What people get wrong about spread assumptions
Spread assumptions are simplifications used to estimate the transaction cost of trading with bid and ask prices. A common mistake is treating the spread as a constant number, even though it can vary with market conditions and execution details. Another mistake is mixing up where the spread appears in calculations (for example, confusing a “quoted spread” with the “effective spread” you actually pay after all trading frictions).
People also often omit important assumptions. If a calculation does not state whether it assumes a fixed spread, an average historical spread, or an all-in cost model, the result becomes hard to interpret and easy to overestimate. Finally, failures are sometimes blamed on “the strategy,” when the root cause was a mismatch between assumed costs and real costs.
The mechanism: what a spread assumption is (and what it isn’t)
In liquid markets, you typically see two prices: a bid (what you can sell for) and an ask (what you can buy for). The spread is the difference between them. A spread assumption is the choice you make for that difference when estimating costs in examples, models, or backtests.
To use spread assumptions correctly, you need to separate stable mechanics from variable factors:
- Stable mechanics: the bid/ask structure and the idea that crossing from one side to the other incurs spread cost.
- Variable factors: whether the spread you assume matches the spread you would face at the time you place orders, and whether other costs also apply.
A clear spread assumption should state the measurement basis (quoted vs effective) and the scope (one trade, many trades, or an average over a period). If your calculation is for multiple entries and exits, assuming one spread value for everything is another frequent misunderstanding.
Evidence and examples of common failure modes
Consider a simple long scenario where you “enter” near an ask price and “exit” near a bid price. If your spread assumption is, say, a tight value that rarely holds when you actually trade, the model understates cost. The consequence is that break-even thresholds shift: the estimated profits can turn into losses once the real effective spread is larger.
A second example is backtesting with historical bid/ask data treated as if it were always tradable at the assumed moments. Even when historical quotes look consistent, execution can differ. If your assumptions ignore slippage or order execution quality, your results may be systematically optimistic.
A third failure mode is mixing spread with other costs. Commissions, financing, or conversion costs (if applicable) may be separate from spread. If you include only spread but ignore other fees, the estimated transaction cost is incomplete. If you instead treat all fees as “spread,” you may still misattribute the cause of differences between estimates and outcomes.
Material limitations and risks in spread assumptions
Spread assumptions are useful for education and rough estimation, but they are limited. Here are common limitations to watch for:
- Spread can change: timing and liquidity can make the real spread larger than the assumed value.
- Assumptions can be inconsistent: quoted spread may not match effective cost after execution.
- Execution introduces uncertainty: order size, speed, and how orders are filled can change costs.
- Historical relationships don’t guarantee future results: past spreads or correlations do not ensure the same behavior later.
A practical verification approach is to explicitly list every cost component your calculation uses and then test whether the assumed spread corresponds to the same definition used in the data or example you are using. If the definitions differ, results may not be comparable.
How to verify your spread assumptions independently
Start with a neutral checklist:
- Define the spread in your calculation: quoted bid-ask difference or effective realized cost.
- State the assumption scope: single trade vs multiple trades vs averaging.
- Include other frictions that your model ignores today (execution quality, commissions, and any additional fees or financing relevant to the scenario).
- Check sensitivity: rerun the estimate using a range of plausible spread values and see how conclusions change.