When traders talk about costs, the conversation almost always revolves around commissions and spreads. Systematic traders may go a step further and model slippage. But one cost is routinely underestimated, simplified, or ignored entirely: swap fees, also known as rollover or overnight financing.
That oversight is not trivial. For many systematic strategies, swap fees represent a meaningful portion of total trading costs. In some cases, they rival commissions. In others, they are the primary reason a strategy that looks profitable in backtests slowly bleeds capital in live trading.
Understanding swap mechanics is therefore not optional for anyone serious about systematic trading or realistic backtesting.
What Swap Really Is (and What It Isn’t)
A swap fee is applied to a position that remains open past the broker’s daily rollover time. Conceptually, it represents the cost of financing that position overnight.
In spot FX trading, every position implicitly involves borrowing one currency and lending another. The swap reflects the interest rate differential between those two currencies, adjusted by the broker’s own financing markup. Even for CFDs on indices, metals, or crypto, the same principle applies: you are paying (or occasionally receiving) financing for holding exposure over time.
If a position is opened and closed within the same trading day before rollover, no swap is applied. The moment it survives rollover, swap becomes part of the PnL.
In simple terms:
- You are borrowing one currency
- And lending another
- The swap reflects the interest rate differential, plus broker markup
- If you hold a position overnight, you pay (or receive) interest.
Why Long and Short Trades Have Different Swaps
A critical detail that many traders overlook is that swap is direction-dependent. Every instrument has two values: one for long positions and one for short positions. These values are almost never symmetrical.
This matters enormously for systematic strategies. A strategy with a directional bias—long-only, short-only, or skewed exposure—may be consistently paying the more expensive side of the financing equation. Over dozens or hundreds of trades, that directional financing cost compounds into a significant drag on performance.
From a modeling perspective, swap must be treated as a function of both instrument and position direction. Ignoring this asymmetry produces systematically optimistic backtests.
The Wednesday Rollover Effect (Why Triple Swap Exists)
One of the most confusing aspects of swap is the so-called “triple swap,” usually applied on Wednesday. At first glance, it feels arbitrary or even unfair. In reality, it is a direct consequence of how spot markets settle.
Spot FX trades settle on a T+2 basis—two business days after execution. Markets do not settle on weekends. If a position is held over Wednesday night, the theoretical settlement date would fall on Saturday, which is not a valid settlement day. To compensate, the broker rolls the position forward to Monday instead.
That single rollover therefore accounts for three calendar days of financing: Friday, Saturday, and Sunday. Rather than charging swap on each of those days individually, the broker applies them all at once. The result is triple swap.
This mechanism is structural, not broker-specific. While FX pairs typically apply triple swap on Wednesday, other instruments—such as indices or CFDs—may use different rollover conventions depending on the broker. The correct rollover day must always be checked at the instrument level.
Important notes:
- FX: usually Wednesday
- Metals: often Wednesday
- Indices / CFDs: broker-defined (can differ)
Swap Is Charged Per Lot, Per Night — and It Adds Up Fast
Swap fees are charged per lot, per night, for every rollover event a position survives. This is where intuition often fails traders.
A single trade held for several days may incur swap multiple times, and if it passes through the triple-swap day, the cost increases sharply. Multiply this by trade size and by the number of trades in a strategy, and what once looked like a “small overnight cost” becomes a recurring structural expense.
Example:
- 1.0 lot
- −5.00 swap per night
- Held 4 nights (including Wednesday)
- Swap cost = −5 × (1 + 1 + 3 + 1) = −30 – (Do this many times in a month — suddenly swap can rival commission.)
This is why traders are often surprised to discover that, over a few weeks of active trading, swap fees can approach—or even exceed—total commissions.
Swap Calculation Modes and Why Backtests Often Get Them Wrong
Another layer of complexity is that brokers calculate swaps using different modes. Depending on the instrument, swap may be quoted in points, in account currency, as an interest rate, or relative to margin used.
For systematic traders, this distinction is critical. A swap quoted in points must be converted into monetary terms using tick value and position size. A swap quoted directly in currency does not. Confusing these modes leads to incorrect cost modeling, often underestimating financing expenses by a wide margin.
This is one of the most common reasons backtests diverge from live results. The trading logic is sound, but the cost model is incomplete.
Why Swap Costs Differ Between Brokers
Swap rates are not universal. Two brokers offering the same instrument can—and often do—apply very different swap values.
These differences arise from broker-specific factors such as liquidity providers, internal risk management, and financing markups. As a result, a strategy may perform acceptably at one broker while slowly deteriorating at another, even if spreads and commissions appear similar.
For systematic traders, this means that broker choice is part of the strategy. Swap must be treated as a broker-dependent parameter, not a market constant.
Swap as a Strategy Killer (or Strategy Driver)
Certain types of strategies are particularly sensitive to swap costs. These include strategies with multi-day holding periods, small average profit per trade, or exposure to high-financing instruments such as indices or crypto CFDs.
A common failure mode looks like this: a strategy shows solid expectancy in backtests, performs reasonably well in forward testing, but gradually underperforms in live trading. The discrepancy is often blamed on slippage or market regime changes. In reality, swap costs—ignored or simplified during development—are quietly eroding the edge.
In some cases, swap is not just a cost but a defining characteristic of the strategy. A system may unintentionally become a negative carry trade, paying financing on nearly every position.
Modeling Swap Correctly in Systematic Backtests
For realistic backtesting, swap must be modeled with the same seriousness as commission.
That means accounting for:
- Instrument-specific swap values
- Long vs short direction
- Rollover frequency
- Triple-swap days
- Correct unit conversion
Best practice is to apply swap daily, while the trade is open, rather than deducting it only at trade close. This allows for accurate drawdown modeling, realistic equity curves, and the possibility of swap-aware exit logic.
Once modeled correctly, many strategies change character. Some lose their edge entirely. Others improve when holding time is reduced or positions are closed before rollover.
Managing Swap in Practice
There are several ways systematic traders can reduce the impact of swap. Strategies can be designed to avoid holding positions through rollover, or specifically avoid the triple-swap day. Instruments with more favorable financing can be preferred, and holding times can often be reduced without materially harming signal quality.
At the broker level, swap tables should be compared just as carefully as commissions. In some cases, alternative instruments or account structures can materially improve financing costs.
The key point is that swap must be addressed deliberately. Ignoring it is not a neutral decision—it is an implicit assumption that financing does not matter.
Final Thoughts
Swap fees are not a rounding error. They are not noise. They are a structural component of trading performance.
For systematic traders and backtesters, understanding and modeling swap is part of doing the job properly. Many strategies fail not because the signal is weak, but because the cost structure was misunderstood. If you do not know how much your strategy pays in swap, you do not fully understand your strategy.
And in systematic trading, incomplete understanding is usually expensive.