Well, you are sort of in the wrong place then, very few people do automated algorithmic execution here. You might have better luck on Nuclear Phynance or something. Sounds like you have the right books though!

Am I missing something here? aren't most users on this site using back testing strategies? what's the difference between algorithmic trading and automated algorithmic execution?

@Simon, I don't think the poster was focused exclusively on OEX algos. Wouldn't it be true that every algo here is an example of algorithmic trading?
No doubt @Aslan H. you discovered that algorithmic trading is a nebulous term that would have a vast range of definitions. I would also gather that you have not exhausted your searching and skimming skills yet. The answers you seek are available with continued searching. Once you learn what tools are available to both use and to build algorithms for trading then you'll be able to answer your own questions. For instance:
A simple household member can trade right here on Quantopian. Or that same member could use various tool sets to construct their own trading system. Your term "simple" means what exactly? My household is simple (rent, two old cars, kids in school, struggle to make ends meet...) A couple of hours of searching and reading will give you a better picture of what algo trading really is.

I am interested in Algorithmic Trading and not High Frequency Trading/BlackBox

I figure every algorithm on here is a BlackBox trading algorithm, and very few have sophisticated OEX sub-algorithms...

@MarketTech, by simple household I simply meant anyone with a computer and an internet connection. Don't get me wrong i've skimmed through multiple books and the one i'm currently using is the only one that mentioned DMA.

"Simple household" and DMA are not even in the same universe...

Algorithmic trading, if you take the pure sense of the word "algorithm", has nothing to do with how you access the market. One could use an algorithm to rebalance quarterly one's IRA portfolio. Or one could use an algorithm to try and arbitrage ETFs on a second by second basis. Both of which one could do from a simple household. Once you start talking about execution speeds faster than a second then you're talking about access to the markets in a way that simple households would not be capable of.

Ok,thanks.

You've come to the right place to learn more about algorithmic investing. Quantopian is a place where any individual, regardless of their background, can research, backtest, and trade in the markets using code.

There are many different forms and flavors of trading strategies, you can read this blog post explaining common strategies. And there is another blog post sharing the top 25 most popular algorithms shared in the community (from last May). This should help clarify the different options.

As mentioned, any person (or household) can trade on Quantopian. If you have a strategy, code it in Python in the IDE, and backtest it to see how the algo would have performed in the past. You can test any US stock and ETF (over 8,000 tickers), against 13 years of minutely historical data. When you find an algorithm you like, you can paper trade it in today's market. Since historical performance doesn't always correlate to today's performance, this is a risk-free way to simulate your algo in current market conditions. When you're ready, you can connect your broker account (we support Interactive Brokers and soon-to-be E*Trade) for real money trading. There are no Quantopian extra-costs for live trading, and the limits are only those imposed by the broker.

Hopefully this helps clarify the platform and capabilities for algorithmic investing.

I would like to reiterate that everything done here is black-box algorithmic investing/quantitative trading. His original question was related to algorithmic execution. He mentioned two of the best books out there for algorithmic execution. Simple households without DMA will only be able to effect the most basic execution algorithms; almost everything he reads in those books does not apply to Quantopian algorithms, and only some of it is possible even when writing one's own stack to a retail broker like IB.

Um, Simon, I read this line:

Hello quantopians, I need to clear some things regarding algorithmic trading.

And took him at his word.
Maybe if you explain the difference to him he could clarify exactly what he thinks he's dealing with.

Well, the definitions are ambiguous, nobody can agree on them anyway. I was working from the context of the rest of his post:

from what I have gathered there are Impact Driven Algorithms,Cost
Driven Algorithms,Opportunistic Algorithms,Other(Multi leg,Volatility
Driven,GWAP).

These are all strictly execution algorithms.

Where do Mean Reversion and Momentum strategies belong?

These are aspects of possible Black Box trading algorithms, or hybrids.

Can a simple household use algorithmic trading or do they rely
brokerage firm to execute their strategies?

A simple household will almost certainly rely on a brokerage firm to execute their strategies; they will not be using a home-made algorithm to jockey for queue position nor will they implement their own VWAP algorithm.

Is algorithmic trading dependent on Direct Market Access(DMA) or is it
not necessary?

Algorithmic trading, assuming we are talking about algorithmic execution, will be severely handicapped without Direct Market Access, but it's not impossible. It is, however, practically impossible on a platform like Quantopian, where you have no event-driven opportunities to change or cancel orders. Algorithmic investing, quantitative/automated trading, these are a different kettle of fish, this stuff is definitely possible without DMA, and Quantopian is proof enough of that.

Are there limits to what you can trade through algorithmic trading?

It's difficult to write automated execution algorithms for assets for which there is not an electronic market. OTC derivatives, for instance.

I'm using Barry Johnson's Algorithmic Trading and DMA book.

Excellent books, I'd say the best, for algorithmic execution.

I am interested in Algorithmic Trading and not High Frequency
Trading/BlackBox

Great, because that's what I have been assuming! Black Box trading is also known as automated trading, or quantitative trading, or algorithmic investing. It is characterized by a completely automated process from trading signal generation to position management. It might not involve any custom algorithmic execution, because, as you say, it's difficult for simple households and retail investors to implement fast-enough event-driven strategies. Typically, people rely on their brokers to execute their orders. High-frequency trading is just Black Box trading except very fast, where other challenges come into play, such as hardware latency, physical location, the speed of light and so forth.

^Thanks a lot for the detailed post, but I have one more question. You mentioned that everything done here is "Blackbox investing/quantitive trading" but isn't most of that stuff done on C++?

The language you choose to write a trading system is irrelevant. The phrase "black box" refers anything of which the innards are unknown or unknowable, and is only judged on its inputs and outputs. In this respect, every fully automated trading system is a black box to everyone except the author, and sometimes not even then (in the case of machine learning systems). The algorithms which I entered into the Quantopian Open are a black box to everyone else in the world; they don't know how they work, just that they do, or don't.

These days blackbox is commonly used as a derisively term referring to HFT where even its creators can't readily explain every trade. This as opposed to something like a long-term systemized trend-following CTA where a narrative is readily available. Even though both may be 100% automated and blackbox is the classical(correct) sense.

Aslan, you might want to clarify wanting strategies by trade and holding frequency. There is algorithmic trading at all levels of execution.