Why Paper Trades Should Include Costs and Slippage

Paper trades should include costs and slippage because real trades always pay them, and leaving them out makes a strategy look more profitable than it can be. Slippage is the gap between the price you expected and the price you actually got; costs include spreads, fees and commissions. Each is small on one trade, but across hundreds of trades they can decide whether a strategy makes or loses money.

This article breaks down where these costs come from, shows with simple arithmetic how quickly they add up, and explains how a paper account can model them sensibly.

What slippage actually is

When you decide to buy, you look at a price. By the time your order reaches the market and finds a seller, the price you pay may be different. That difference is slippage.

Slippage happens for a few ordinary reasons:

• The spread. A market order to buy fills at the ask, which sits above the last trade. A market order to sell fills at the bid, which sits below it.
• Time. Prices move between the moment a signal is generated and the moment the order executes. For a strategy that decides after the close and trades at the next open, that gap is overnight.
• Gaps at the open. News released overnight can make a stock or fund open well above or below where it closed.
• Size. A large order in a thinly traded stock can use up all the shares on offer at one price and move on to the next, worse price.
• Fast markets. During sharp sell-offs, quotes change quickly and fills can land far from the last price seen.

Slippage can be positive as well as negative. Sometimes the open is better than you expected. Over many trades, though, the spread alone means the average is a small cost, not a gain.

The costs that sit on top of slippage

Not every trader pays every cost. A US investor buying large index funds with a commission-free broker mostly pays the spread and the fund's expense ratio. A gold trader using leverage on MetaTrader 5 may pay a spread on every trade and a financing charge on positions held overnight.

How small costs become large ones

The arithmetic is simple, and it is the reason this topic matters. These are illustrative numbers, not data about any particular strategy.

Suppose every fill costs 0.05% once spread and slippage are combined. A round trip (buy then sell) is two fills, so about 0.1%.

These are rough because compounding and partial position changes alter the exact figures, but the shape is right. A strategy that trades a few times a year barely notices costs. A strategy that trades every day can have its entire edge eaten by them.

That is why a paper account showing no costs is especially misleading for active strategies. It does not just round the result up a little; it can turn a losing strategy into an apparently winning one.

A real example of cost sensitivity

AlgolaneBots publishes a cost warning for one of its strategies for exactly this reason. Gold Breakout checks gold every hour and trades when a quiet period ends in a breakout. In its historical backtest from June 2004 to November 2026 at 2% risk per trade, a computer test on past prices and not money anyone earned, it returned +15.1% a year with a worst fall of 41%. Simply holding gold over the same period returned +11.0% a year with a worst fall of 45%. The strategy's deepest fall ran from May 2018 to October 2019 and took 41 months after the bottom to recover.

Its caveat is the point of this section: doubling trading costs cuts its yearly result roughly in half. In other words, the difference between a modest cost assumption and a pessimistic one is the difference between beating gold held and trailing it. Future results will differ from the backtest, and costs are one of the biggest reasons they might.

If you paper trade a strategy like this on an account with zero costs, the paper result will look better than both the backtest and any live result. You can read more about the strategy on the strategies overview, and about trading gold this way in how to trade gold with a bot.

How a paper account can model costs honestly

There is no perfect way to simulate every fill, but some approaches are clearly better than others.

Fill at a realistic time

The most important choice is when an order fills. Filling at the next real market open, rather than at the close that generated the signal, captures the overnight gap. This matters a lot for strategies that make their decision after the close.

Charge a fixed percentage per fill

A flat charge on every fill is a simple, transparent way to approximate spread and slippage together. It will be too high on some trades and too low on others, but it does not systematically flatter results the way zero does.

Adjust for corporate actions

A stock split changes the share count and price without changing the value of the holding. A simulator that misses this can record a fake crash or a fake windfall, both of which distort the cost picture too.

Be conservative where in doubt

If you are unsure whether costs will be 0.03% or 0.08% per fill, testing with the higher number is the safer habit. A strategy that only works with optimistic costs is fragile.

For reference, AlgolaneBots's paper accounts fill at the next real market open, take 0.05% off every fill, and adjust for stock splits. That is one reasonable set of assumptions, not a guarantee of what live fills will cost.

How to check costs yourself

Even if your paper account models costs, it is worth a quick sanity check:

1. Pick five recent paper trades. Note the fill price and time.
2. Look up the real opening price for those days, or the real price at that time.
3. Compare. Are paper fills consistently better than the real quotes? If so, the simulator is generous.
4. Count trades per month. Multiply by your estimated cost per fill to get a monthly drag.
5. Subtract it from the paper result if the account does not already.

This takes ten minutes and often changes how you read a paper record.

A worked check, with illustrative numbers

Suppose your paper account recorded a buy of 50 shares at $200.00 at the open on a given day. You look up that day's real opening print and find the first trades happened around $200.30, with the quoted ask a little above that. On the real market, 50 shares at $200.30 would have cost $10,015 instead of $10,000, a difference of $15, or about 0.15%. If your simulator charged 0.05%, it was kinder than reality on this trade. One trade proves nothing, but if four or five spot checks all lean the same way, discount your paper results accordingly.

What costs mean for choosing and testing strategies

Costs should shape how you think about any strategy, not just how you test it:

• High-frequency strategies need strong edges. If a strategy trades daily, ask how sensitive it is to costs. Honest providers will tell you.
• Low-frequency strategies are more forgiving. A monthly rebalancer like Big Tech Trend makes few trades, so per-trade costs matter less, though fund fees still apply.
• Leveraged funds carry higher fees. Their expense ratios are typically higher than plain index funds, and that cost is invisible in a simple price chart.
• Costs differ by route. A gold strategy on a leveraged MT5 account pays different costs from a stock strategy on a commission-free US broker.

For more on how paper and live results diverge overall, see paper trading vs live trading. And for how backtests handle the same problem, read what is backtesting.

Key takeaways

• Slippage is the gap between the expected and actual price; spreads, fees and financing add further costs.
• Small per-fill costs compound quickly for strategies that trade often, and can erase an apparent edge.
• Gold Breakout's own backtest shows that doubling costs roughly halves its yearly result.
• Honest paper accounts fill at realistic times, charge a cost on every fill and adjust for splits.
• Spot-check paper fills against real prices and subtract costs yourself if the simulator does not.

Test with costs built in

AlgolaneBots's paper trial charges 0.05% on every fill and fills at the next real market open, so what you see is closer to what live trading would cost. Start a free 90-day trial with $10,000 of pretend money and no card. Please read our risk disclosure before trading with real money.