There is of course one sure-fire way to make a fortune from reliably predicting tomorrow's share price for any type of share. First obtain an email list of 20 million middle-class individuals. One Day One, email 10 million telling them stock X will go up tomorrow, and email the other 10 million telling them stock X will go down tomorrow. Then wait to see what happens. If the stock goes up, discard the second batch of 10 million. Now break the remaining 10 million into two groups of 5 million; tell the first group stock Y will go up tomorrow, tell the second group it will go down. As before, wait & see, discard the failed prediction group, split the successful prediction group and repeat. After 5 days, the 625,000 people remaining in the "successful predictions" group will believe you're the most gifted stock-picker in history. Now send an email saying "now I've proved how great I am at predicting stock price movements, send me $20k and I'll return you $40k within 5 business days." At least 1% will fall for this and send you the cash so you walk off with $1.25 million for five days of trivial work. This is basically the only way to make serious cash out of stock price predictions on a day-to-day basis where random fluctuations make even the most sophisticated model worthless.