Big Data, Big Profit
In This Week’s Issue:
- Stockscores’ Market Minutes Video – Trading Channels
- Stockscores Trader Training – Big Data for Big Profit
- Stock Features of the Week – Stockscores Abnormal Breaks
Stockscores Market Minutes – Trading Channels
This week, I look at the importance of understanding price channels and how it made me profits this past week. Plus, my regular weekly market analysis and the trade of the week on DOMO.
Click here to view
To get instant updates when I upload a new video, subscribe to the Stockscores YouTube Channel
Stockscores Trader Training – Big Data for Big Profit
We have all heard the story about the King who asks a group of blind men to feel an elephant and report back on what an elephant is. Each feels a different part and as a result, each has a very different perception of what the elephant is. They fail to accurately understand the Elephant because each does not touch the entire animal.
Many traders fall in to a similar mistake when evaluating their approach to the market. It is very easy to misjudge the effectiveness of trading rules by looking at the result of the last trade. If you buy a stock because it is breaking to new highs and the trade ends up failing, it is easy to say that buying stocks breaking to new highs is not an effective strategy.
This concept is referred to by some as being fooled by randomness. By drawing conclusions from a small sample of data, the trader makes an incorrect assessment of cause and effect.
To really judge the effectiveness of strategy rules requires they be tested over a large sample size, at least 30 trades but more is better. Only then can you start to see patterns and correlations. Only then can you assess cause and effect.
Suppose you are sitting in front of your computer and you decide that you will buy shares in Herbalife (HLF) if the next car that drives past your window is blue. The next car that drives by is blue so you buy HLF and the trade ends up making you a $1000 profit.
Encouraged by your result, you take a look at Pfizer (PFE) and again determine that you will buy the stock if the next car that drives past your window is blue. The next car surprises you by being blue so you buy and again, you make a profit. Trading seems easy!
What do you think would happen if you carried out this rule for your next 30 trades? Since most will realize that there can be no cause and effect between a blue car and a winning trade, most will say that the overall result should not be positive. Intuitively, you know what there can be no correlation between the color of the car that drives past your window and the performance of your trades.
However, what if your test actually finds that 25 out of the 30 trades you do end up being winners? Is there now reason to believe that blue cars predict strong stocks?
The problem is that even when there seems to be a correlation between one factor and a result, it could simply be that there is another cause at work. The reason that there was 25 winners out of 30 could simply be due to a strong trending market that makes most stocks rise.
This example highlights two important considerations when assessing the effectiveness of strategy rules.
First, make sure you test a rule over a large sample to get data that is reliable.
Second, test your strategy rules over varying market conditions so you can remove bias.
When testing the rules of a strategy, do not stop at the entry rules. Evaluate the exit strategy and how you size positions and do risk management. Small changes in any of these areas can have dramatic effect on your profitability. I recently completed a 5-month test of one of my day trading strategies and found that adding one rule to the strategy more than doubled the profitability during the test period.
If you want to truly understand how well your trading strategy works, take the time to compile data on a large number of trades across varying market conditions. Avoid looking at just one factor or the results of your last trade.