Slippage and commissions: why your real result is never the backtest result

The Rentabilio backtest shows $274,406 in gross profit over more than seven years on a $50,000 account. Your broker statement will never show that number. Not because the backtest is dishonest, but because a backtest is a model of trading and trading has a toll booth on every trip.

The gap has two parts. One is fixed, known in advance, and easy to calculate. The other is variable, occasionally ugly, and impossible to know exactly until it happens. Here is what both are worth, in dollars, on this specific system.

The two costs that separate a model from a statement

Commission is what you pay to have a trade executed: a known number, published before you trade, charged per contract per side. Nothing about it is a surprise.

Slippage is the difference between the price the model assumed and the price you got. Sometimes it is zero. Sometimes it goes your way. Over enough trades it tends to go against you, because the moments when the market moves through your order fastest are the moments your order most needed to be filled.

Backtesting software fills you at a price the software chose. The real market fills you at a price the market chose. That is the whole subject.

What a round turn on a micro contract actually costs

A "round turn" is one contract in and back out: an entry and an exit. It is the unit costs are quoted in, and it stacks up from three separate places.

That figure (roughly a dollar in and out per micro contract) is the number used throughout this site, and it is representative of what a retail or funded account is charged on micro index futures. Shop hard and you might do slightly better; take the default rate at a full-service broker and you will do worse. Either way it is not five dollars and it is not twenty cents.

Note what is not in there: platform fees, market data, and your VPS if you rent one. Those are monthly costs that do not scale with how much you trade. Keep them in a separate column in your head.

What that does across 4,557 trades

The published backtest contains 4,557 trades. At roughly $1 per round turn, one contract at a time, pure commission comes to about $4,600. Against $274,406 of gross profit that is about 1.7%.

But the net figure published on this site is ≈$260,700, about 5% off the gross. The haircut is roughly $13,700, or about $3 per trade, which is deliberately more than commission alone.

The extra two dollars a trade is a slippage allowance. It is not a measured number, it is a conservative assumption: a third of the haircut covers what we know we will be charged, and two thirds is padding for fills that do not match the model. On micro index futures a tick is worth between $0.50 and $1.25 depending on the contract, so a $2 allowance across an entire round turn is generous on products that trade a one-tick spread most of the day. Done this way, the cost assumption errs on the expensive side rather than the flattering one. That is the direction an honest number should lean.

The arithmetic, written out

The whole thing reconciles from the average trade, which is worth seeing because it shows how thin the per-trade edge really is.

Sixty dollars a trade, gross. That is the entire edge. It survives a three-dollar toll comfortably. It would not survive a fifteen-dollar one, which is exactly the situation a lot of higher-frequency systems quietly live in.

Why the backtest is published gross

Because it is the number you can reproduce. Load the strategy into the Strategy Analyzer in NinjaTrader 8, set the same period, hit Run, and the report prints the gross figure. If we published only the net, you would run the test, get a different number, and have no way to tell whether the difference was our cost assumption or something worse.

So gross is the verifiable number, and the net sits next to it with the assumption spelled out: about 5% off, at roughly $1 per micro round turn. Plug your own broker's rate into the calculator for your own version; the steps to reproduce the report are on the performance page.

A number you can reproduce and adjust is worth more than a number that has already been adjusted for you.

What slippage actually is, and when it shows up

Slippage is what happens when the price you wanted is not the price available at the instant your order reaches the exchange. There are only a few situations where it is material, and it is worth knowing which ones apply to you.

• A thin book. If only a few contracts rest at the best price and you want more, you take the next price up, and the next. On micro index futures during the US morning this is rare. On a quiet product at 2 AM it is the normal state of affairs.
• Data releases. In the seconds around a scheduled release the book empties, refills, and prices jump levels rather than walking through them. Everyone knows the number is coming, so nobody leaves size resting into it.
• Stop orders in a violent move. This is the big one. A stop is not a price guarantee: it is an instruction that becomes a market order the moment it is touched. If the market gaps through your level, you get filled wherever the market is, and your risk on that trade is briefly larger than planned.
• Contract roll. Around quarterly expiration, liquidity migrates from the front contract to the next one. Trading the wrong one is self-inflicted slippage.

Here is the uncomfortable part, said out loud: Rentabilio trades at 8:30 AM ET, which is precisely when US economic data is released. That is the second bullet on the list. The system deliberately operates in the window where slippage is most likely, because that is also the window where the movement it needs happens.

Two things offset it. That window has the deepest liquidity of the pre-open by a distance: volume around the release is enormous, which is the difference between a fast market and a thin one. And the system does not fire the instant a number prints; it reads what happens after, which is a calmer place to transact than the first two hundred milliseconds. None of that eliminates slippage. It is why the allowance is two dollars and not twenty cents.

Why a wide stop and a 2-to-1 target barely notices

This is the part most people get backwards. Cost sensitivity has almost nothing to do with how good a system is and almost everything to do with how big its trades are.

A scalper hunting four ticks hands most of the gross edge to the toll booth, and one tick of slippage on one leg wipes out the trade. A system with a wide stop, an average loser of $193 and an average winner of $354 is playing a different game: a bad fill makes a trade slightly worse, it does not turn a winner into a loser.

This is the practical reason a 2-to-1 reward-to-risk structure is worth having even though it means losing more often than you win. The wider the trade, the less friction matters. And friction is the one thing in trading you can count on.

Four ways to keep your own costs down

1. Shop the broker rate, once. Commission is the only component you can negotiate, and the spread between a cheap futures broker and an expensive one is real money over four thousand trades. Do it before you start.
2. Read the funded account's schedule. Prop firms charge their own commissions and are not always the cheapest. That cost sits alongside the evaluation fee. See funded capital.
3. Trade the liquid contract. Front month, roll when volume rolls, stay in products with real depth. Cleverness about instrument selection is usually slippage in disguise.
4. Do not add trades. The most expensive habit is manually adding entries the system did not take. Every discretionary trade pays the same toll with none of the tested edge behind it.

Frequently asked questions

Is slippage always bad for me?

No. Slippage is a deviation, and it goes both directions: sometimes you get filled at a better price than the model assumed, which is called positive slippage. Over a large number of trades, though, the average tends to run slightly against you, because your orders are most likely to be filled instantly precisely when the market is moving away from you. Assume it is a small net cost and plan for it, rather than hoping it nets to zero.

Why not just use limit orders and avoid slippage entirely?

A limit order guarantees your price but not your fill. If the market never trades back to your level, you simply do not get in, and the trades you miss will disproportionately be the ones that ran hardest in your direction. Exits are worse: a stop has to become a market order or it is not a stop at all, and refusing to accept a bad fill on a stop means accepting an unlimited loss instead. Slippage is the price of certainty about being in and out.

Do commissions get worse when I trade more contracts?

Commission is charged per contract, so two contracts cost twice as much as one. But the gross profit scales the same way, so the percentage stays flat. What can change is slippage: a larger order may need more than the size resting at the best price and fill across several levels. On micro index futures at one to five contracts that effect is negligible; it becomes real at institutional size, not here.

Does the backtest include slippage?

The gross figure of $274,406 is the raw Strategy Analyzer result, which does not model slippage. The net figure of approximately $260,700 applies a 5% haircut that covers both commission at roughly $1 per micro round turn and a deliberately generous slippage provision on top. Both numbers come from a simulation over historical data; they are hypothetical results and no real money was at risk.

How do I know my broker's real all-in rate?

Ask for the rate per side including exchange and NFA fees, in writing, for the specific micro contract you will trade, and then check it against a real filled trade on your statement once you start. Advertised commission rates frequently exclude the exchange and regulatory components, which makes the headline look better than the charge. The number to write into the calculator is the all-in one.