Forward testing with small real money

[Amanhasnoname]

Demo results and backtests have limits. The step between them and full size is a small live account, and many skip it.

A small live account captures things nothing else can: real fills, real slippage, real emotions. Even a tiny size makes losses feel real, and decisions change.

A reasonable plan is trading the minimum size for a few hundred trades, tracking every detail, and comparing the outcomes with the backtest. If the live result lags by a stable amount, that gap is the true cost of execution. If it lags by a lot or in an inconsistent way, something in the model is wrong.

The amount at risk is tiny compared with the information gained. It also reveals practical problems: platform quirks, broker limits, and personal habits that only show up with money on the line.

Scaling up should be gradual, with each step confirming that results hold.

How long was the forward test period before scaling up, and what surprised you most?