5 most important things I consider when choosing the trading system
Probably you are the trader with automated strategies in your investment portfolio. Or you considering to be one of us. And you it know well, there are so many criterias around when you choosing the right automated trading system.
Profit factor, Sharpe ratio, drawdown, stop loss, slippage, intraday system vs. swing system and myriads of others. But how to select the right criterias and don´t immerse yourself into the ocean of parameters?
I focus on my 5 metrics first. Supposing the system I want to test in detail passes elementary attributes as the overall profitability, reliability of the developer, system provider etc.
My favourite metrics when evaluate prosperous trading system are followings:
1/ Average trade
Average trade tells me, how the strategy handles the risk of bad fill and overall trade expenses.
Every trade bear trade expenses. When put aside strategy subscription, I have to count with trade slippage and trade commission. If the average trade is not big enough, trade expenses could easily destroy total strategy performance.
I like the average profit to be at least 3x trade expenses.
For example. When trading e-mini S&P futures:
– @ES, 1 tick value is $12.50 I have to count with 1 tick slippage
– my broker commission for 1 contract is $2.5
My trade expenses are $12.5 + $2.5 in and $12.5 + $2.5 out. $30 total trade expenses round turn.
I wouldn´t trade anything below $75 avg. trade profit and I would choose those strategies with avg. trade profit over $90. So when I am not filled correctly, bad things happen, and I suffer 1 or 2 tick more slippage, I still close the position green.
2/ Number of total trades
The strategy should record enough trades to be statistically significant. I consider as absolute minimum 350 total trades in total history. And I prefer strategies with at least 700 trades in the history. More trades here gives better statistical significance.
Mostly I choose the strategy with 4 – 10 trades per month. I am not the fan of strategies trading daily or almost daily. I like my strategy to select the right market conditions first before enter the trade.
Scalping strategies trading daily make small average trade profit and few ticks more slippage could damage the strategy completely.
I like to trade the strategy „edge“ which is not visible so often. Too many trades in month increase the chance of being finded by others easily. And massive trading erases the strategy „edge“.
The gold rule for me here is 2 – 10 trades per month.
3/ Annual Profit distribution
I like to see at least 10 years history (about in-sample and out-of-sample periods I´ll write in some future post). In the other side, the rule „more history = better system“ I don´t apply here.
History since 2006, 2007 is enough for me, it covers the 2007 – 2008 financial crisis. And history before 2002 I don´t consider to be relevant for today´s system.
Annual profit distribution should show ALL YEARS PROFITABLE with possible ONE exeption. I have to see every year profitable with smooth distribution. Best when most profitable years are double of those years with less profits. I can allow one year with limited loss, up to 10% of average profit.
4/ Monthly profit distribution
In monthly profit distribution I want to see trades in most months, at least in 10 months in a year. Good system doesn´t need to trade every month. 70% months should be profitable. Not long loosing strikes.
3 consequent loosing months should be rare exception, only once of few times in all history. Best when a loosing month is bordered by more profitable months.
And again, smooth distribution is the key. Target is the profit to be spread during most part of the year. In the other side you have to count with losing months too, 1 – 2 months in the red per year I consider as acceptable.
5/ Average annual profit / worst case drawdown ratio
This my own proprietary ratio, you have to calculate it manually in most cases. It cannot be found in Strategy Performance Reports normally.
It is my most favourite parameter. It tells me, how long could worst case drawdown lasts.
Every strategy inevitably passes periods with drawdowns. Some brokers use to refer to mont-to-month drawdown. But it doesn´t tell, how big is the max. historical drawdown. There could be several consequent month-to-month drawdowns, which changes the picture completely. Make sure you always know the worst case historical drawdown achieved by the strategy in 10 years history.
Most strategy reports shows average monthly return. I take that number, multiple by 12 to get average year and divide by worst case drawdown.
Best strategies have Avg. Annual profit / Worst Case Drawdown > 2.0 It tells me, that worst backtested historical drawdown would be erased in 6 months.
I never trade strategies with AAP/WCDD worse than 1.3, my target is to select those with the ratio over 1.5
And that´s it. You may ask where is the Profit factor, Equity, Sharpe ratio and the others?
Well, when strategy passes my 5 criterias, I consider it as tradable. I find then Profit factor over 1.6, nice equity and Sharpe over 0.35 or 0.4.
This setup works for me. Is anything else what works for you?