In simple words, if there is a chance that you loose all your money, it is speculation and you should not do it. If you can gain a lot but only only in case of a very rare event (as in the lottery), it is also speculation.
Day trading is costly and risky. It is costly because you'll pay a large amount in commission fees, and you will pay a lot of tax. You need a 10% gain on every transaction just to break even. One study showed that people who traded at least 20% of their portfolio per month underperformed the market by 6.4%, after brokerage costs. Those who traded less than 0.2% a month beat the market slightly (Barber and Odean, http://faculty.haas.berkeley.edu/odean/Current%20Research.htm. Tax: If you trade on a daily basis, the IRS counts your gains as income, instead of as long term gain, which means paying more tax.
If you could time the market, you would get out of stocks just before the market crashes, but history has shown that even experienced analysts are not good at this.
Don't try buying IPO's (initial public offerings). It is almost impossible to get the stock for the initial offering price, because those shares go to investment banks. Some stocks are small in number, but then you fight with other small investors, which can be just as difficult. Most investors get access to the IPO only after the stock price has peaked. Also, many IPO's increase in price the first month, but go out of business after that.
This is a share that is something in between a common stock and a bond. Pays a fixed dividend, so you don't get more when the business goes well. Also, if the company goes out of business, it's the bondholders that have the legal claim on the company's assets, not the preferred stock holders. Further, companies get a tax-break for interest payments on bonds, but not on preferred stock, which is not an advantage. Finally, a company may issue preferred stocks because there wasn't much demand for the bonds, which obviously has a reason. Long story short: Don't buy them.
A forward P/E ratio is obtained by dividing the current price by predicted future earnings. The problem is that you cannot predict the future. You have expectations, like everybody else. Earnings forecasts very often turn out to be very wrong, and it is not of any use to divide the current price by earnings or losses that have not been realized yet.