Market participants mostly use the S&P 500 index as their benchmark to compare performance against the overall stock market. This often creates behavioral biases – for example, looking backward and saying that if one didn’t beat the index, then they should just park their money in it, assuming the future will be the same. Obviously, that mindset shows up in an uptrending market because people are performance chasers by nature. Nobody says that in a bear market.
It’s somewhat expected behavior from the crowd since the US stock market has been the standout performer of the past decade. My benchmark, however, is the historical performance of my own strategies – that’s what I need to compare my results to. It’s my job as a quant to make sure the edges are still valid and performing within expectations.
Noise vs statistical validity
Any performance metric shorter than three years is noise for me. Market cycles that affect my strategies can take years to play out, so I adjust my benchmarks accordingly. That’s why I monitor key performance metrics for my strategies on a rolling three-year period, reviewed every six months. Every half-year, the oldest six months drop off and get replaced by the most recent six months of results.
This robust approach applies even to day-trading, because a short-term strategy is still part of a long-term business. It keeps me ignorant to daily predictions and market stories – the psychological traps and noise the retail crowd tends to fall for.
One might assume that reviewing performance twice a year means I do it in January and July. Nope. I do it in March and September – purely random, because my trading business doesn’t care about the 12-month calendar window. I’m a private trader, not a fund manager. The last review was at the beginning of September, covering trades from September 2022 through August 2025.
Three-year stats for my stock trades (excluding options and futures):
- Trades: 4022
- Win rate: 54%
- Avg win/loss ratio: 1.48
- Profit factor: 1.72
One of the main metrics I use is profit factor, calculated by dividing the total dollar amount won by the total dollar amount lost. It tells me how many dollars I make for every dollar I lose. For example, a profit factor of 2 means that for every $1 lost, I made back $2. That’s a great business – exchanging a dollar for two.
Of course, it depends on how many dollars I can exchange like that. So profit factor alone doesn’t say much without considering trade size and frequency. I could exchange $1 for $2 through ten trades a day with $100 risk per trade at 50% win rate, meaning 5 × $200 − 5 × $100 = $500 profit in a day. Or I could do it through twenty trades a week with $1000 risk per trade at 50% win rate, making 10 × $2000 − 10 × $1000 = $10k profit in a week. The profit factor of 2 is the same in both cases – it shows relative profitability, not absolute. That’s why I also look at the average trade size and how much a strategy makes per 1, 10 or 100 trades depending on its frequency.
In the stats above, a profit factor of 1.72 might look low, but that’s over 4000 trades. My long-term trend strategy alone has a profit factor of 4, but with fewer than 200 trades. The MOMO strategy I wrote about recently has a profit factor of 1.9 – and it’s been solidly profitable.
Stats need at least a hundred trades to mean anything. Otherwise, I could take just two trades – one big NVDA win and one small stopped-out loss to show a profit factor of 100 at a 50% win rate. That wouldn’t be valid.
I don’t care about my portfolio’s percentage return anymore, and I don’t even track it. My main profitability goals are:
- To multiply capital through long-term position trading by capturing big moves and compounding
- To maximize turnover through short-term swing and day-trading with positive expectancy models
So, on one hand, I wait to capture big long-term moves that create size, and on the other, I churn dollars quickly through short-term models that create frequency.
Now let’s keep losing these $1 bills to make more money!
Share this post