Paper Trading: What It Proves and What It Doesn't

Paper trading is running a strategy on live, real-time market data without risking real money. It answers one question well: does the strategy's logic hold up outside of historical data, on prices nobody has seen before today. It answers almost nothing about how you, personally, will behave when a real account is on the line.

What is paper trading?

Paper trading means feeding a strategy live market data and letting it generate orders, but routing those orders to a simulator instead of a broker. The prices are real. The fills are not. Most futures platforms, including the ones built on Rithmic feeds, offer a simulated account for exactly this purpose.

It sits between backtesting and going live. A backtest runs the strategy against years of historical bars, all at once, with the benefit of hindsight baked into the data structure. Paper trading runs the same rules forward, one bar at a time, in real time, with no hindsight at all. That distinction matters more than most traders give it credit for.

What paper trading actually proves

Paper trading confirms that a strategy's code does what its backtest said it would do. It catches the bugs a backtest hides: a data feed that lags during high volume, an order type that behaves differently in the live session than in the historical replay, a signal that fires twice because of a timestamp quirk. These are mechanical failures, and they only show up when the strategy is running against a live clock.

It also proves the strategy can handle a real trading session end to end. Market open volatility, midday chop, the scramble around economic releases, the close. A backtest sees these as rows in a spreadsheet. A forward test sees them as they actually unfold, with the strategy reacting in real time to data it has never processed before.

For a rules-based system like the one Karani runs, this stage is not optional. Every strategy gets paper-tested on live data before it ever touches a client's funded AMP account. It is the last mechanical check before real capital is at risk.

Paper trading can prove a strategy works. It cannot prove you will follow it when the money is real.

The gap paper trading can't close

Here is the part that gets skipped in most explanations of paper trading: it cannot test you. When the strategy paper-trades a losing streak, nothing is actually lost. There is no discomfort, no second-guessing, no urge to override the next signal. That absence of consequence is precisely what makes paper trading safe to run and precisely what makes its psychology unrepresentative of live trading.

A trader who watches a strategy lose eight simulated trades in a row reacts very differently than the same trader watching it lose eight real trades in a row. The eighth real loss comes with an account balance that has actually shrunk, a broker statement that actually reflects it, and a very human impulse to intervene. Paper trading will never generate that impulse, because there is nothing real to protect.

This is exactly why automation matters more once a strategy goes live than it did during testing. A system that follows its own rules regardless of how the last eight trades felt removes the one variable paper trading can't simulate: a person deciding, mid-drawdown, to deviate from the plan.

Slippage and fills: the other blind spot

Simulated fills are usually kinder than real ones. A paper account often fills your order at the exact price the strategy requested, or close to it. A live order in a fast market can fill several ticks worse, especially in the E-mini S&P when volume spikes around an economic release or the cash open.

One ES tick is 0.25 points, worth $12.50 per contract. That sounds small until a strategy trades often and slippage eats a tick or two on both entry and exit, every time. A strategy that looks marginally profitable in simulation can turn marginally unprofitable once real fill quality is subtracted, which is one reason live results are watched closely even after a strategy clears paper trading.

How long should you paper trade?

Long enough to see the strategy operate across more than one kind of session. A single calm week proves very little. You want to see it handle a volatile week, a quiet week, a week with a major data release, and ideally a stretch that includes a losing streak, because a strategy that has never lost on paper has not been tested against the condition that matters most.

There is no universal number of days that makes a forward test complete. What matters is variety of conditions, not just duration. A month of identical, low-volatility sessions teaches less than two weeks that happen to include a sharp selloff and a sharp reversal.

The honest use of paper trading is as a mechanical filter, not a promise. It weeds out strategies with real execution bugs and confirms the logic survives contact with live data. It cannot tell you whether you'll follow the rules when your own money is on the line, and it cannot fully price what a live fill will cost you. Those two gaps are why a forward test is a step, not a finish line.