An algorithmic trader explains why back-tested strategies fail in live trading and introduces the Monte Carlo reshuffle test to stress-test them. He compares managing a multi-strategy portfolio to running a sports team: keep the winners, cut the losers.
As an algo trader, you have to expect that the strategy that you will allocate will fail. There is a concept we call Monte Carlo. So basically the Monte Carlo test is you reshuffle your trade to see what is your worst outcome. Because nowadays every software will generate a beautiful equity curve. But it doesn't mean this is the equity curve that you will have in real time with your real money. You have to have some process to turn off strategy. When you manage a portfolio of strategy, it's like managing a sports team. Your goal is to let your plus player on the field and removing the less performant one.
