At my never-to-be-revealed age I’m more risk-adverse than I was 30 years ago (oops! I may have just inadvertently dated myself) and the metrics that are most important to me may not be the same as yours. With that in mind, let’s talk performance.
Percent profit per trade
Beware of too little percent profit per trade (%P/T) so that any slippage and commissions would wipe out all your gains. Strategies that trade more frequently (like dip buyers that enter and exit positions after 1 or 2 days) tend to have lower values for %P/T, perhaps 0.6% or less. That strategy trading $10K position sizes would then make an average of $60 profit per trade.
But imagine you traded it with an account allocation of only one-fifth that size, i.e., $2,000 positions. Your average profit would be only $12. Add in $2 to $20 of commissions (depending on your broker) and a probable negative effect of trade slippage, your expected gain may even be negative! While strategies with low $P/T are often able to generate significant profit due to many trade opportunities, they may not be a profitable endeavor for smaller allocations.
Win/loss ratio of trades (or Winning Trades Percent)
While it’s not required for profitable strategy, trading is easier, and it just plain feels better if you have more winning trades than losers. You can take this with a grain a salt though. What’s important is that the percent profit of winning trades is [significantly] greater than that of trade losers. That’s the first bit of evidence that a strategy author allows winners to run and more quickly cuts losses. In the end, higher ratios of both add to the robustness of a strategy.
Equity and Drawdown curves
Equity and drawdown curves are interrelated in the following way: Drawdown is the value, expressed in either a dollar or percentage amount, that your portfolio has fallen since achieving the last peak equity value. The max drawdown (Max DD) then is the greatest negative value of one of these “dips”. It’s typical to find strategies with the best overall profit metrics with a Max DD in excess of -40%. Don’t forget that you need to gain +67% to recover from a -40% drop and +100% to recover from a -50% drawdown! You may be different, but these are not the kind of strategies that I can stick with while trading.
Warren Buffett has written (and spoken) several times that he prefers “a lumpy 15 percent return over time than a smooth 12 percent.” Everyone who trades wants to end up with more money; that just makes sense. But if you’re watching your account value daily – and maybe even count on some returns to “keep the lights on” – it may not be too sensical. (I do not advocate that you trade with money that you cannot affort to lose. Trading stocks is one of the riskiest things you can do with your money.) It’s not as easy to stay the course while experiencing lumpy -30% and -40% drawdowns as it is with an account that has a smoother, upwardly sloping equity curve. And, if you can’t stay the course, you’re likely to quit precisely at the wrong time.
Something that I really like and is unique about WealthSignals is the ability to see a backtest/research equity curve connected to the WealthSignals out-of-sample equity curve. You can easily get a feel if a strategy is continuing to perform in a way expected or “observed” in the past. If performance isn’t par with the observations in the past, there’s a fair chance that the strategy has been over-optimized, or possibly the market conditions have changed so much that the strategy is just no longer effective.
Monthly performance table
How many months in the year [or in a row] did I make money? How many months [in a row] did I lose money? A monthly performance report, especially when graphed as a histogram, it lets you see in a glance how constant (or variable) returns are over time and how often you can expect no or negative returns – a measure that may not be obvious in an apparent ever-increasing equity curve that spans a decade or two.
I like to compare the backtested (research) numbers and the WealthSignals out-of-sample numbers. If the discrepancy is too high, I would like to find out why it is and contact the system author.
Feel free to post your questions!