size under a quote is the **displayed** size — the part somebody chose to show you. Three things sit behind it. **Hidden** orders are real but unpublished. **Reserve** or iceberg orders publish a small tip and keep the rest out of sight, refilling the tip as it trades. And the displayed tip itself can be **cancelled** in the time it takes you to read it: order-to-trade ratios on U.S. exchanges run in the tens to the low hundreds to one, which tells you that most orders are never executed at all. So displayed size is a weak estimate of available liquidity, and it is biased in a specific direction. In a calm, liquid name it *understates* liquidity, because reserve orders and hidden flow will trade with you that you cannot see. In a stressed name it *overstates* it, because displayed quotes get pulled exactly when they are most needed. That asymmetry is worth internalising: liquidity you can see is most reliable when you need it least. The practical consequence is a two-part cost estimate you can do in your head before any order. If your size is comfortably inside the displayed depth at the touch, your cost is about half the spread — you are taking what is offered. If your size is larger, your cost is half the spread **plus** the price impact of walking the book, and the further you go the worse each additional level gets. Estimating cost before you send it (bid $42.10 × ask $42.13, 1,200 shown at the ask) — Buy 500 shares — inside the displayed size: about half the spread: ≈$1.50 · Buy 1,200 shares — exactly the displayed size: still ≈half the spread: ≈$3.60 · Buy 4,000 shares — through the touch and beyond: half-spread plus impact, and the impact term dominates ← · The same 4,000 shares as a resting limit order: no spread paid, no certainty of a fill ← This is why the size you type matters more than the venue you choose. Below the displayed size the market absorbs you at a predictable price; above it, you are the one setting the price you pay. A displayed quote is not a commitment to trade at that price for that size. It can be cancelled, and it can be cancelled because of you — a quote that disappears the moment your order arrives is the cost of being visible in a thin book.

The rest of the bill

The spread is the largest cost in most retail trades, but it is not the only one, and a trade that ignores the rest looks cheaper than it is. There are four lines, and they scale differently — which is why the honest way to think about them is not a single number but a rule for which one dominates at which position size. **Regulatory and exchange fees** are per-share or per-dollar and tiny: a sale pays the SEC’s Section 31 fee, computed on proceeds at a rate the agency resets each year and which currently works out to tens of dollars per million sold, and a sale also pays the FINRA trading activity fee of a fraction of a cent per share with a small per-trade cap. Clearing and exchange fees are in the same range. The total is a few dollars on a ten-thousand-dollar trade — small in absolute terms, and worth knowing because the *absence* of commission has made these the only fixed costs left, and they are absent from most backtests. **Slippage from size** is the term that grows with how much you trade. Inside the displayed depth you pay about half the spread; beyond it you walk the book, and the marginal cost of each additional level rises because the levels themselves get thinner and further apart. This is the cost that makes a strategy break at scale, and it is invisible in any per-trade estimate computed on a hundred shares. **Opportunity cost** is the one nobody counts and the one that decides the outcome of patient strategies. An order that rests at your price and never fills has a cost equal to the move you missed, and because that cost does not appear on a statement it is systematically underestimated. A limit order that improves your price by five cents fails maybe a quarter of the time in a moving market, and when it fails it usually fails because the stock left without you. Pricing that correctly is what makes the choice between crossing the spread and resting in it an actual decision rather than a preference. • Spread: paid on every round trip, and the largest cost at retail size. • Regulatory and exchange fees: a few dollars per trade, unavoidable, and missing from most backtests. • Slippage: grows with size, dominates once your order is larger than the depth at the touch. • Opportunity cost: the price of the fills you never got, which is the real cost of always resting and waiting. A useful default for a small account: estimate total round-trip cost as the spread in basis points, add a couple of basis points for fees, and add impact only when the order exceeds the displayed size at the touch. That gets you within the right order of magnitude, which is all a pre-trade estimate needs to do.

Benchmarking your own fills

Everything so far has been the cost the market advertises. The number that decides whether a strategy works is the cost your own orders actually pay, and there is exactly one way to learn it: compare each fill with a price you recorded *before* the order left. That reference is the **arrival price** — the midpoint of the quote at the instant you submitted, not the price you hoped for and not the last print on the tape (which can be old, or from the wrong side of the market). For a buy, slippage is your fill minus the arrival midpoint, expressed as a fraction of the midpoint so it is comparable across names. This is the first term of what institutions call implementation shortfall, and for a retail order it is nearly the whole story. It helps to know which benchmark answers which question, because using the wrong one makes good execution look like bad execution. **Arrival** isolates the cost of getting your order done. **A day’s VWAP** asks whether your patient order did better or worse than the volume-weighted average everybody else traded at — useful for comparing a sliced institutional order, meaningless for a single retail fill. **The close** asks whether simply waiting to the end of the day would have been better; it is a hindsight benchmark, and it is the one people reach for when they want to feel bad about a trade that went their way. A retail trader can run the arrival test with a notebook. Log the bid, the ask and the midpoint the moment you click, then log the fill and the size. Fifty trades later, average the signed slippage in basis points. A positive average on buys means you are consistently paying up — which is what market orders into a rising tape do — and a negative average means your resting orders are quietly earning you money. Neither is a moral judgement. Both are facts about your execution that you cannot get from the broker’s disclosure, because the disclosure does not know when you decided. The value of the number is that it sets a floor under everything else. If your fills average 12 bps of cost and you hold a position for three days, the trade has to clear that hurdle before the thesis even starts to matter. If your average is 30 bps, the same idea needs a far larger expected move to be worth expressing, or a longer hold to spread one crossing over more return. A strategy that looks marginal on paper is often clearly unprofitable once the measured number is subtracted — and the measured number is usually worse than the estimate, because the estimate is computed from the calm quotes you happen to look at while the fills happen in the moments you do not. One caution keeps the exercise honest. A single order in a thinly traded name can legitimately cost hundreds of basis points, and a single auction print can be worse than any quote you ever saw. The benchmark is about the **average across many fills**, because that average is the thing a strategy has to survive. Read it monthly, compare it against your estimate, and treat a persistent gap as a question about your order types rather than about the market. • Arrival midpoint: recorded when you submit, and the only reference that measures your execution rather than the market’s. • VWAP and the close: useful for slicing and hindsight, wrong for a single retail fill. • Signed slippage in bps, averaged over dozens of trades: your real cost, and the hurdle every idea must clear. This is the same arrival benchmark the SEC asks brokers to report against in their order-routing disclosures. You can reconstruct it yourself from quotes you already saw; you just have to write them down before the order goes.

Sources

• Market structure and quoted spreadsSEC — Equity Market Structure literature review
• Trading and Exchanges: Market Microstructure for PractitionersLarry Harris, Oxford University Press (2003)
• Tick size, access fees and 2024 amendmentsSEC — Regulation NMS amendments (Rule 612 and access fee caps)