That's in a diversified portfolio of long-term investments
Dogma, most often repeated to dazzle small fish into thinking the stock market is a place where money is generated, instead of stolen. It's roughly equivalent to a poker table, and it's a zero-sum game. There's a great illusion where you put in a million dollars and a $5 stock with 200,000 shares issued becomes a $50 stock and woo there's ten million dollars!!
The trick is the idiot does come back with $10 million, because he sees the price going up and up, and wants in on that; you sell it back to him because you see the climb slowing and showing distress, and then it collapses and he cries and sells it back to you for $1 million, and you sell it to the next moron with $10 million in his pocket when it climbs back, and now you have $19 million in your pocket and two poor single-millionaires across the table from you.
The biggest argument against diversification isn't the most obvious. The most obvious is that investing 100% into SPY (the S&P500 tracker) is an instant diversified portfolio; as corollary, any selection of multiple securities--stocks, bonds, options (though they expire), commodities (they get delivered, so you have to keep trading contracts in practice; exchange funds that handle this for you are a close substitute)--behaves like a single security, fluctuating up and down as the market does and as their representative sectors in total do.
A diversified portfolio doesn't magically make money; it simply lowers risk. You must make good buying and selling decisions to make money. In the extreme, a single-stock strategy has the potential to gain *much* more than a diversified portfolio; it can also *lose* much more. In a diversification strategy, you try to select a bunch of securities--stocks or funds which represent a strategy (sector investment, profile investment)--based on what you think would make a good single-security investment, scaling their proportion to their relative risks versus return and your risk tolerance.
That all sounds complex, but it's easily illustrated. Let's say you think BGWNR is likely to climb sharply, making you a 7% profit in the next 6 weeks, but that it's of course more risky than SFBT which you believe is near-guaranteed to make you a 2% profit in the next 6 weeks. You have a pile of money with which to purchase these investments to fill a gap in your portfolio. If your strategy is an aggressive, higher-risk affair, you might put 75% of your money into BGWNR, and 25% into SFBT; if the market betrays you, BGWNR losses should be partially or wholly mitigated by SFBT losses, and the loss in total should be less even if both lose (because BGWNR would lose more of its value than SFBT). If your strategy is a conservative, lower-risk affair, you might put 10% into BGWNR, and 90% into SFBT, because a loss in both would be almost as small as a loss in SFBT, and a win in both would be significantly (but not greatly) larger than a 100% investment in SFBT, and a loss in BGWNR is way more likely than a loss in SFBT and so would probably leave you with *most* of the gain from SFBT if SFBT went up and BGWNR went down.
Does diversification magically mean profit? No. It means less profit, and less loss; it means you don't wake up one day finding out you got wiped out at the race track, and so can keep playing. If you're not a good trader in the first place, you might consistently lose money until you bleed to death slowly. For that matter, the market as a whole has a strong influence on individual securities: MSFT or AAPL can experience a 1.2% drop for the day for absolutely no other reason than the S&P having a shitty -3.2% day (an example of a strategy of all eggs in one basket managing better than a diversified strategy, which, again, isn't a magical rule and shouldn't be taken as an argument that one-bet-on-the-table is a good strategy, either). Your diversified portfolio should move around pretty much like everything else, just less extremely.
Long-term investments are also shit. To be more specific: buy-and-hold is shit; it's used along with diversification so that rich people can buy into small movements, hope for a modicum of stability over a much longer period than you really want to hold any security for, and then sell off and not have to pay income tax (if you hold for less than 1 year, you pay income tax--which can be 39.6% for rich people--versus 15% capital gains). Buy-and-hold is a great and powerful marketing tool used by investment funds who don't care if you make or lose money, as long as you keep it in their fund and let them suck 1% or 0.1% per year out of it.
This is why I no longer play the stock market. I made 1% per day for about a month, then decided to run as far the fuck away as I possibly can. I was waking up at 4am, checking foreign markets, checking foreign exchanges(!), checking commodities, reading news (MarketWatch, Seeking Alpha, anything eTrade brought through, anything UpDown brought through, etc.), reading charts, performing projections, manually calculating metrics (not all the metrics I use are standard-issue technical analysis), etc etc. My mind was in the stock market for 18 hours per day. Fuck that. If I'm not playing that hard, I'm just giving charity to Warren Buffet and Donald Trump; and there's no way I'm playing that hard ever again. There's nothing magic about it; it's stab-out-your-eyes obsession, and you still don't win 100% of your bets.
Maybe I'll take up day trading one day. That's much more casual, believe it or not: the exact same rules as swing trading and trend trading apply, but only over a period of hours. You don't need to make big projections, and you get much faster feedback and instant gratification; the problem is you have to invest more time (you have a few minutes of turn-around before you have to pull your investments, make shorts, buy out your shorts, make new long positions...), which, as I said, is kind of a farce when I'm spending almost zero time trading and 18 hours every fucking day researching. At least when the god damn market closes you're done day trading.