Every execution algorithm is optimizing between cost, alpha, and risk. This part builds that framework. It is the most conceptual chapter of the series and the one everything later refers back to.

Explicit Costs

Explicit costs are the ones that appear on an invoice: broker commissions, exchange fees, and transaction taxes. Commission rates typically reflect how expensive a client is to serve and how many bundled services (research, prime brokerage) they consume.

In crypto, explicit costs take the form of tiered maker/taker fees, withdrawal fees, and - on-chain - network gas. Because fee tiers step down with volume, the same strategy can have materially different explicit costs for different accounts.

Implicit Costs

Implicit costs never appear on an invoice; they show up as worse fill prices. They usually dwarf explicit costs for institutional-size orders, and they come in five flavors.

Spread Cost

Crossing the spread with a marketable order means paying a premium over the midpoint - half the spread each way, the full spread for a round-trip. But the quoted spread overstates what careful traders actually pay, because marketable orders often receive price improvement: executing against hidden midpoint liquidity, for example, or via wholesalers who improve retail flow to win it.

The standard measure that accounts for this is the effective spread: twice the signed difference between your execution price and the prevailing midpoint. If you achieve no price improvement it equals the quoted spread; every basis point of improvement narrows it.

Market Impact

Market impact is the cost of your own trading moving the price against you. It has a mechanical component - consuming the best-priced liquidity means the next fill happens at a worse level - and an informational component: other participants reprice when they detect one-sided flow.

Why should prices move at all if your trades carry no news about fundamental value? Because a market price does not just represent the fundamental value of an asset - it embeds the premium that liquidity suppliers demand for taking the other side. Professional market makers charge for inventory risk; natural counterparties (investors who happened to want the opposite trade anyway) charge little or nothing. Your realized impact depends on the mix available to you. That is also why patience helps: trading slowly recruits more natural liquidity and gives market makers time to recycle inventory, both of which lower the premium you pay. We quantify these effects in Part 13.

Information Leakage

A large unexecuted order is valuable private information - its future impact is predictable price drift that others can front-run. Leakage is any behavior that gives it away: displaying size that scares the far side into repricing, trading with recognizable aggression, or getting detected by pinging - tiny probe orders sent into dark venues to sense whether a whale is resting there. Serious execution doesn’t just hide orders; it avoids looking like a large trader at all. Camouflage tactics run throughout Part 7 and Part 11.

Opportunity Cost

The cost of not trading. Suppose a trader is given discretion on a 10,000-share buy, expects a dip, and waits - then the stock rips higher and the order is cancelled. No commission was paid, no spread crossed, no impact created, yet the trader is worse off by the entire missed move. Unfilled quantity is a real cost, and any honest performance framework charges for it (see Part 12).

Adverse Selection

Adverse selection is the disadvantage of being the one who posts a firm price. Whoever quotes first gets picked off selectively: counterparties trade only when it profits them, which is precisely when it hurts the quoter. A dealer whose bids are hit mainly by better-informed sellers loses steadily and must recover those losses from uninformed, liquidity-motivated flow - which is why market makers prize retail order flow and invest heavily in the low-latency infrastructure needed to update quotes before being run over. The same dynamic drives the latency-arbitrage debate: faster traders adversely select slower quoters whose prices are momentarily stale. If you place passive limit orders, adverse selection is a cost you carry too.

Alpha

Asset managers use “alpha” to mean benchmark-relative outperformance. Traders use it more loosely: trading alpha is the expected price drift over the execution horizon itself. The distinction matters. The portfolio manager’s investment thesis may play out over months; the execution algorithm only cares whether the price will drift for or against the order during the next few hours.

Sometimes the two coincide - in short-horizon stat-arb, the trading window is the investment window. But generally they are separate signals. Execution-horizon alpha can come from short-term momentum or reversion factors, or from liquidity events (stepping in after someone else’s aggressive flow pushed price away from equilibrium). Whether a human or a machine is trading, an estimate of near-term drift should shape how urgent the execution is.

Risk

Price Risk

The dominant execution risk is simply the volatility of the asset over the trading horizon. Volatility differs enormously across assets - a small-cap or a long-tail altcoin versus a major FX pair - and trading models range from a simple market-plus-idiosyncratic split to granular factor models (size, style, sector, and residual). Whatever the model, baseline volatility is the input that most changes the right strategy.

Duration

Risk scales with time in the market. The slower you trade, the longer the market can drift away from your decision price. Pace is therefore a risk dial: speeding up buys certainty and pays impact; slowing down saves impact and accepts uncertainty.

Portfolio Effects and Hedging

For basket trades, what matters is aggregate risk, not the sum of the parts. Correlated buys and sells hedge each other; diversified baskets are calmer than concentrated ones. Traders can also hedge actively: a manager who receives a large cash inflow into a small-cap strategy can buy index futures immediately - neutralizing market beta within minutes - and then unwind the hedge gradually as the underlying positions are built over days. Basket-aware execution is the subject of Part 8.

Risk Aversion

Finally, risk aversion converts objective volatility into a subjective penalty. It doesn’t change the asset’s volatility - it changes how much cost a given desk will pay to avoid it. Two desks handed the identical order can rationally trade it very differently: the risk-averse one fast and expensive, the risk-tolerant one slow and cheap.

The Central Trade-Off

Put the pieces together and execution strategy reduces to one persistent tension: trading fast costs impact; trading slow costs risk (and forgone alpha, if the price is drifting against you). Every algorithm in this series is an answer to that tension - from rigid schedules (TWAP, VWAP) to the explicit cost-risk optimization of implementation shortfall.