Paper Trading and Going Live Gradually

5.1 The backtest passed. Now what?

Mint's trend system had passed every check in Chapter 4: robust parameters, decent out-of-sample results, realistic costs. Mint was ready to commit a large portion of the account.

"Not yet," said Hash. "A backtest tells you how the rules behaved on data you already had. It doesn't tell you how they'll behave on data that hasn't happened yet, or how you'll behave when real money is involved. That takes forward testing."

5.2 The stages from idea to live

Stage	What happens	What it tests
1. Backtest	Rules applied to historical data	Whether the idea ever worked
2. Paper trading	Rules applied in real time, without real money	Whether it works on new data, and whether you can follow it
3. Small live trading	Real money, very small size	Execution, costs, and real emotions
4. Gradual scaling	Increase size step by step as evidence builds	Whether results hold at larger size

Chapter 5 · From idea to live

A strategy passes through a pipeline: backtesting on historical data, paper trading in real time, trading live with small size, and finally scaling up. Each stage checks what the previous one could not, including real costs, ambiguous rules, and emotional discipline, and problems send the strategy back a step.

5.3 Paper trading

Paper trading, also called forward testing, means following your system in real time and recording every trade as if it were real, without risking money. Many exchanges offer demo accounts or testnets (Module 2, Chapter 5); a simple spreadsheet also works.

Paper trading reveals	Example
Rule ambiguities	"The rule says 'daily close,' but which exchange's close?"
Practical issues	Signals arriving at inconvenient times
Whether results match the backtest	Similar win rate, trade frequency, and R-multiples
Your ability to follow the system	Did you skip any signals?

Weakness: paper trading doesn't involve real emotions, real fills, or real slippage. Many people follow rules perfectly on paper and differently with money at stake.

5.4 Comparing live and backtest results

As paper and live trades accumulate, compare them with the backtest.

Metric	Backtest (illustrative)	Forward test after 3 months (illustrative)	Concern?
Trades per month	~0.7	1	No
Win rate	35%	33%	No; within normal range
Average win	4.2R	Not enough wins to judge	Need more data
Average cost per trade	0.1%	0.15%	Yes: costs higher than assumed

Mint discovered that real costs were nearly double the backtest assumption, because of slippage on the exchange Mint used. That was a valuable finding before scaling up.

Early forward results will vary from the backtest, especially with few trades. Look for systematic differences, like costs or missed signals, rather than reacting to short-term results.

5.5 Going live small

When moving to real money, start much smaller than your intended final size.

Phase	Risk per trade (illustrative)	Move to next phase when
Starter	0.25%	At least 10 to 20 trades executed correctly, with no operational problems
Intermediate	0.5%	Results broadly consistent with expectations; rules followed
Full	1%	Sustained consistency over several months

Small live trading tests what paper can't: real fills, real costs, and real emotions.

"It feels slow," said Mint.

"It is slow," said Hash. "And it's much cheaper than discovering a problem at full size."

5.6 Rules for following a live system

1. Take every signal. Skipping signals you don't like changes the strategy, and often skips the best trades.
2. Don't override exits. If the rules say exit, exit (Module 12, Chapter 4).
3. Log deviations. If you do deviate, record why, and review it (Module 12, Chapter 13).
4. Don't change rules mid-stream. Make changes only during scheduled reviews, after testing.
5. Set a system-level stop. Define in advance the drawdown at which you'll pause the system and investigate (Chapter 11).

5.7 The hardest part

Mint's system took four small losing trades in a row in its first live month, during a choppy market. Mint felt the urge to skip the fifth signal.

The fifth signal became the biggest winner of the year.

"The system did what the backtest said it would," said Mint. "I'm the part that almost broke."

"That's the most common story in systematic trading," said Hash. "Most systems fail because people stop following them at the worst moment, not because the rules were wrong."

5.8 Try it yourself

Paper trade the system you backtested in Chapters 3 and 4 for at least one month, logging every signal and whether you would have taken it. Record realistic fills using the actual prices at signal times. Then compare trade frequency, win rate, and costs with your backtest.

Key takeaways from this chapter

1. After backtesting, systems should be forward tested before full commitment.
2. The path is backtest, paper trading, small live trading, and gradual scaling.
3. Paper trading reveals rule ambiguities, practical issues, and whether you can follow the system, but not real emotions or fills.
4. Compare forward results with the backtest, looking for systematic differences like costs or missed signals.
5. Go live at a fraction of intended size, scaling up only as evidence builds.
6. Take every signal, don't override exits, log deviations, and change rules only during scheduled reviews.
7. Systems most often fail because people stop following them during losing streaks.