Technically you can use that algorithm to make money and it will work but it is discouraged because that is someone else's intellectual property and not your own idea. People want to develop their own algos because using someone else's is basically cheating.
Algorithmic trading is designed to be faster, more scientific and more efficient than discretionary(manual) trading. Let's be honest here - who do you think will have better performance?- John and Sally on an E-Trade account looking at 15-minute delayed data trying hard to find a stock or two to invest in, OR; You or me with a very fast and efficient algorithm that can find stocks to buy and sell within a minute and place hundreds of orders for thousands of shares in about five seconds, while you sip coffee and thank God that you aren't John and Sally. The answer is obvious. Algorithmic trading is much easier and often times more lucrative than discretionary trading. Welcome to the new era of investing.

Owen

Lot of the algos shared in community are to demonstrate a principle. You can always use the same principle
in real life trading (at your own risk).

Most of these algos (as writen) are not having specific technical protections needed in live trading.
Even the winning algo in contest can not go live as it is. You can find here what are such pitfalls.

Thanks Yagnesh and Owen!...

@Owen - i thought if it is shared to be accessed by public, then we can use it at our own risk. It is like open source kind of thing if the author/creator of the algorithm has willingly shared to public. How can it be cheating or wrong doing?... And i understand the efficiency of it Vs the manual trading.

Btw, i 'm planning to use it to generate buy/sell signal and let me act on it in my existing brokerage account, rather than automatic live trading direct from the Quantopian. Is that something possible and effective?.. I 'm not intending to do high frequency trading, nor day trading. But some sort of positional/speculative trading leveraging the technical parameters and underlying trend?... Is there any way to send some alert msg or e-mail whenever the algo generates buy/sell action ie, instead of placing the order i need the order info in an email?...

Your inputs/suggestions are most welcome!..

Thanks again!

Vad

I think anyone is welcome to use algos which are posted however they like, but I would strongly caution against trading something that you have not created.

If you didn't make it, and it starts to lose money, you won't necessarily know why, and you might not have a good idea whether you should stick through it or whether the system is permanently broken.

There's the additional problem that most of the algos posted here do not have much in the way of error handling, since they are shared to share an idea, and hundreds of lines of tedious error handling code will detract from the message/algo.

There's also the problem, of course, that any algos which look great on paper may have been overfitted within an inch of their life, and so performance in the future will be disappointing, to say the least.

Hey Vad

Welcome to Quantopian! I'd like to give you some input on your questions. First, your question about using other people's algorithms. We highly encourage you to use other peoples ideas! You'll notice that when someone posts a backtest there is the little "clone" button, that is there for a very important reason. If you look back at history you'll see that most scientific advances were not major revelations but an incremental improvement on existing work. At Quantopian we believe that this iterative clone and tweak process will yield great algorithms. As you are looking through the forum take special note of the posts with a long comment chain, you'll often see a lively discussion about the pros and cons of a strategy. Nestled within those comments will usually be backtests of people who cloned the OP's algorithm and improved it, or perhaps they made it fail on purpose. It is debate and criticism that are the crucible of idea creation, and the end result is where we find brilliance.

In response to your question about why people seem to dismiss algorithms with high returns, it's not about returns so much as it's about risk. Any one can through darts and maybe find a real winner with incredible returns, but in doing so they take on a awesome amount of risk. As quants we think we can do better, consequently many high return algos are either tossed or tweaked to bring risk to a level that is comfortable to the investor. The game is about returns and risk management. Part of that risk management is knowing how your algo works and so blindly copying an algo, and throwing real money at it presents risk as you have no good knowledge of its economic underpinnings nor what to do when it fails.