Hi everyone. I want to ask professionals and statisticians, how do I know if I have enough data in my journal to make a decision? I've done about 70 trades with a new strategy and in my journal I'm seeing a pattern (see screenshot). The graph is about my trade distribution across different float categories of the stock. I've had nearly equal number of trades in 10-50M, 50-100M and 100-500M float - which is 20. The results for 100-500M float is clearly not good though. I want to stop trading this float range but I wonder if I'm making the decision too soon, do I need more data to confirm this pattern?
There’s a rule of thumb … Sample 10% of the population. How many trade do you expect to make over your lifetime ? If the answer is 200 or less then here you have it but anything below 100 is usually meaningless. I even prefer 1,000 ! The underlying distribution is also determining because the more tails and variance, the more data you’ll need. We can’t even tell with a high degree of conviction if a coin is fair after 20 observations. We might get 7 heads and 13 tails or worse. Does it mean the coin is uneven ? Nope.
It is not possible to put it in 'one number'. Next, certainly doesn't excists. Google for an hour on sampling and stratified sampling and you will have your answer. Your backtest trades should ideally represent your future trades. And ultimo you never can know for sure. Just start small.
I like the cut of your jib. What is your gut telling you at this point? Apply the 70% rule When making decisions under uncertainty, we have a natural tendency to want 100% of the information we need to make the best possible decision. The problem is, it takes time and effort to gather all the relevant information, and often the delay in making the decision is more costly than any advantage you get from having access to 100% of the data. Jeff Bezos understands the power of speed in decision making, and he encourages employees at Amazon to use the “70% rule:” “Most decisions should probably be made with somewhere around 70% of the information you wish you had. If you wait for 90%, in most cases, you’re probably being slow. Plus, either way, you need to be good at quickly recognizing and correcting bad decisions. If you’re good at course correcting, being wrong may be less costly than you think, whereas being slow is going to be expensive for sure.” – Jeff Bezos (2) Data-Driven Decision Making https://www.dogpile.com/serp?q=Data-Driven+Decision+Making
Have you looked at performance by instrument price range? Best Buy Price https://elitetrader.com/et/threads/...st-right-here-baby.335635/page-8#post-4990185
Someone said 1000,someone else said 2000. You could read up on stratified sampling,as suggested,and come up with your own number. You could apply,as suggested,a trusted protocol of one of the most successful men to have ever lived.