Crypto Execution Algorithms: A Practical Guide for Institutional Traders

Sweep Algorithms: Aggressive Immediate Execution

A sweep algorithm executes an order as quickly as possible by simultaneously hitting the best available prices across all connected venues. When the SOR identifies liquidity at favorable prices on multiple exchanges, a sweep sends orders to all of them at once, "sweeping" the order book.

Sweep is the most aggressive execution strategy. It prioritizes speed and certainty of execution over price optimization. The full order is exposed to the market immediately, which means there is no opportunity to benefit from favorable price movements during execution. However, there is also no risk of adverse price movement between the decision to trade and the completion of execution.

Sweep algorithms are best suited for several scenarios. Small to medium orders (where the total quantity is within the combined top-of-book depth across venues) benefit from sweep because the price impact is minimal and the execution is instantaneous. Urgent orders where time sensitivity outweighs price sensitivity, such as risk-reducing trades or trades tied to expiring opportunities, are natural candidates. Highly liquid assets like BTC and ETH, where the order book depth across 50+ venues is substantial, can absorb sweep orders up to meaningful sizes without significant slippage.

The primary risk of sweep execution is information leakage. When orders hit multiple venues simultaneously, the market can infer the presence of a large buyer or seller. Sophisticated participants may adjust their quoting accordingly. For very large orders relative to available liquidity, sweep can cause temporary price dislocations that work against the trader. In these cases, a time-slicing algorithm like TWAP is generally more appropriate.

TWAP and VWAP: Time-Based Execution Strategies

Time-Weighted Average Price (TWAP) and Volume-Weighted Average Price (VWAP) are the workhorses of institutional execution. Both algorithms break a large order into smaller slices and execute them over a defined time window, but they differ in how they pace execution.

TWAP divides the order into equal-sized slices and executes one slice at each regular interval (for example, one slice every 30 seconds over a 30-minute window). The goal is to achieve an average execution price close to the arithmetic average of prices observed over the execution window. TWAP is agnostic to volume patterns; it executes the same quantity regardless of whether the market is active or quiet.

VWAP, by contrast, weights execution toward periods of higher market activity. It aims to achieve an average price close to the volume-weighted average price. In equity markets, VWAP algorithms use historical volume profiles to predict when volume will be highest (typically near market open and close) and concentrate execution during those periods. In crypto, 24/7 trading means there are no opening or closing auctions, but volume still follows patterns (activity tends to peak during U.S. and European business hours and around major economic data releases). A crypto VWAP algorithm must learn and adapt to these patterns.

For most institutional crypto use cases, TWAP is the more common choice. The 24/7 nature of crypto markets makes volume prediction less reliable than in equities, and TWAP's simplicity makes it easier to reason about expected execution quality. TWAP is ideal for large orders (exceeding 5-10% of average daily volume) where minimizing market impact is the primary objective, and where the trader does not have strong views on short-term price direction.

Both algorithms integrate with SOR to determine venue selection for each individual slice. Each time a slice is ready to execute, the SOR evaluates current liquidity conditions and routes the slice to the optimal venue or combination of venues.

Chase and Adaptive Algorithms: Responding to Market Movement

While TWAP and VWAP execute on a predetermined schedule, chase and adaptive algorithms dynamically adjust their behavior based on real-time market conditions. This makes them particularly valuable in crypto, where price movements can be sudden and substantial.

A chase algorithm monitors the market price relative to a reference point (typically the price at the start of execution) and adjusts its aggression accordingly. When the market moves in a favorable direction (prices dropping for a buy order, rising for a sell order), the algorithm accelerates execution to capture the favorable move. When the market moves unfavorably, the algorithm slows down or pauses, waiting for conditions to improve.

The rationale behind chase is rooted in short-term momentum. Crypto markets exhibit pronounced momentum effects, where price moves in one direction tend to persist for brief periods before reverting. A chase algorithm exploits this by executing more aggressively during favorable momentum and conserving capacity during adverse moves.

The risk of chase algorithms is that they can become trapped in an unfavorable position. If the market moves adversely and the algorithm repeatedly pauses, the order may remain largely unexecuted, and the eventual completion price may be worse than if a simple TWAP had been used. To mitigate this, well-designed chase algorithms include "catch-up" logic that increases aggression if the order falls too far behind schedule.

More broadly, adaptive algorithms encompass any strategy that modifies its behavior based on observed market conditions. Some adaptive algorithms adjust execution speed based on realized volatility (slowing down during quiet periods and speeding up during volatile ones to avoid being adversely selected). Others monitor order book imbalance signals and adjust their passive/aggressive mix accordingly.

These algorithms require more sophisticated parameter tuning than TWAP. Traders must set thresholds for when the algorithm should accelerate or decelerate, maximum and minimum execution rates, and the catch-up parameters that prevent the order from falling too far behind schedule.

Market Making and Spreader Algorithms

Market making and spreader algorithms differ from the execution strategies described above in that they are designed to both buy and sell simultaneously, earning the bid-ask spread while gradually building or reducing a net position.

A market making algorithm continuously places limit orders on both sides of the order book across multiple venues, quoting a bid price and an ask price with a spread between them. When both sides fill, the algorithm earns the spread as profit. The algorithm dynamically adjusts its quotes based on inventory position, volatility, and order flow signals. When inventory becomes unbalanced (for example, if the algorithm has accumulated a long position), it skews its quotes to encourage trades that rebalance the position.

For institutional firms, market making algorithms serve two primary purposes. First, they provide a way to accumulate or distribute large positions passively, earning the spread along the way rather than paying it. A firm looking to build a 1,000 BTC position can run a market making algorithm that gradually accumulates inventory while generating spread revenue. Second, they provide a hedging mechanism for OTC flow. When a firm executes a large OTC trade, it can use a market making algorithm to gradually hedge the resulting position on exchanges.

Spreader algorithms are a specialized variant that focus on cross-venue spread capture. A spreader posts bids on one exchange and offers on another, profiting from persistent price discrepancies between venues. This is closely related to arbitrage but is structured as a continuous, automated strategy rather than an opportunistic one.

Both market making and spreader algorithms require careful risk management. Inventory limits prevent the algorithm from accumulating excessive directional exposure. Kill switches halt the algorithm if losses exceed predefined thresholds. Latency monitoring ensures that stale quotes are cancelled before they can be adversely selected.

Selecting the Right Algorithm: A Decision Framework

Choosing the optimal execution algorithm requires evaluating several trade-specific and market-specific factors.

Order size relative to available liquidity is the most important factor. For orders that are small relative to the combined order book depth across connected venues (roughly less than 1% of the total displayed depth), a sweep is typically the best choice. The execution is instantaneous, the price impact is negligible, and there is no timing risk. For larger orders, time-slicing strategies (TWAP, VWAP, Chase) become necessary to manage market impact.

Urgency determines how aggressively the algorithm should execute. A risk-reducing trade or a trade tied to an expiring opportunity warrants more aggressive execution (sweep or short-duration TWAP). A position-building trade with no deadline can be spread over hours or days using TWAP or market making strategies.

Volatility conditions affect algorithm selection in two ways. During low-volatility periods, passive strategies (limit orders, market making) are more effective because adverse selection risk is lower. During high-volatility events, more aggressive strategies may be preferred to avoid the risk of prices moving substantially during a slow execution.

Information leakage sensitivity matters for orders where other market participants' awareness of the trade could move prices. If the order is based on proprietary research or represents a known large player's activity, minimizing information leakage is critical. In these cases, dark pool routing, smaller slice sizes, and randomized timing help obscure the order's presence.

A practical decision framework follows this logic: if the order is small and urgent, use sweep. If the order is large and the timeline is flexible, use TWAP with SOR. If the market is trending favorably, consider chase to capture the move. If the order is very large and the firm can be patient, market making strategies that earn the spread while building position offer the best total cost of execution.

• Small + urgent = Sweep across all venues
• Large + flexible timeline = TWAP with SOR optimization
• Large + favorable trend = Chase algorithm with catch-up logic
• Very large + patient = Market making to accumulate while earning spread
• High information sensitivity = Smaller slices, randomized timing, dark pool routing
• High volatility = More aggressive execution to reduce timing risk