Pre-Trade Analysis: The Complete Guide for Stock Investors

What pre-trade analysis covers

A stock can look attractive at its displayed price yet be costly or difficult to trade in the intended quantity. Pre-trade analysis separates the investment idea from the proposed transaction: what is being traded, how much, under what conditions, and what could change between the decision and the fill? It applies to buys and sells, and to a single stock or a basket.

A useful analysis has three outputs. First, it estimates transaction costs, including the bid-ask spread, potential market impact and broker charges. Second, it describes execution feasibility: whether the quantity is large relative to available trading volume and whether the proposed timing is realistic. Third, it places the trade in portfolio context, including volatility, concentration and exposure to known events. Estimates should be labeled as estimates, especially when live quotes or complete order-book data are unavailable.

The process does not determine whether a company is fairly valued. It also does not tell you what a stock will do next. Its narrower question is whether the trade, as currently specified, deserves a closer look before an order reaches a broker.

Where the practice came from

Pre-trade analysis developed around institutional program-trading desks, which needed to plan orders involving many stocks and quantities large enough to affect execution prices. Transaction cost analysis (TCA) gave those desks a framework for comparing an intended trade with what actually happened. Models associated with Kyle (1985), Glosten and Milgrom (1985), and Almgren and Chriss (2000) help explain, in different ways, why liquidity, information and order size matter to trading costs.

The distinction is timing:

A pre-trade estimate is not a post-trade measurement. The latter needs actual executions and a clearly defined benchmark, such as the price when the decision was made. For a private investor, the same distinction is useful even when an order is small: planning first and reviewing actual fills afterward answer different questions.

Start with the exact trade, not just a ticker

Write down the side (buy or sell), number of shares, order conditions and intended timing for each line. A 50-share purchase and a 50,000-share purchase in the same stock face different liquidity constraints. A market order and a limit order also create different trade-offs: a limit price constrains the price at which the order may execute but does not guarantee a fill.

For a basket, add the intended dollar value of each line. Check whether the figures represent trade amounts or post-trade holdings; those are not interchangeable. If the goal is to assess the portfolio after the trades, include existing positions and cash as appropriate. Specify the currency and the date or time of the market data used. These small definitions prevent misleading comparisons later.

Estimate the cost before placing an order

The visible quote contains a bid—the price at which a buyer is currently willing to buy—and an ask—the price at which a seller is currently willing to sell. The quoted spread is:

Spread per share = ask price − bid price, in US$ per share. Relative spread = (ask − bid) ÷ midpoint × 10,000, in basis points (bps), where midpoint = (ask + bid) ÷ 2 and 1 bp = 0.01%.

For example, a hypothetical US$20.00 bid and US$20.04 ask produce a US$0.04 spread and a US$20.02 midpoint. The relative spread is about 20 bps. A buyer who immediately pays the ask rather than the midpoint crosses half that quoted spread—US$0.02 per share, or about 10 bps—before other costs. A later sale that also crosses the spread could incur another half-spread if quotes were unchanged. Actual costs can differ because quotes, available shares and fills change.

Market impact is the price effect associated with executing an order, especially when its size is meaningful relative to available liquidity. It is distinct from the quoted spread. Larger or more urgent orders can face higher estimated impact; a displayed best price may cover only part of the intended quantity. Treat an impact model as an estimate, not a charge you can read directly from a quote.

Add any applicable commissions and fees in dollars, then express them in a common unit. For an order with approximate notional value N US$:

Estimated one-way cost (bps) = 10,000 × (estimated spread cost + estimated impact cost + commissions and fees) ÷ N.

Use one-way costs for one proposed buy or sell. Do not silently double them into round-trip costs, or count the spread twice by adding both a full quoted spread and a separately calculated half-spread for the same immediate execution. Taxes, currency conversion and other broker-specific charges may also matter; include them when applicable and known.

Check liquidity and participation

Average daily volume (ADV) is the average number of shares traded per day over a stated lookback period. It provides scale, not a promise that those shares will be available at the desired price or time. Compare the proposed quantity with ADV in matching units:

Participation rate = proposed shares ÷ ADV shares × 100%.

A hypothetical 10,000-share order in a stock with 500,000 shares of ADV is 2% of ADV. That does not mean it will take 2% of a day to fill: trading is uneven, current volume may differ from its average, and participation is normally assessed against volume over the period in which execution is attempted. If a stock trades thinly or has a wide spread, consider what happens if only part of the quantity fills. Check for trading halts or suspensions separately; an ADV figure cannot make a suspended stock executable.

Different data sets can produce different ADV estimates. Record the lookback window, whether volume is adjusted consistently for share splits, and whether the relevant exchange session is open. For a cross-market basket, avoid comparing raw share volume alone as if every stock had the same share price and market structure.

Separate market movement from execution cost

Volatility measures variation in returns; it is not itself a transaction fee. Daily volatility can help put a proposed trade’s short-horizon risk in context. An estimate may use close-to-close returns or daily high, low and opening prices. The method and lookback period matter, particularly around sharp moves or sparse data.

Value at risk (VaR) summarizes a modeled loss threshold over a stated horizon and confidence level. Under a simplified normal-return assumption, with portfolio value V US$, daily return volatility σ in decimal units, and no assumed expected daily return:

1-day 95% VaR ≈ 1.645 × σ × V, in US$.

For a hypothetical US$10,000 portfolio with estimated daily volatility of 2% (σ = 0.02), this gives approximately US$329. Interpreted within the model, losses exceeding US$329 would occur on about 5% of days. It is not a maximum possible loss, a guaranteed boundary or an estimate of the size of losses beyond that threshold. The calculation also depends on the portfolio’s combined volatility; adding individual position VaRs generally does not produce the basket’s VaR because holdings can move together or differently.

Check risks a single volatility figure can miss:

• Gap risk: a stock can open far from its previous close, leaving no opportunity to trade at intervening prices.
• Event risk: earnings, company announcements and other scheduled events can alter liquidity and price behavior.
• Model risk: historical estimates may be poor guides when market conditions change, especially during unusual events.

A risk estimate is more useful when its assumptions and data date are visible than when it is presented as one unexplained number.

Measure the basket you would actually hold

A basket can spread capital across several tickers yet remain concentrated in one sector, country, currency or company. Calculate each holding’s post-trade weight using a consistent portfolio value: position market value divided by total portfolio market value. A 40% position has a different effect on overall exposure from a 4% position, even if both trades have similar estimated execution costs.

The Herfindahl-Hirschman index (HHI) offers a compact measure of position concentration:

HHI = Σ wᵢ², where each position weight wᵢ is a decimal and the weights sum to 1 across the included holdings. A three-position basket weighted 50%, 30% and 20% has HHI = 0.50² + 0.30² + 0.20² = 0.38, or 3,800 if weights are entered as whole percentages. For comparison, three equal weights produce about 0.333 (or 3,333 on the percentage scale). State which scale you use.

HHI describes how uneven the weights are; it does not measure correlations, sector overlap or the chance of loss. Two holdings can have different tickers but similar economic exposures. Look at the largest single-position weight alongside sector, country, region and currency breakdowns, rather than relying on one concentration number.

Check data quality and context

Pre-trade calculations inherit the limitations of their inputs. Before trusting a precise-looking output, verify that prices and volumes refer to the right security, currency and session; that a stock has not been suspended; and that corporate actions have been handled consistently. End-of-day data is useful for screening but is not a live executable quote. A model-based spread estimate is not the current bid and ask.

Recent news and analyst ratings can add context, but neither is a substitute for checking the original announcement or understanding what an estimate means. Note publication dates, potential event timing and whether an analyst view predates material new information. If data is missing or stale, record that uncertainty instead of treating the resulting number as equally reliable.

A 10-point pre-trade checklist

Use this list for each proposed line and then repeat the portfolio checks for the combined basket:

1. Trade definition: Have I recorded the stock, buy or sell side, share quantity, intended timing and order conditions?
2. Price and spread: What are the available bid, ask and spread, or—if live quotes are unavailable—what is clearly labeled as an estimate?
3. Full cost: Have I considered spread crossing, market impact, commissions and other applicable charges without double-counting?
4. Liquidity: What is ADV over a stated period, and how does the proposed quantity compare with it?
5. Execution feasibility: Could a partial fill, changed quote, halt or suspension affect the proposed plan?
6. Volatility: What is the estimated daily volatility, and what data and method produced it?
7. Downside and events: What does a stated 1-day VaR assumption show, and what gap or scheduled-event risks might it miss?
8. Concentration: What are post-trade position weights, the largest weight and HHI? Do holdings cluster by sector or country?
9. Data quality: Are identifiers, prices, volumes, currencies and timestamps current and consistent enough for the calculation?
10. Decision record: Which assumptions or conditions would require a fresh analysis before an order is placed?

There is no universal spread, participation or HHI cutoff suitable for every investor and stock. A checklist makes the inputs and trade-offs visible; it does not make the decision automatically.

Worked example: a hypothetical three-stock basket

Suppose an investor is examining three US-listed stocks at illustrative prices. The figures below are hypothetical round numbers, not market quotes. Assume each proposed purchase would become the investor’s only stock holdings, so trade weights and post-trade stock-basket weights are the same. The example excludes cash, commissions and taxes to isolate the mechanics.

Notional is quantity × hypothetical midpoint. Participation divides each quantity by its own stock’s ADV; it is not calculated by adding unrelated stocks’ share volumes. The final column consists of assumed one-way cost estimates, not observed fills. Multiplying each line’s notional by its cost in bps and dividing by 10,000 gives US$15 for A, US$15 for B and US$25 for C. The estimated basket total is US$55, or 27.5 bps of US$20,000, before excluded charges. Stock C is only one-quarter of the basket’s value but accounts for US$25 of the estimated US$55 cost.

The post-trade weights are 50%, 25% and 25%. Their HHI is 0.50² + 0.25² + 0.25² = 0.375, or 3,750 on the whole-percentage scale. Technology represents 75% of this stock basket by value because A and C share that illustrative sector. Those observations say something different from the cost estimate: a low-cost fill would not remove the basket’s sector concentration.

If a separately estimated basket daily volatility were hypothetically 2%, the simplified normal-model 1-day 95% VaR would be 1.645 × 0.02 × US$20,000 = US$658. That volatility cannot be derived from the table’s spreads and ADV; it would require return data and a basket-level treatment of how the holdings move together. The example supplies it only to show units and interpretation, not to certify a risk limit or a verdict.

How PreTrAIde applies the framework

PreTrAIde is an AI pre-trade analysis platform for stocks, covering US stocks and major European exchanges. A user builds a basket of one or more stocks, specifying the side, quantity and execution conditions for each. Before any order is placed, a PreTrade analysis returns a GO, NO-GO or MODIFY verdict for the basket and for each line, with a success probability, risk assessment, suggested execution strategy and plain-language reasoning. These verdicts are decision support, not recommendations.

Its data enrichment draws on end-of-day prices, volumes and company profiles, plus recent ticker news and analyst ratings—about 57 fields per stock. A deterministic quantitative layer estimates spreads using a liquidity-tier model, volatility uplift and one-tick floor; it also calculates participation in daily volume, Almgren-Chriss-style market impact, daily volatility using blended estimators, 1-day 95% VaR, liquidity and data-quality measures, portfolio breakdowns and HHI. It produces 0–100 liquidity, volatility and execution scores for each stock and the basket. AI agents reason over those scores and the news and analyst context.

Guardrails can correct an AI conclusion that conflicts with the data: a suspended stock is always NO-GO, a GO with a success probability below 50% is downgraded, and probabilities are constrained to 5%–95%. Guardrails move verdicts only in a more conservative direction. The success probability is a model estimate, not a guaranteed outcome or a forecast of investment returns.

When a verdict is MODIFY, the optional Let AI Modify feature can rework the basket once for that analysis at no additional analysis credit. It can make up to five AI passes that trim, exit or replace lines, subject to limits including at most two new lines, retention of at least half the original lines, and no single position above 40% of the basket. The user still decides what, if anything, to do.

PreTrAIde is analysis software, not a broker, investment adviser or portfolio manager. It does not receive, transmit or execute orders; users place any trades themselves with their own broker. A paper-trading simulation is available after a verdict on every plan, without real money. Users may also log actual executions and fills so a per-account memory can personalize later AI analysis; the quantitative scores remain unchanged. For a lighter first pass, PreTrAIde’s free tools include calculators for position size, trading cost and VaR, a portfolio concentration checker and a pre-trade checklist. Those tools and a full basket analysis serve different levels of detail, but both begin with the same principle: define the proposed trade before interpreting the numbers.