Win-Lose
It's easy to get caught up in "win rate" as a metric to measure success rates. But taken by itself, it is a very poor metric to measure success in trading.
In the world of sports, if a team has a win rate of 65%, they are likely making the playoffs.
But, what if we considered the number of points scored versus the number of points allowed per game? What if we determined the teams that made the playoffs not by win/loss record, but by the overall point differential in games played?
This would result in a different set of teams making the playoffs. Sure, there would be some of the same teams, but other teams would make the playoffs that had a lower win rate. Is it also feasible that a team with a higher win rate could completely miss the playoffs because they had a negative point differential? Absolutely. All it takes is a couple of really bad games.
Small Winners, Big Losers
The analogy in trading would be making 10 trades and having a 90% win rate. That's 9 winners and 1 loser. Typically that would be a good thing, but what if each of those 9 winners made 1% while the loser lost 10%? Well, you just posted a 1% loss.
More realistically would be a 75% win rate over 100 trades where the 75 winners resulted in a net gain of 25%, but the 25 losers resulted in a net loss of 40%.
That's a net loss to you of 15%. But aren't you happy with that 75% win rate?
Making Winners Work
Our current win rate for our index of 50 stocks for Day Trading is 68% (YTD) and 68% (YTD) for Swing Trading.
However, that has resulted in gains of 45.4% (YTD) and 191.6% (YTD) respectively.
Like win rate, gain/loss across a large number of trades is not a perfect metric either. Unless you can enter every trade, you may have to be more selective about which trades you enter or have a very large amount of money to trade with.
That's where our algorithm helps.
Our algorithm does the hard part - finding patterns that are more likely than not to predict both direction and duration. However, looking at each alert is also important to determine which trades to enter.
While our algorithm learns every day by looking at both successful and unsuccessful patterns as well as patterns that were abandoned and not alerted, it is not perfect. But, since it is learning, the win rate is constantly improving.
Technically, over time, this will result in less overall alerts, but it will learn to ignore patterns that are less likely to be predictable.
We are also working to improve our algorithm and looking for correlations to add new parameters that will allow our algorithm to be more accurate. We are also working on a rating system that will rate the confidence in each alert based on various factors.
A high win rate is only as good as the gains that are made and a high gain rate is only as good as the number of trades you can actually enter without dilluting your funds.
