Partial Fills and Leg Risk in Crypto Arbitrage

Partial Fills and Leg Risk in Crypto Arbitrage

Understand partial fills and leg risk when one side of an arbitrage completes before the other, including measurement, prevention and predefined exit plans.

This topic should not be treated as a shortcut to guaranteed profit or as an automated trading instruction. Crypto prices, order books, network status, exchange rules and fees can change quickly. A sound approach treats an observed difference as the start of research, verifies current conditions on official exchange interfaces and includes a downside scenario.

Definition and scope

A partial fill occurs when only part of a submitted order executes. Leg risk is the open market exposure created when one step of a multi-leg arbitrage completes but the other is delayed, partially filled or not filled at all.

In practice, no single indicator is sufficient. The same signal can produce a different outcome when order size, account tier, regional restrictions, network choice or data age changes. The analysis must therefore cover executable conditions rather than only a theoretical percentage.

Why does this matter?

An arbitrage may look direction-neutral in theory without remaining neutral during execution. If a buy fills on one venue and the sale does not, the user becomes exposed to the coin’s price. With a narrow expected margin, even a short unhedged interval can exceed the projected spread.

A large displayed spread does not prove that both sides of a transaction can be completed. Skipping one control layer may create a partial fill, an unexpected cost, a transfer block or an unhedged market position. A systematic review is useful mainly because it filters false positives before capital is exposed.

Key factors to evaluate

Filled quantity

Track the actual filled coin amount on both sides; risk should be based on fills rather than submitted order size.

Evaluate this factor for the intended transaction size. Conditions that look acceptable for a small order can change rapidly at a larger size.

Verification question: Is this factor supported by current official data rather than a stale snapshot?

Order-book depth

Cumulative quantity behind the best price must be sufficient for the intended size on both venues.

Do not limit the check to a scanner screen. Reconfirm the official exchange data, timestamp and account restrictions immediately before a decision.

Verification question: Is this factor supported by current official data rather than a stale snapshot?

Order timing

Submission sequence and latency determine which side may create temporary exposure.

Even when this indicator looks favourable, read it together with cost and risk layers. The goal is not to chase the largest number but to make assumptions visible.

Verification question: Is this factor supported by current official data rather than a stale snapshot?

Cancel and reprice logic

Define when the remaining order will be cancelled, repriced or closed with a marketable order.

Evaluate this factor for the intended transaction size. Conditions that look acceptable for a small order can change rapidly at a larger size.

Verification question: Is this factor supported by current official data rather than a stale snapshot?

Maximum open exposure

Set limits for unmatched quantity, acceptable loss, elapsed time and adverse price movement.

Do not limit the check to a scanner screen. Reconfirm the official exchange data, timestamp and account restrictions immediately before a decision.

Verification question: Is this factor supported by current official data rather than a stale snapshot?

Logs and monitoring

Record fill events, timestamps, errors and weighted execution prices for later review.

Even when this indicator looks favourable, read it together with cost and risk layers. The goal is not to chase the largest number but to make assumptions visible.

Verification question: Is this factor supported by current official data rather than a stale snapshot?

Step-by-step verification workflow

The following workflow helps review similar signals with consistent criteria. The sequence may be compressed when conditions move quickly, but critical checks should not be removed.

1. Define the route and intended size

Specify the coin, trading pair, buy venue, sell venue and intended amount. Confirm that the asset identity and account conditions match the route.

Record the data source and timestamp at this stage. If a small change in assumptions turns the result negative, consider a wider safety margin or a smaller transaction size.

2. Verify data sources and timestamps

Compare the scanner update with official exchange data. Do not use the displayed percentage as a decision input when the source is delayed or incomplete.

Record the data source and timestamp at this stage. If a small change in assumptions turns the result negative, consider a wider safety margin or a smaller transaction size.

3. Review the first two critical factors together

Assess Filled quantity and Order-book depth for the same timestamp and size. A strong reading in one and a weak reading in the other may make the gross difference misleading.

Record the data source and timestamp at this stage. If a small change in assumptions turns the result negative, consider a wider safety margin or a smaller transaction size.

4. Add costs and execution effects to the model

Combine Order timing with trading fees, withdrawal cost, slippage and, where relevant, conversion or rebalancing effects in one calculation.

Record the data source and timestamp at this stage. If a small change in assumptions turns the result negative, consider a wider safety margin or a smaller transaction size.

5. Run an adverse-scenario stress test

Use worse assumptions for Cancel and reprice logic and Maximum open exposure. Test whether the estimate remains acceptable if price moves adversely, liquidity declines or execution is delayed.

Record the data source and timestamp at this stage. If a small change in assumptions turns the result negative, consider a wider safety margin or a smaller transaction size.

6. Complete a final check on official exchange interfaces

Reconfirm Logs and monitoring, network status, order book, account limits, maintenance notices and fees on official exchange interfaces. Where data conflicts, rely on the official venue.

Record the data source and timestamp at this stage. If a small change in assumptions turns the result negative, consider a wider safety margin or a smaller transaction size.

7. Record the outcome and update thresholds

Record execution prices, elapsed time, fees, partial fills and the net outcome. Improve future thresholds with observed results rather than only theoretical assumptions.

Record the data source and timestamp at this stage. If a small change in assumptions turns the result negative, consider a wider safety margin or a smaller transaction size.

Worked example

The figures below are hypothetical and are used only to explain the method. Actual exchange fees, limits and market conditions may differ.

Assume buy and sell orders for 1,000 coins are submitted on exchanges A and B. The buy fills fully, while only 650 coins sell on B.

The user is left net long 350 coins. A 1% price decline on that residual amount can erase much of the intended spread. The exit plan should define whether to wait, reprice the remaining limit order or close exposure with a controlled marketable order.

Estimated net difference = gross price difference − trading fees − transfer/network cost − slippage − conversion and rebalancing cost − safety buffer

The formula does not guarantee an outcome; it shows which cost layers belong in the same model. Fixed charges should be divided by transaction value, while percentage fees should be applied to executable prices.

Main risks

The central mistake is assuming that current conditions will remain unchanged until completion. The following risks can reinforce one another and turn an initially positive estimate negative.

• Unexpected change in Filled quantity: Track the actual filled coin amount on both sides; risk should be based on fills rather than submitted order size. This risk may be reduced through smaller test sizes, fresh data, a defined cancellation plan and final verification, but it cannot be eliminated.

• Unexpected change in Order-book depth: Cumulative quantity behind the best price must be sufficient for the intended size on both venues. This risk may be reduced through smaller test sizes, fresh data, a defined cancellation plan and final verification, but it cannot be eliminated.

• Unexpected change in Order timing: Submission sequence and latency determine which side may create temporary exposure. This risk may be reduced through smaller test sizes, fresh data, a defined cancellation plan and final verification, but it cannot be eliminated.

• Unexpected change in Cancel and reprice logic: Define when the remaining order will be cancelled, repriced or closed with a marketable order. This risk may be reduced through smaller test sizes, fresh data, a defined cancellation plan and final verification, but it cannot be eliminated.

• Unexpected change in Maximum open exposure: Set limits for unmatched quantity, acceptable loss, elapsed time and adverse price movement. This risk may be reduced through smaller test sizes, fresh data, a defined cancellation plan and final verification, but it cannot be eliminated.

Common mistakes

The following mistakes widen the gap between a theoretical spread and an actual outcome:

• Ignoring Filled quantity: Track the actual filled coin amount on both sides; risk should be based on fills rather than submitted order size.

• Ignoring Order-book depth: Cumulative quantity behind the best price must be sufficient for the intended size on both venues.

• Ignoring Order timing: Submission sequence and latency determine which side may create temporary exposure.

• Ignoring Cancel and reprice logic: Define when the remaining order will be cancelled, repriced or closed with a marketable order.

• Ignoring Maximum open exposure: Set limits for unmatched quantity, acceptable loss, elapsed time and adverse price movement.

• Ignoring Logs and monitoring: Record fill events, timestamps, errors and weighted execution prices for later review.

• Using the last-traded price instead of the executable buy ask and sell bid.

• Treating one successful example as evidence of permanent performance.

How Exarbi supports this analysis

Exarbi is designed to present supported exchange price differences together with decision-support signals such as data status, risk level, transfer readiness and fee impact. This helps users narrow the routes worth researching instead of treating a raw price difference as a decision by itself.

Information shown in the dashboard is not an automated trading instruction, personalised investment advice or a profit guarantee. Exarbi does not trade for users, hold funds or request exchange API keys. Final verification and execution remain with the user.

Pre-transaction checklist

Before acting on a route, make sure every question below has a clear answer:

• Has Filled quantity been verified with current official data?

• Has Order-book depth been verified with current official data?

• Has Order timing been verified with current official data?

• Has Cancel and reprice logic been verified with current official data?

• Has Maximum open exposure been verified with current official data?

• Has Logs and monitoring been verified with current official data?

• Are the executable ask for buying and bid for selling being used?

• Have weighted average prices been calculated for the intended size?

• Do the coin, contract and network match on both venues?

• Are deposits and withdrawals currently available?

• Are all costs and an adverse-scenario buffer included?

• Is there an exit plan for a partial fill or delay?

• Does the content avoid profit guarantees and personalised calls to trade?

Frequently asked questions

What is partial fills and leg risk in crypto arbitrage?

A partial fill occurs when only part of a submitted order executes. Leg risk is the open market exposure created when one step of a multi-leg arbitrage completes but the other is delayed, partially filled or not filled at all.

Is partial fills and leg risk in crypto arbitrage sufficient on its own for a trading decision?

No. Price, liquidity, fees, data freshness, transfer status and account restrictions must be assessed together.

Can this analysis be fully automated?

Data collection and initial filtering can be automated, but exchange status, account limits and the final order book should still be verified before execution.

How often should the checks be refreshed?

Refresh them when the signal first appears, immediately before placing orders and, where transfers are involved, again before initiating a withdrawal.

How can Exarbi be used for this topic?

Exarbi helps users research price differences and related risk signals in a readable dashboard; it does not execute transactions or decide for the user.

Conclusion

Partial fills and leg risk can turn an apparently two-sided arbitrage into directional market exposure. The maximum unmatched quantity, permitted time and exit rule should be defined before orders are submitted.

You can explore how Exarbi presents market data, price differences, transfer conditions and risk indicators. Exarbi does not recommend or execute transactions.

Risk and responsibility notice

This content is for general education and information only. It is not investment advice, a personal recommendation or an invitation to trade. Cryptoassets are highly volatile and involve a risk of capital loss. Examples are hypothetical. Independently verify official exchange conditions, fees, network status and your legal or tax obligations before making any decision.

======================================================================